RFC1716

From RFC-Wiki

Network Working Group P. Almquist, Author Request for Comments: 1716 Consultant Category: Informational F. Kastenholz, Editor

                                                  FTP Software, Inc.
                                                       November 1994
              Towards Requirements for IP Routers

Status of this Memo

This memo provides information for the Internet community. This memo does not specify an Internet standard of any kind. Distribution of this memo is unlimited.

Almquist & Kastenholz [Page i]

Almquist & Kastenholz [Page ii]

Almquist & Kastenholz [Page iii]

Almquist & Kastenholz [Page iv]

Contents

INTERMEDIATE SYSTEM TO INTERMEDIATE SYSTEM - DUAL IS-IS

Almquist & Kastenholz [Page v]

Almquist & Kastenholz [Page vi]

PREFACE

This document is a snapshot of the work of the Router Requirements working group as of November 1991. At that time, the working group had essentially finished its task. There were some final technical matters to be nailed down, and a great deal of editing needed to be done in order to get the document ready for publication. Unfortunately, these tasks were never completed.

At the request of the Internet Area Director, the current editor took the last draft of the document and, after consulting the mailing list archives, meeting minutes, notes, and other members of the working group, edited the document to its current form. This effort included the following tasks: 1) Deleting all the parenthetical material (such as editor's comments). Useful information was turned into DISCUSSION sections, the rest was deleted. 2) Completing the tasks listed in the last draft's To be Done sections. As a part of this task, a new "to be done" list was developed and included as an appendix to the current document. 3) Rolling Philip Almquist's "Ruminations on the Next Hop" and "Ruminations on Route Leaking" into this document. These represent significant work and should be kept. 4) Fulfilling the last intents of the working group as determined from the archival material. The intent of this effort was to get the document into a form suitable for publication as an Historical RFC so that the significant work which went into the creation of this document would be preserved.

The content and form of this document are due, in large part, to the working group's chair, and document's original editor and author: Philip Almquist. Without his efforts, this document would not exist.

INTRODUCTION

The goal of this work is to replace RFC-1009, Requirements for Internet Gateways ([INTRO:1]) with a new document.

This memo is an intermediate step toward that goal. It defines and discusses requirements for devices which perform the network layer forwarding function of the Internet protocol suite. The Internet community usually refers to such devices as IP routers or simply routers; The OSI community refers to such devices as intermediate systems. Many older Internet documents refer to these devices as gateways, a name which more recently has largely passed out of favor to avoid confusion with application gateways.

An IP router can be distinguished from other sorts of packet switching devices in that a router examines the IP protocol header as part of the switching process. It generally has to modify the IP header and to strip off and replace the Link Layer framing.

The authors of this memo recognize, as should its readers, that many routers support multiple protocol suites, and that support for multiple protocol suites will be required in increasingly large parts of the Internet in the future. This memo, however, does not attempt to specify Internet requirements for protocol suites other than TCP/IP.

This document enumerates standard protocols that a router connected to the Internet must use, and it incorporates by reference the RFCs and other documents describing the current specifications for these protocols. It corrects errors in the referenced documents and adds additional discussion and guidance for an implementor.

For each protocol, this final version of this memo also contains an explicit set of requirements, recommendations, and options. The reader must understand that the list of requirements in this memo is incomplete by itself; the complete set of requirements for an Internet protocol router is primarily defined in the standard protocol specification documents, with the corrections, amendments, and supplements contained in this memo.

This memo should be read in conjunction with the Requirements for Internet Hosts RFCs ([INTRO:2] and [INTRO:3]). Internet hosts and routers must both be capable of originating IP datagrams and receiving IP datagrams destined for them. The major distinction between Internet hosts and routers is that routers are required to implement forwarding algorithms and Internet hosts do not require forwarding capabilities. Any Internet host acting as a router must adhere to the requirements contained in the final version of this memo.

The goal of open system interconnection dictates that routers must function correctly as Internet hosts when necessary. To achieve this, this memo provides guidelines for such instances. For simplification and ease of document updates, this memo tries to avoid overlapping discussions of host requirements with [INTRO:2] and [INTRO:3] and incorporates the relevant requirements of those documents by reference. In some cases the requirements stated in [INTRO:2] and [INTRO:3] are superseded by the final version of this document.

A good-faith implementation of the protocols produced after careful reading of the RFCs, with some interaction with the Internet technical community, and that follows good communications software engineering practices, should differ from the requirements of this memo in only minor ways. Thus, in many cases, the requirements in this document are already stated or implied in the standard protocol documents, so that their inclusion here is, in a sense, redundant. However, they were included because some past implementation has made the wrong choice, causing problems of interoperability, performance, and/or robustness.

This memo includes discussion and explanation of many of the requirements and recommendations. A simple list of requirements would be dangerous, because:

o Some required features are more important than others, and some features are optional.

o Some features are critical in some applications of routers but irrelevant in others.

o There may be valid reasons why particular vendor products that are designed for restricted contexts might choose to use different specifications.

However, the specifications of this memo must be followed to meet the general goal of arbitrary router interoperation across the diversity and complexity of the Internet. Although most current implementations fail to meet these requirements in various ways, some minor and some major, this specification is the ideal towards which we need to move.

These requirements are based on the current level of Internet architecture. This memo will be updated as required to provide additional clarifications or to include additional information in those areas in which specifications are still evolving.

Reading this Document

Organization

  This memo emulates the layered organization used by [INTRO:2] and
  [INTRO:3].  Thus, Chapter 2 describes the layers found in the
  Internet architecture.  Chapter 3 covers the Link Layer.  Chapters
  4 and 5 are concerned with the Internet Layer protocols and
  forwarding algorithms.  Chapter 6 covers the Transport Layer.
  Upper layer protocols are divided between Chapter 7, which
  discusses the protocols which routers use to exchange routing
  information with each other, Chapter 8, which discusses network
  management, and Chapter 9, which discusses other upper layer
  protocols.  The final chapter covers operations and maintenance
  features.  This organization was chosen for simplicity, clarity,
  and consistency with the Host Requirements RFCs.  Appendices to
  this memo include a bibliography, a glossary, and some conjectures
  about future directions of router standards.
  In describing the requirements, we assume that an implementation
  strictly mirrors the layering of the protocols.  However, strict
  layering is an imperfect model, both for the protocol suite and
  for recommended implementation approaches.  Protocols in different
  layers interact in complex and sometimes subtle ways, and
  particular functions often involve multiple layers.  There are
  many design choices in an implementation, many of which involve
  creative breaking of strict layering.  Every implementor is urged
  to read [INTRO:4] and [INTRO:5].
  In general, each major section of this memo is organized into the
  following subsections:
  (1)  Introduction
  (2)  Protocol Walk-Through - considers the protocol specification
       documents section-by-section, correcting errors, stating
       requirements that may be ambiguous or ill-defined, and
       providing further clarification or explanation.
  (3)  Specific Issues - discusses protocol design and
       implementation issues that were not included in the walk-
       through.
  Under many of the individual topics in this memo, there is
  parenthetical material labeled DISCUSSION or IMPLEMENTATION. This
  material is intended to give a justification, clarification or
  explanation to the preceding requirements text.  The
  implementation material contains suggested approaches that an
  implementor may want to consider.  The DISCUSSION and
  IMPLEMENTATION sections are not part of the standard.

Requirements

  In this memo, the words that are used to define the significance
  of each particular requirement are capitalized.  These words are:
  o  MUST
     This word means that the item is an absolute requirement of the
     specification.
  o  MUST IMPLEMENT
     This phrase means that this specification requires that the
     item be implemented, but does not require that it be enabled by
     default.
  o  MUST NOT
     This phrase means that the item is an absolute prohibition of
     the specification.
  o  SHOULD
     This word means that there may exist valid reasons in
     particular circumstances to ignore this item, but the full
     implications should be understood and the case carefully
     weighed before choosing a different course.
  o  SHOULD IMPLEMENT
     This phrase is similar in meaning to SHOULD, but is used when
     we recommend that a particular feature be provided but does not
     necessarily recommend that it be enabled by default.
  o  SHOULD NOT
     This phrase means that there may exist valid reasons in
     particular circumstances when the described behavior is
     acceptable or even useful, but the full implications should be
     understood and the case carefully weighed before implementing
     any behavior described with this label.
  o  MAY
     This word means that this item is truly optional.  One vendor
     may choose to include the item because a particular marketplace
     requires it or because it enhances the product, for example;
     another vendor may omit the same item.

Compliance

  Some requirements are applicable to all routers.  Other
  requirements are applicable only to those which implement
  particular features or protocols.  In the following paragraphs,
  Relevant refers to the union of the requirements applicable to all
  routers and the set of requirements applicable to a particular
  router because of the set of features and protocols it has
  implemented.
  Note that not all Relevant requirements are stated directly in
  this memo.  Various parts of this memo incorporate by reference
  sections of the Host Requirements specification, [INTRO:2] and
  [INTRO:3].  For purposes of determining compliance with this memo,
  it does not matter whether a Relevant requirement is stated
  directly in this memo or merely incorporated by reference from one
  of those documents.
  An implementation is said to be conditionally compliant if it
  satisfies all of the Relevant MUST, MUST IMPLEMENT, and MUST NOT
  requirements.  An implementation is said to be unconditionally
  compliant if it is conditionally compliant and also satisfies all
  of the Relevant SHOULD, SHOULD IMPLEMENT, and SHOULD NOT
  requirements.  An implementation is not compliant if it is not
  conditionally compliant (i.e., it fails to satisfy one or more of
  the Relevant MUST, MUST IMPLEMENT, or MUST NOT requirements).
  For any of the SHOULD and SHOULD NOT requirements, a router may
  provide a configuration option that will cause the router to act
  other than as specified by the requirement.  Having such a
  configuration option does not void a router's claim to
  unconditional compliance as long as the option has a default
  setting, and that leaving the option at its default setting causes
  the router to operate in a manner which conforms to the
  requirement.
  Likewise, routers may provide, except where explicitly prohibited
  by this memo, options which cause them to violate MUST or MUST NOT
  requirements.  A router which provides such options is compliant
  (either fully or conditionally) if and only if each such option
  has a default setting which causes the router to conform to the
  requirements of this memo.  Please note that the authors of this
  memo, although aware of market realities, strongly recommend
  against provision of such options.  Requirements are labeled MUST
  or MUST NOT because experts in the field have judged them to be
  particularly important to interoperability or proper functioning
  in the Internet.  Vendors should weigh carefully the customer
  support costs of providing options which violate those rules.
  Of course, this memo is not a complete specification of an IP
  router, but rather is closer to what in the OSI world is called a
  profile.  For example, this memo requires that a number of
  protocols be implemented.  Although most of the contents of their
  protocol specifications are not repeated in this memo,
  implementors are nonetheless required to implement the protocols
  according to those specifications.

Relationships to Other Standards

There are several reference documents of interest in checking the current status of protocol specifications and standardization:

 o  INTERNET OFFICIAL PROTOCOL STANDARDS
    This document describes the Internet standards process and lists
    the standards status of the protocols.  As of this writing, the
    current version of this document is STD 1, RFC 1610, [ARCH:7].
    This document is periodically re-issued.  You should always
    consult an RFC repository and use the latest version of this
    document.
 o  Assigned Numbers
    This document lists the assigned values of the parameters used
    in the various protocols.  For example, IP protocol codes, TCP
    port numbers, Telnet Option Codes, ARP hardware types, and
    Terminal Type names.  As of this writing, the current version of
    this document is STD 2, RFC 1700, [INTRO:7].  This document is
    periodically re-issued.  You should always consult an RFC
    repository and use the latest version of this document.
 o  Host Requirements
    This pair of documents reviews the specifications that apply to
    hosts and supplies guidance and clarification for any
    ambiguities.  Note that these requirements also apply to
    routers, except where otherwise specified in this memo.  As of
    this writing (December, 1993) the current versions of these
    documents are RFC 1122 and RFC 1123, (STD 3) [INTRO:2], and
    [INTRO:3] respectively.
 o  Router Requirements (formerly Gateway Requirements)
    This memo.
 Note that these documents are revised and updated at different
 times; in case of differences between these documents, the most
 recent must prevail.
 These and other Internet protocol documents may be obtained from
 the:
 The InterNIC
 DS.INTERNIC.NET
 InterNIC Directory and Database Service
 +1 (800) 444-4345 or +1 (619) 445-4600
 [email protected]

General Considerations

There are several important lessons that vendors of Internet software have learned and which a new vendor should consider seriously.

Continuing Internet Evolution

  The enormous growth of the Internet has revealed problems of
  management and scaling in a large datagram-based packet
  communication system.  These problems are being addressed, and as
  a result there will be continuing evolution of the specifications
  described in this memo.  New routing protocols, algorithms, and
  architectures are constantly being developed.  New and additional
  internet-layer protocols are also constantly being devised.
  Because routers play such a crucial role in the Internet, and
  because the number of routers deployed in the Internet is much
  smaller than the number of hosts, vendors should expect that
  router standards will continue to evolve much more quickly than
  host standards.  These changes will be carefully planned and
  controlled since there is extensive participation in this planning
  by the vendors and by the organizations responsible for operation
  of the networks.
  Development, evolution, and revision are characteristic of
  computer network protocols today, and this situation will persist
  for some years.  A vendor who develops computer communications
  software for the Internet protocol suite (or any other protocol
  suite!) and then fails to maintain and update that software for
  changing specifications is going to leave a trail of unhappy
  customers.  The Internet is a large communication network, and the
  users are in constant contact through it.  Experience has shown
  that knowledge of deficiencies in vendor software propagates
  quickly through the Internet technical community.

Robustness Principle

  At every layer of the protocols, there is a general rule (from
  [TRANS:2] by Jon Postel) whose application can lead to enormous
  benefits in robustness and interoperability:
                   Be conservative in what you do,
              be liberal in what you accept from others.
  Software should be written to deal with every conceivable error,
  no matter how unlikely; sooner or later a packet will come in with
  that particular combination of errors and attributes, and unless
  the software is prepared, chaos can ensue.  In general, it is best
  to assume that the network is filled with malevolent entities that
  will send packets designed to have the worst possible effect.
  This assumption will lead to suitably protective design.  The most
  serious problems in the Internet have been caused by unforeseen
  mechanisms triggered by low probability events; mere human malice
  would never have taken so devious a course!
  Adaptability to change must be designed into all levels of router
  software.  As a simple example, consider a protocol specification
  that contains an enumeration of values for a particular header
  field - e.g., a type field, a port number, or an error code; this
  enumeration must be assumed to be incomplete.  If the protocol
  specification defines four possible error codes, the software must
  not break when a fifth code shows up.  An undefined code might be
  logged, but it must not cause a failure.
  The second part of the principle is almost as important: software
  on hosts or other routers may contain deficiencies that make it
  unwise to exploit legal but obscure protocol features.  It is
  unwise to stray far from the obvious and simple, lest untoward
  effects result elsewhere.  A corollary of this is watch out for
  misbehaving hosts; router software should be prepared to survive
  in the presence of misbehaving hosts.  An important function of
  routers in the Internet is to limit the amount of disruption such
  hosts can inflict on the shared communication facility.

Error Logging

  The Internet includes a great variety of systems, each
  implementing many protocols and protocol layers, and some of these
  contain bugs and misfeatures in their Internet protocol software.
  As a result of complexity, diversity, and distribution of
  function, the diagnosis of problems is often very difficult.
  Problem diagnosis will be aided if routers include a carefully
  designed facility for logging erroneous or strange events.  It is
  important to include as much diagnostic information as possible
  when an error is logged.  In particular, it is often useful to
  record the header(s) of a packet that caused an error.  However,
  care must be taken to ensure that error logging does not consume
  prohibitive amounts of resources or otherwise interfere with the
  operation of the router.
  There is a tendency for abnormal but harmless protocol events to
  overflow error logging files; this can be avoided by using a
  circular log, or by enabling logging only while diagnosing a known
  failure.  It may be useful to filter and count duplicate
  successive messages.  One strategy that seems to work well is to
  both:
  o  Always count abnormalities and make such counts accessible
     through the management protocol (see Chapter 8); and
  o  Allow the logging of a great variety of events to be
     selectively enabled.  For example, it might useful to be able
     to log everything or to log everything for host X.
  This topic is further discussed in [MGT:5].

Configuration

  In an ideal world, routers would be easy to configure, and perhaps
  even entirely self-configuring.  However, practical experience in
  the real world suggests that this is an impossible goal, and that
  in fact many attempts by vendors to make configuration easy
  actually cause customers more grief than they prevent.  As an
  extreme example, a router designed to come up and start routing
  packets without requiring any configuration information at all
  would almost certainly choose some incorrect parameter, possibly
  causing serious problems on any networks unfortunate enough to be
  connected to it.
  Often this memo requires that a parameter be a configurable
  option.  There are several reasons for this.  In a few cases there
  currently is some uncertainty or disagreement about the best value
  and it may be necessary to update the recommended value in the
  future.  In other cases, the value really depends on external
  factors - e.g., the distribution of its communication load, or the
  speeds and topology of nearby networks - and self-tuning
  algorithms are unavailable and may be insufficient.  In some
  cases, configurability is needed because of administrative
  requirements.
  Finally, some configuration options are required to communicate
  with obsolete or incorrect implementations of the protocols,
  distributed without sources, that persist in many parts of the
  Internet.  To make correct systems coexist with these faulty
  systems, administrators must occasionally misconfigure the correct
  systems.  This problem will correct itself gradually as the faulty
  systems are retired, but cannot be ignored by vendors.
  When we say that a parameter must be configurable, we do not
  intend to require that its value be explicitly read from a
  configuration file at every boot time.  For many parameters, there
  is one value that is appropriate for all but the most unusual
  situations.  In such cases, it is quite reasonable that the
  parameter default to that value if not explicitly set.
  This memo requires a particular value for such defaults in some
  cases.  The choice of default is a sensitive issue when the
  configuration item controls accommodation of existing, faulty,
  systems.  If the Internet is to converge successfully to complete
  interoperability, the default values built into implementations
  must implement the official protocol, not misconfigurations to
  accommodate faulty implementations.  Although marketing
  considerations have led some vendors to choose misconfiguration
  defaults, we urge vendors to choose defaults that will conform to
  the standard.
  Finally, we note that a vendor needs to provide adequate
  documentation on all configuration parameters, their limits and
  effects.

Algorithms

In several places in this memo, specific algorithms that a router ought to follow are specified. These algorithms are not, per se, required of the router. A router need not implement each algorithm as it is written in this document. Rather, an implementation must present a behavior to the external world that is the same as a strict, literal, implementation of the specified algorithm.

Algorithms are described in a manner that differs from the way a good implementor would implement them. For expository purposes, a style that emphasizes conciseness, clarity, and independence from implementation details has been chosen. A good implementor will choose algorithms and implementation methods which produce the same results as these algorithms, but may be more efficient or less general.

We note that the art of efficient router implementation is outside of the scope of this memo.

INTERNET ARCHITECTURE

This chapter does not contain any requirements. However, it does contain useful background information on the general architecture of the Internet and of routers.

General background and discussion on the Internet architecture and supporting protocol suite can be found in the DDN Protocol Handbook [ARCH:1]; for background see for example [ARCH:2], [ARCH:3], and [ARCH:4]. The Internet architecture and protocols are also covered in an ever-growing number of textbooks, such as [ARCH:5] and [ARCH:6].

Introduction

The Internet system consists of a number of interconnected packet networks supporting communication among host computers using the Internet protocols. These protocols include the Internet Protocol (IP), the Internet Control Message Protocol (ICMP), the Internet Group Management Protocol (IGMP), and a variety transport and application protocols that depend upon them. As was described in Section [1.2], the Internet Engineering Steering Group periodically releases an Official Protocols memo listing all of the Internet protocols.

All Internet protocols use IP as the basic data transport mechanism. IP is a datagram, or connectionless, internetwork service and includes provision for addressing, type-of-service specification, fragmentation and reassembly, and security. ICMP and IGMP are considered integral parts of IP, although they are architecturally layered upon IP. ICMP provides error reporting, flow control, first-hop router redirection, and other maintenance and control functions. IGMP provides the mechanisms by which hosts and routers can join and leave IP multicast groups.

Reliable data delivery is provided in the Internet protocol suite by Transport Layer protocols such as the Transmission Control Protocol (TCP), which provides end-end retransmission, resequencing and connection control. Transport Layer connectionless service is provided by the User Datagram Protocol (UDP).

Elements of the Architecture

Protocol Layering

  To communicate using the Internet system, a host must implement
  the layered set of protocols comprising the Internet protocol
  suite.  A host typically must implement at least one protocol from
  each layer.
  The protocol layers used in the Internet architecture are as
  follows [ARCH:7]:
  o  Application Layer
     The Application Layer is the top layer of the Internet protocol
     suite.  The Internet suite does not further subdivide the
     Application Layer, although some application layer protocols do
     contain some internal sub-layering.  The application layer of
     the Internet suite essentially combines the functions of the
     top two layers - Presentation and Application - of the OSI
     Reference Model [ARCH:8].  The Application Layer in the
     Internet protocol suite also includes some of the function
     relegated to the Session Layer in the OSI Reference Model.
     We distinguish two categories of application layer protocols:
     user protocols that provide service directly to users, and
     support protocols that provide common system functions.  The
     most common Internet user protocols are:
     - Telnet (remote login)
     - FTP (file transfer)
     - SMTP (electronic mail delivery)
     There are a number of other standardized user protocols and
     many private user protocols.
     Support protocols, used for host name mapping, booting, and
     management, include SNMP, BOOTP, TFTP, the Domain Name System
     (DNS) protocol, and a variety of routing protocols.
     Application Layer protocols relevant to routers are discussed
     in chapters 7, 8, and 9 of this memo.
  o  Transport Layer
     The Transport Layer provides end-to-end communication services.
     This layer is roughly equivalent to the Transport Layer in the
     OSI Reference Model, except that it also incorporates some of
     OSI's Session Layer establishment and destruction functions.
     There are two primary Transport Layer protocols at present:
     - Transmission Control Protocol (TCP)
     - User Datagram Protocol (UDP)
     TCP is a reliable connection-oriented transport service that
     provides end-to-end reliability, resequencing, and flow
     control.  UDP is a connectionless (datagram) transport service.
     Other transport protocols have been developed by the research
     community, and the set of official Internet transport protocols
     may be expanded in the future.
     Transport Layer protocols relevant to routers are discussed in
     Chapter 6.
  o  Internet Layer
     All Internet transport protocols use the Internet Protocol (IP)
     to carry data from source host to destination host.  IP is a
     connectionless or datagram internetwork service, providing no
     end-to-end delivery guarantees. IP datagrams may arrive at the
     destination host damaged, duplicated, out of order, or not at
     all.  The layers above IP are responsible for reliable delivery
     service when it is required.  The IP protocol includes
     provision for addressing, type-of-service specification,
     fragmentation and reassembly, and security.
     The datagram or connectionless nature of IP is a fundamental
     and characteristic feature of the Internet architecture.
     The Internet Control Message Protocol (ICMP) is a control
     protocol that is considered to be an integral part of IP,
     although it is architecturally layered upon IP, i.e., it uses
     IP to carry its data end-to-end.  ICMP provides error
     reporting, congestion reporting, and first-hop router
     redirection.
     The Internet Group Management Protocol (IGMP) is an Internet
     layer protocol used for establishing dynamic host groups for IP
     multicasting.
     The Internet layer protocols IP, ICMP, and IGMP are discussed
     in chapter 4.
  o  Link Layer
     To communicate on its directly-connected network, a host must
     implement the communication protocol used to interface to that
     network.  We call this a Link Layer layer protocol.
     Some older Internet documents refer to this layer as the
     Network Layer, but it is not the same as the Network Layer in
     the OSI Reference Model.
     This layer contains everything below the Internet Layer.
     Protocols in this Layer are generally outside the scope of
     Internet standardization; the Internet (intentionally) uses
     existing standards whenever possible.  Thus, Internet Link
     Layer standards usually address only address resolution and
     rules for transmitting IP packets over specific Link Layer
     protocols.  Internet Link Layer standards are discussed in
     chapter 3.

Networks

  The constituent networks of the Internet system are required to
  provide only packet (connectionless) transport.  According to the
  IP service specification, datagrams can be delivered out of order,
  be lost or duplicated, and/or contain errors.
  For reasonable performance of the protocols that use IP (e.g.,
  TCP), the loss rate of the network should be very low.  In
  networks providing connection-oriented service, the extra
  reliability provided by virtual circuits enhances the end-end
  robustness of the system, but is not necessary for Internet
  operation.
  Constituent networks may generally be divided into two classes:
    o  Local-Area Networks (LANs)
       LANs may have a variety of designs.  In general, a LAN will
       cover a small geographical area (e.g., a single building or
       plant site) and provide high bandwidth with low delays.  LANs
       may be passive (similar to Ethernet) or they may be active
       (such as ATM).
    o  Wide-Area Networks (WANs)
       Geographically-dispersed hosts and LANs are interconnected by
       wide-area networks, also called long-haul networks.  These
       networks may have a complex internal structure of lines and
       packet-switches, or they may be as simple as point-to-point
       lines.

Routers

  In the Internet model, constituent networks are connected together
  by IP datagram forwarders which are called routers or IP routers.
  In this document, every use of the term router is equivalent to IP
  router.  Many older Internet documents refer to routers as
  gateways.
  Historically, routers have been realized with packet-switching
  software executing on a general-purpose CPU.  However, as custom
  hardware development becomes cheaper and as higher throughput is
  required, but special-purpose hardware is becoming increasingly
  common.  This specification applies to routers regardless of how
  they are implemented.
  A router is connected to two or more networks, appearing to each
  of these networks as a connected host.  Thus, it has (at least)
  one physical interface and (at least) one IP address on each of
  the connected networks (this ignores the concept of un-numbered
  links, which is discussed in section [2.2.7]).  Forwarding an IP
  datagram generally requires the router to choose the address of
  the next-hop router or (for the final hop) the destination host.
  This choice, called routing, depends upon a routing database
  within the router.  The routing database is also sometimes known
  as a routing table or forwarding table.
  The routing database should be maintained dynamically to reflect
  the current topology of the Internet system.  A router normally
  accomplishes this by participating in distributed routing and
  reachability algorithms with other routers.
  Routers provide datagram transport only, and they seek to minimize
  the state information necessary to sustain this service in the
  interest of routing flexibility and robustness.
  Packet switching devices may also operate at the Link Layer; such
  devices are usually called bridges. Network segments which are
  connected by bridges share the same IP network number, i.e., they
  logically form a single IP network.  These other devices are
  outside of the scope of this document.
  Another variation on the simple model of networks connected with
  routers sometimes occurs: a set of routers may be interconnected
  with only serial lines, to form a network in which the packet
  switching is performed at the Internetwork (IP) Layer rather than
  the Link Layer.

Autonomous Systems

  For technical, managerial, and sometimes political reasons, the
  routers of the Internet system are grouped into collections called
  autonomous systems.  The routers included in a single autonomous
  system (AS) are expected to:
  o  Be under the control of a single operations and maintenance
     (O&M) organization;
  o  Employ common routing protocols among themselves, to
     dynamically maintain their routing databases.
  A number of different dynamic routing protocols have been
  developed (see Section [7.2]); the routing protocol within a
  single AS is generically called an interior gateway protocol or
  IGP.
  An IP datagram may have to traverse the routers of two or more ASs
  to reach its destination, and the ASs must provide each other with
  topology information to allow such forwarding.  An exterior
  gateway protocol (generally BGP or EGP) is used for this purpose.

Addresses and Subnets

  An IP datagram carries 32-bit source and destination addresses,
  each of which is partitioned into two parts - a constituent
  network number and a host number on that network.  Symbolically:
     IP-address  ::=  { <Network-number>, <Host-number> }
  To finally deliver the datagram, the last router in its path must
  map the Host-number (or rest) part of an IP address into the
  physical address of a host connection to the constituent network.
  This simple notion has been extended by the concept of subnets,
  which were introduced in order to allow arbitrary complexity of
  interconnected LAN structures within an organization, while
  insulating the Internet system against explosive growth in network
  numbers and routing complexity.  Subnets essentially provide a
  multi-level hierarchical routing structure for the Internet
  system.  The subnet extension, described in [INTERNET:2], is now a
  required part of the Internet architecture.  The basic idea is to
  partition the <Host-number> field into two parts: a subnet number,
  and a true host number on that subnet:
     IP-address  ::=
       { <Network-number>, <Subnet-number>, <Host-number> }
  The interconnected physical networks within an organization will
  be given the same network number but different subnet numbers.
  The distinction between the subnets of such a subnetted network is
  normally not visible outside of that network.  Thus, routing in
  the rest of the Internet will be based only upon the <Network-
  number> part of the IP destination address; routers outside the
  network will combine <Subnet-number> and <Host-number> together to
  form an uninterpreted rest part of the 32-bit IP address.  Within
  the subnetted network, the routers must route on the basis of an
  extended network number:
     { <Network-number>, <Subnet-number> }
  Under certain circumstances, it may be desirable to support
  subnets of a particular network being interconnected only via a
  path which is not part of the subnetted network.  Even though many
  IGP's and no EGP's currently support this configuration
  effectively, routers need to be able to support this configuration
  of subnetting (see Section [4.2.3.4]).  In general, routers should
  not make assumptions about what are subnets and what are not, but
  simply ignore the concept of Class in networks, and treat each
  route as a { network, mask }-tuple.
  DISCUSSION:
     It is becoming clear that as the Internet grows larger and
     larger, the traditional uses of Class A, B, and C networks will
     be modified in order to achieve better use of IP's 32-bit
     address space.  Classless Interdomain Routing (CIDR)
     [INTERNET:15] is a method currently being deployed in the
     Internet backbones to achieve this added efficiency.  CIDR
     depends on the ability of assigning and routing to networks
     that are not based on Class A, B, or C networks.  Thus, routers
     should always treat a route as a network with a mask.
  Furthermore, for similar reasons, a subnetted network need not
  have a consistent subnet mask through all parts of the network.
  For example, one subnet may use an 8 bit subnet mask, another 10
  bit, and another 6 bit.  Routers need to be able to support this
  type of configuration (see Section [4.2.3.4]).
  The bit positions containing this extended network number are
  indicated by a 32-bit mask called the subnet mask; it is
  recommended but not required that the <Subnet-number> bits be
  contiguous and fall between the <Network-number> and the <Host-
  number> fields.  No subnet should be assigned the value zero or -1
  (all one bits).
  Although the inventors of the subnet mechanism probably expected
  that each piece of an organization's network would have only a
  single subnet number, in practice it has often proven necessary or
  useful to have several subnets share a single physical cable.
  There are special considerations for the router when a connected
  network provides a broadcast or multicast capability; these will
  be discussed later.

IP Multicasting

  IP multicasting is an extension of Link Layer multicast to IP
  internets.  Using IP multicasts, a single datagram can be
  addressed to multiple hosts. This collection of hosts is called a
  multicast group.  Each multicast group is represented as a Class D
  IP address.  An IP datagram sent to the group is to be delivered
  to each group member with the same best-effort delivery as that
  provided for unicast IP traffic.  The sender of the datagram does
  not itself need to be a member of the destination group.
  The semantics of IP multicast group membership are defined in
  [INTERNET:4].  That document describes how hosts and routers join
  and leave multicast groups.  It also defines a protocol, the
  Internet Group Management Protocol (IGMP), that monitors IP
  multicast group membership.
  Forwarding of IP multicast datagrams is accomplished either
  through static routing information or via a multicast routing
  protocol.  Devices that forward IP multicast datagrams are called
  multicast routers. They may or may not also forward IP unicasts.
  In general, multicast datagrams are forwarded on the basis of both
  their source and destination addresses.  Forwarding of IP
  multicast packets is described in more detail in Section [5.2.1].
  Appendix D discusses multicast routing protocols.

Unnumbered Lines and Networks and Subnets

  Traditionally, each network interface on an IP host or router has
  its own IP address.  Over the years, people have observed that
  this can cause inefficient use of the scarce IP address space,
  since it forces allocation of an IP network number, or at least a
  subnet number, to every point-to-point link.
  To solve this problem, a number of people have proposed and
  implemented the concept of unnumbered serial lines.  An unnumbered
  serial line does not have any IP network or subnet number
  associated with it.  As a consequence, the network interfaces
  connected to an unnumbered serial line do not have IP addresses.
  Because the IP architecture has traditionally assumed that all
  interfaces had IP addresses, these unnumbered interfaces cause
  some interesting dilemmas.  For example, some IP options (e.g.
  Record Route) specify that a router must insert the interface
  address into the option, but an unnumbered interface has no IP
  address.  Even more fundamental (as we shall see in chapter 5) is
  that routes contain the IP address of the next hop router.  A
  router expects that that IP address will be on an IP (sub)net that
  the router is connected to.  That assumption is of course violated
  if the only connection is an unnumbered serial line.
  To get around these difficulties, two schemes have been invented.
  The first scheme says that two routers connected by an unnumbered
  serial line aren't really two routers at all, but rather two
  half-routers which together make up a single (virtual) router.
  The unnumbered serial line is essentially considered to be an
  internal bus in the virtual router.  The two halves of the virtual
  router must coordinate their activities in such a way that they
  act exactly like a single router.
  This scheme fits in well with the IP architecture, but suffers
  from two important drawbacks.  The first is that, although it
  handles the common case of a single unnumbered serial line, it is
  not readily extensible to handle the case of a mesh of routers and
  unnumbered serial lines.  The second drawback is that the
  interactions between the half routers are necessarily complex and
  are not standardized, effectively precluding the connection of
  equipment from different vendors using unnumbered serial lines.
  Because of these drawbacks, this memo has adopted an alternative
  scheme, which has been invented multiple times but which is
  probably originally attributable to Phil Karn.  In this scheme, a
  router which has unnumbered serial lines also has a special IP
  address, called a router-id in this memo.  The router-id is one of
  the router's IP addresses (a router is required to have at least
  one IP address).  This router-id is used as if it is the IP
  address of all unnumbered interfaces.

Notable Oddities

Embedded Routers
     A router may be a stand-alone computer system, dedicated to its
     IP router functions.  Alternatively, it is possible to embed
     router functions within a host operating system which supports
     connections to two or more networks.  The best-known example of
     an operating system with embedded router code is the Berkeley
     BSD system.  The embedded router feature seems to make
     internetting easy, but it has a number of hidden pitfalls:
     (1)  If a host has only a single constituent-network interface,
          it should not act as a router.
          For example, hosts with embedded router code that
          gratuitously forward broadcast packets or datagrams on the
          same net often cause packet avalanches.
     (2)  If a (multihomed) host acts as a router, it must implement
          ALL the relevant router requirements contained in this
          document.
          For example, the routing protocol issues and the router
          control and monitoring problems are as hard and important
          for embedded routers as for stand-alone routers.
          Since Internet router requirements and specifications may
          change independently of operating system changes, an
          administration that operates an embedded router in the
          Internet is strongly advised to have the ability to
          maintain and update the router code (e.g., this might
          require router code source).
     (3)  Once a host runs embedded router code, it becomes part of
          the Internet system.  Thus, errors in software or
          configuration can hinder communication between other
          hosts.  As a consequence, the host administrator must lose
          some autonomy.
          In many circumstances, a host administrator will need to
          disable router code embedded in the operating system, and
          any embedded router code must be organized so that it can
          be easily disabled.
     (4)  If a host running embedded router code is concurrently
          used for other services, the O&M (Operation and
          Maintenance) requirements for the two modes of use may be
          in serious conflict.
          For example, router O&M will in many cases be performed
          remotely by an operations center; this may require
          privileged system access which the host administrator
          would not normally want to distribute.
Transparent Routers
     There are two basic models for interconnecting local-area
     networks and wide-area (or long-haul) networks in the Internet.
     In the first, the local-area network is assigned a network
     number and all routers in the Internet must know how to route
     to that network.  In the second, the local-area network shares
     (a small part of) the address space of the wide-area network.
     Routers that support this second model are called address
     sharing routers or transparent routers.  The focus of this memo
     is on routers that support the first model, but this is not
     intended to exclude the use of transparent routers.
     The basic idea of a transparent router is that the hosts on the
     local-area network behind such a router share the address space
     of the wide-area network in front of the router.  In certain
     situations this is a very useful approach and the limitations
     do not present significant drawbacks.
     The words in front and behind indicate one of the limitations
     of this approach: this model of interconnection is suitable
     only for a geographically (and topologically) limited stub
     environment.  It requires that there be some form of logical
     addressing in the network level addressing of the wide-area
     network.  All of the IP addresses in the local environment map
     to a few (usually one) physical address in the wide-area
     network.  This mapping occurs in a way consistent with the { IP
     address <-> network address } mapping used throughout the
     wide-area network.
     Multihoming is possible on one wide-area network, but may
     present routing problems if the interfaces are geographically
     or topologically separated.  Multihoming on two (or more)
     wide-area networks is a problem due to the confusion of
     addresses.
     The behavior that hosts see from other hosts in what is
     apparently the same network may differ if the transparent
     router cannot fully emulate the normal wide-area network
     service.  For example, the ARPANET used a Link Layer protocol
     that provided a Destination Dead indication in response to an
     attempt to send to a host which was powered off.  However, if
     there were a transparent router between the ARPANET and an
     Ethernet, a host on the ARPANET would not receive a Destination
     Dead indication if it sent a datagram to a host that was
     powered off and was connected to the ARPANET via the
     transparent router instead of directly.

Router Characteristics

An Internet router performs the following functions:

(1) Conforms to specific Internet protocols specified in this

    document, including the Internet Protocol (IP), Internet Control
    Message Protocol (ICMP), and others as necessary.

(2) Interfaces to two or more packet networks. For each connected

    network the router must implement the functions required by that
    network.  These functions typically include:
    o  Encapsulating and decapsulating the IP datagrams with the
       connected network framing (e.g., an Ethernet header and
       checksum),
    o  Sending and receiving IP datagrams up to the maximum size
       supported by that network, this size is the network's Maximum
       Transmission Unit or MTU,
    o  Translating the IP destination address into an appropriate
       network-level address for the connected network (e.g., an
       Ethernet hardware address), if needed, and
    o  Responding to the network flow control and error indication,
       if any.
    See chapter 3 (Link Layer).

(3) Receives and forwards Internet datagrams. Important issues in

    this process are buffer management, congestion control, and
    fairness.
    o  Recognizes various error conditions and generates ICMP error
       and information messages as required.
    o  Drops datagrams whose time-to-live fields have reached zero.
    o  Fragments datagrams when necessary to fit into the MTU of the
       next network.
    See chapter 4 (Internet Layer - Protocols) and chapter 5
    (Internet Layer - Forwarding) for more information.

(4) Chooses a next-hop destination for each IP datagram, based on

    the information in its routing database.  See chapter 5
    (Internet Layer - Forwarding) for more information.

(5) (Usually) supports an interior gateway protocol (IGP) to carry

    out distributed routing and reachability algorithms with the
    other routers in the same autonomous system.  In addition, some
    routers will need to support an exterior gateway protocol (EGP)
    to exchange topological information with other autonomous
    systems.  See chapter 7 (Application Layer - Routing Protocols)
    for more information.

(6) Provides network management and system support facilities,

    including loading, debugging, status reporting, exception
    reporting and control.  See chapter 8 (Application Layer -
    Network Management Protocols) and chapter 10 (Operation and
    Maintenance) for more information.

A router vendor will have many choices on power, complexity, and features for a particular router product. It may be helpful to observe that the Internet system is neither homogeneous nor fully- connected. For reasons of technology and geography it is growing into a global interconnect system plus a fringe of LANs around the edge. More and more these fringe LANs are becoming richly interconnected, thus making them less out on the fringe and more demanding on router requirements.

o The global interconnect system is comprised of a number of wide-

  area networks to which are attached routers of several Autonomous
  Systems (AS); there are relatively few hosts connected directly to
  the system.

o Most hosts are connected to LANs. Many organizations have

  clusters of LANs interconnected by local routers.  Each such
  cluster is connected by routers at one or more points into the
  global interconnect system.  If it is connected at only one point,
  a LAN is known as a stub network.

Routers in the global interconnect system generally require:

o Advanced Routing and Forwarding Algorithms

  These routers need routing algorithms which are highly dynamic and
  also offer type-of-service routing.  Congestion is still not a
  completely resolved issue (see Section [5.3.6]).  Improvements in
  these areas are expected, as the research community is actively
  working on these issues.

o High Availability

  These routers need to be highly reliable, providing 24 hours a
  day, 7 days a week service.  Equipment and software faults can
  have a wide-spread (sometimes global) effect.  In case of failure,
  they must recover quickly.  In any environment, a router must be
  highly robust and able to operate, possibly in a degraded state,
  under conditions of extreme congestion or failure of network
  resources.

o Advanced O&M Features

  Internet routers normally operate in an unattended mode.  They
  will typically be operated remotely from a centralized monitoring
  center.  They need to provide sophisticated means for monitoring
  and measuring traffic and other events and for diagnosing faults.

o High Performance

  Long-haul lines in the Internet today are most frequently 56 Kbps,
  DS1 (1.4Mbps), and DS3 (45Mbps) speeds.  LANs are typically
  Ethernet (10Mbps) and, to a lesser degree, FDDI (100Mbps).
  However, network media technology is constantly advancing and even
  higher speeds are likely in the future.  Full-duplex operation is
  provided at all of these speeds.

The requirements for routers used in the LAN fringe (e.g., campus networks) depend greatly on the demands of the local networks. These may be high or medium-performance devices, probably competitively procured from several different vendors and operated by an internal organization (e.g., a campus computing center). The design of these routers should emphasize low average latency and good burst performance, together with delay and type-of-service sensitive resource management. In this environment there may be less formal O&M but it will not be less important. The need for the routing mechanism to be highly dynamic will become more important as networks become more complex and interconnected. Users will demand more out of their local connections because of the speed of the global interconnects.

As networks have grown, and as more networks have become old enough

that they are phasing out older equipment, it has become increasingly imperative that routers interoperate with routers from other vendors.

Even though the Internet system is not fully interconnected, many parts of the system need to have redundant connectivity. Rich connectivity allows reliable service despite failures of communication lines and routers, and it can also improve service by shortening Internet paths and by providing additional capacity. Unfortunately, this richer topology can make it much more difficult to choose the best path to a particular destination.

Architectural Assumptions

The current Internet architecture is based on a set of assumptions about the communication system. The assumptions most relevant to routers are as follows:

o The Internet is a network of networks.

  Each host is directly connected to some particular network(s); its
  connection to the Internet is only conceptual.  Two hosts on the
  same network communicate with each other using the same set of
  protocols that they would use to communicate with hosts on distant
  networks.

o Routers don't keep connection state information.

  To improve the robustness of the communication system, routers are
  designed to be stateless, forwarding each IP packet independently
  of other packets.  As a result, redundant paths can be exploited
  to provide robust service in spite of failures of intervening
  routers and networks.
  All state information required for end-to-end flow control and
  reliability is implemented in the hosts, in the transport layer or
  in application programs.  All connection control information is
  thus co-located with the end points of the communication, so it
  will be lost only if an end point fails.  Routers effect flow
  control only indirectly, by dropping packets or increasing network
  delay.
  Note that future protocol developments may well end up putting
  some more state into routers.  This is especially likely for
  resource reservation and flows.

o Routing complexity should be in the routers.

  Routing is a complex and difficult problem, and ought to be
  performed by the routers, not the hosts.  An important objective
  is to insulate host software from changes caused by the inevitable
  evolution of the Internet routing architecture.

o The system must tolerate wide network variation.

  A basic objective of the Internet design is to tolerate a wide
  range of network characteristics - e.g., bandwidth, delay, packet
  loss, packet reordering, and maximum packet size.  Another
  objective is robustness against failure of individual networks,
  routers, and hosts, using whatever bandwidth is still available.
  Finally, the goal is full open system interconnection: an Internet
  router must be able to interoperate robustly and effectively with
  any other router or Internet host, across diverse Internet paths.
  Sometimes implementors have designed for less ambitious goals.
  For example, the LAN environment is typically much more benign
  than the Internet as a whole; LANs have low packet loss and delay
  and do not reorder packets.  Some vendors have fielded
  implementations that are adequate for a simple LAN environment,
  but work badly for general interoperation.  The vendor justifies
  such a product as being economical within the restricted LAN
  market.  However, isolated LANs seldom stay isolated for long;
  they are soon connected to each other, to organization-wide
  internets, and eventually to the global Internet system.  In the
  end, neither the customer nor the vendor is served by incomplete
  or substandard routers.
  The requirements spelled out in this document are designed for a
  full-function router.  It is intended that fully compliant routers
  will be usable in almost any part of the Internet.

LINK LAYER

Although [INTRO:1] covers Link Layer standards (IP over foo, ARP, etc.), this document anticipates that Link-Layer material will be covered in a separate Link Layer Requirements document. A Link-Layer requirements document would be applicable to both hosts and routers. Thus, this document will not obsolete the parts of [INTRO:1] that deal with link-layer issues.

INTRODUCTION

Routers have essentially the same Link Layer protocol requirements as other sorts of Internet systems. These requirements are given in chapter 3 of Requirements for Internet Gateways [INTRO:1]. A router MUST comply with its requirements and SHOULD comply with its recommendations. Since some of the material in that document has become somewhat dated, some additional requirements and explanations are included below.

DISCUSSION:

  It is expected that the Internet community will produce a
  Requirements for Internet Link Layer standard which will supersede
  both this chapter and chapter 3 of [INTRO:1].

LINK/INTERNET LAYER INTERFACE

Although this document does not attempt to specify the interface between the Link Layer and the upper layers, it is worth noting here that other parts of this document, particularly chapter 5, require various sorts of information to be passed across this layer boundary.

This section uses the following definitions:

o Source physical address

  The source physical address is the Link Layer address of the host
  or router from which the packet was received.

o Destination physical address

  The destination physical address is the Link Layer address to
  which the packet was sent.

The information that must pass from the Link Layer to the Internetwork Layer for each received packet is:

(1) The IP packet [5.2.2],

(2) The length of the data portion (i.e., not including the Link-

    Layer framing) of the Link Layer frame [5.2.2],

(3) The identity of the physical interface from which the IP packet

    was received [5.2.3], and

(4) The classification of the packet's destination physical address

    as a Link Layer unicast, broadcast, or multicast [4.3.2],
    [5.3.4].

In addition, the Link Layer also should provide:

(5) The source physical address.

The information that must pass from the Internetwork Layer to the Link Layer for each transmitted packet is:

(1) The IP packet [5.2.1]

(2) The length of the IP packet [5.2.1]

(3) The destination physical interface [5.2.1]

(4) The next hop IP address [5.2.1]

In addition, the Internetwork Layer also should provide:

(5) The Link Layer priority value [5.3.3.2]

The Link Layer must also notify the Internetwork Layer if the packet to be transmitted causes a Link Layer precedence-related error [5.3.3.3].

SPECIFIC ISSUES

Trailer Encapsulation

  Routers which can connect to 10Mb Ethernets MAY be able to receive
  and forward Ethernet packets encapsulated using the trailer
  encapsulation described in [LINK:1].  However, a router SHOULD NOT
  originate trailer encapsulated packets.  A router MUST NOT
  originate trailer encapsulated packets without first verifying,
  using the mechanism described in section 2.3.1 of [INTRO:2], that
  the immediate destination of the packet is willing and able to
  accept trailer-encapsulated packets.  A router SHOULD NOT agree
  (using these same mechanisms) to accept trailer-encapsulated
  packets.

Address Resolution Protocol - ARP

  Routers which implement ARP MUST be compliant and SHOULD be
  unconditionally compliant with the requirements in section 2.3.2
  of [INTRO:2].
  The link layer MUST NOT report a Destination Unreachable error to
  IP solely because there is no ARP cache entry for a destination.
  A router MUST not believe any ARP reply which claims that the Link
  Layer address of another host or router is a broadcast or
  multicast address.

Ethernet and 802.3 Coexistence

  Routers which can connect to 10Mb Ethernets MUST be compliant and
  SHOULD be unconditionally compliant with the requirements of
  Section [2.3.3] of [INTRO:2].

Maximum Transmission Unit - MTU

  The MTU of each logical interface MUST be configurable.
  Many Link Layer protocols define a maximum frame size that may be
  sent.  In such cases, a router MUST NOT allow an MTU to be set
  which would allow sending of frames larger than those allowed by
  the Link Layer protocol.  However, a router SHOULD be willing to
  receive a packet as large as the maximum frame size even if that
  is larger than the MTU.
  DISCUSSION:
     Note that this is a stricter requirement than imposed on hosts
     by [INTRO:2], which requires that the MTU of each physical
     interface be configurable.
     If a network is using an MTU smaller than the maximum frame
     size for the Link Layer, a router may receive packets larger
     than the MTU from hosts which are in the process of
     initializing themselves, or which have been misconfigured.
     In general, the Robustness Principle indicates that these
     packets should be successfully received, if at all possible.

Point-to-Point Protocol - PPP

  Contrary to [INTRO:1], the Internet does have a standard serial
  line protocol: the Point-to-Point Protocol (PPP), defined in
  [LINK:2], [LINK:3], [LINK:4], and [LINK:5].
  A serial line interface is any interface which is designed to send
  data over a telephone, leased, dedicated or direct line (either 2
  or 4 wire) using a standardized modem or bit serial interface
  (such as RS-232, RS-449 or V.35), using either synchronous or
  asynchronous clocking.
  A general purpose serial interface is a serial line interface
  which is not solely for use as an access line to a network for
  which an alternative IP link layer specification exists (such as
  X.25 or Frame Relay).
  Routers which contain such general purpose serial interfaces MUST
  implement PPP.
  PPP MUST be supported on all general purpose serial interfaces on
  a router.  The router MAY allow the line to be configured to use
  serial line protocols other than PPP, all general purpose serial
  interfaces MUST default to using PPP.
Introduction
     This section provides guidelines to router implementors so that
     they can ensure interoperability with other routers using PPP
     over either synchronous or asynchronous links.
     It is critical that an implementor understand the semantics of
     the option negotiation mechanism.  Options are a means for a
     local device to indicate to a remote peer what the local device
     will *accept* from the remote peer, not what it wishes to send.
     It is up to the remote peer to decide what is most convenient
     to send within the confines of the set of options that the
     local device has stated that it can accept.  Therefore it is
     perfectly acceptable and normal for a remote peer to ACK all
     the options indicated in an LCP Configuration Request (CR) even
     if the remote peer does not support any of those options.
     Again, the options are simply a mechanism for either device to
     indicate to its peer what it will accept, not necessarily what
     it will send.
Link Control Protocol (LCP) Options
     The PPP Link Control Protocol (LCP) offers a number of options
     that may be negotiated.  These options include (among others)
     address and control field compression, protocol field
     compression, asynchronous character map, Maximum Receive Unit
     (MRU), Link Quality Monitoring (LQM), magic number (for
     loopback detection), Password Authentication Protocol (PAP),
     Challenge Handshake Authentication Protocol (CHAP), and the
     32-bit Frame Check Sequence (FCS).
     A router MAY do address/control field compression on either
     synchronous or asynchronous links.  A router MAY do protocol
     field compression on either synchronous or asynchronous links.
     A router MAY indicate that it can accept these compressions,
     but MUST be able to accept uncompressed PPP header information
     even if it has indicated a willingness to receive compressed
     PPP headers.
     DISCUSSION:
        These options control the appearance of the PPP header.
        Normally the PPP header consists of the address field (one
        byte containing the value 0xff), the control field (one byte
        containing the value 0x03), and the two-byte protocol field
        that identifies the contents of the data area of the frame.
        If a system negotiates address and control field compression
        it indicates to its peer that it will accept PPP frames that
        have or do not have these fields at the front of the header.
        It does not indicate that it will be sending frames with
        these fields removed.  The protocol field may also be
        compressed from two to one byte in most cases.
     IMPLEMENTATION:
        Some hardware does not deal well with variable length header
        information.  In those cases it makes most sense for the
        remote peer to send the full PPP header.  Implementations
        may ensure this by not sending the address/control field and
        protocol field compression options to the remote peer.  Even
        if the remote peer has indicated an ability to receive
        compressed headers there is no requirement for the local
        router to send compressed headers.
     A router MUST negotiate the Async Control Character Map (ACCM)
     for asynchronous PPP links, but SHOULD NOT negotiate the ACCM
     for synchronous links.  If a router receives an attempt to
     negotiate the ACCM over a synchronous link, it MUST ACKnowledge
     the option and then ignore it.
     DISCUSSION:
        There are implementations that offer both sync and async
        modes of operation and may use the same code to implement
        the option negotiation.  In this situation it is possible
        that one end or the other may send the ACCM option on a
        synchronous link.
     A router SHOULD properly negotiate the maximum receive unit
     (MRU).  Even if a system negotiates an MRU smaller than 1,500
     bytes, it MUST be able to receive a 1,500 byte frame.
     A router SHOULD negotiate and enable the link quality
     monitoring (LQM) option.
     DISCUSSION:
        This memo does not specify a policy for deciding whether the
        link's quality is adequate.  However, it is important (see
        Section [3.3.6]) that a router disable failed links.
     A router SHOULD implement and negotiate the magic number option
     for loopback detection.
     A router MAY support the authentication options (PAP - password
     authentication protocol, and/or CHAP - challenge handshake
     authentication protocol).
     A router MUST support 16-bit CRC frame check sequence (FCS) and
     MAY support the 32-bit CRC.
IP Control Protocol (ICP) Options
     A router MAY offer to perform IP address negotiation.  A router
     MUST accept a refusal (REJect) to perform IP address
     negotiation from the peer.
     A router SHOULD NOT perform Van Jacobson header compression of
     TCP/IP packets if the link speed is in excess of 64 Kbps.
     Below that speed the router MAY perform Van Jacobson (VJ)
     header compression.  At link speeds of 19,200 bps or less the
     router SHOULD perform VJ header compression.

Interface Testing

  A router MUST have a mechanism to allow routing software to
  determine whether a physical interface is available to send
  packets or not.  A router SHOULD have a mechanism to allow routing
  software to judge the quality of a physical interface.  A router
  MUST have a mechanism for informing the routing software when a
  physical interface becomes available or unavailable to send
  packets because of administrative action.  A router MUST have a
  mechanism for informing the routing software when it detects a
  Link level interface has become available or unavailable, for any
  reason.
  DISCUSSION:
     It is crucial that routers have workable mechanisms for
     determining that their network connections are functioning
     properly, since failure to do so (or failure to take the proper
     actions when a problem is detected) can lead to black holes.
     The mechanisms available for detecting problems with network
     connections vary considerably, depending on the Link Layer
     protocols in use and also in some cases on the interface
     hardware chosen by the router manufacturer.  The intent is to
     maximize the capability to detect failures within the Link-
     Layer constraints.

INTERNET LAYER - PROTOCOLS

INTRODUCTION

This chapter and chapter 5 discuss the protocols used at the Internet Layer: IP, ICMP, and IGMP. Since forwarding is obviously a crucial topic in a document discussing routers, chapter 5 limits itself to the aspects of the protocols which directly relate to forwarding. The current chapter contains the remainder of the discussion of the Internet Layer protocols.

INTERNET PROTOCOL - IP

INTRODUCTION

  Routers MUST implement the IP protocol, as defined by
  [INTERNET:1].  They MUST also implement its mandatory extensions:
  subnets (defined in [INTERNET:2]), and IP broadcast (defined in
  [INTERNET:3]).
  A router  MUST be compliant, and SHOULD be unconditionally
  compliant, with the requirements of sections 3.2.1 and 3.3 of
  [INTRO:2], except that:
  o  Section 3.2.1.1 may be ignored, since it duplicates
     requirements found in this memo.
  o  Section 3.2.1.2 may be ignored, since it duplicates
     requirements found in this memo.
  o  Section 3.2.1.3 should be ignored, since it is superseded by
     Section [4.2.2.11] of this memo.
  o  Section 3.2.1.4 may be ignored, since it duplicates
     requirements found in this memo.
  o  Section 3.2.1.6 should be ignored, since it is superseded by
     Section [4.2.2.4] of this memo.
  o  Section 3.2.1.8 should be ignored, since it is superseded by
     Section [4.2.2.1] of this memo.
  In the following, the action specified in certain cases is to
  silently discard a received datagram.  This means that the
  datagram will be discarded without further processing and that the
  router will not send any ICMP error message (see Section [4.3]) as
  a result.  However, for diagnosis of problems a router SHOULD
  provide the capability of logging the error (see Section [1.3.3]),
  including the contents of the silently-discarded datagram, and
  SHOULD record the event in a statistics counter.

PROTOCOL WALK-THROUGH

  RFC 791 is [INTERNET:1], the specification for the Internet
  Protocol.
Options: RFC-791 Section 3.2
     In datagrams received by the router itself, the IP layer MUST
     interpret those IP options that it understands and preserve the
     rest unchanged for use by higher layer protocols.
     Higher layer protocols may require the ability to set IP
     options in datagrams they send or examine IP options in
     datagrams they receive.  Later sections of this document
     discuss specific IP option support required by higher layer
     protocols.
     DISCUSSION:
        Neither this memo nor [INTRO:2] define the order in which a
        receiver must process multiple options in the same IP
        header.  Hosts and routers originating datagrams containing
        multiple options must be aware that this introduces an
        ambiguity in the meaning of certain options when combined
        with a source-route option.
     Here are the requirements for specific IP options:
     (a)  Security Option
          Some environments require the Security option in every
          packet originated or received.  Routers SHOULD IMPLEMENT
          the revised security option described in [INTERNET:5].
          DISCUSSION:
             Note that the security options described in
             [INTERNET:1] and RFC 1038 ([INTERNET:16]) are obsolete.
     (b)  Stream Identifier Option
          This option is obsolete; routers SHOULD NOT place this
          option in a datagram that the router originates.  This
          option MUST be ignored in datagrams received by the
          router.
     (c)  Source Route Options
          A router MUST be able to act as the final destination of a
          source route.  If a router receives a packet containing a
          completed source route (i.e., the pointer points beyond
          the last field and the destination address in the IP
          header addresses the router), the packet has reached its
          final destination; the option as received (the recorded
          route) MUST be passed up to the transport layer (or to
          ICMP message processing).
          In order to respond correctly to source-routed datagrams
          it receives, a router MUST provide a means whereby
          transport protocols and applications can reverse the
          source route in a received datagram and insert the
          reversed source route into datagrams they originate (see
          Section 4 of [INTRO:2] for details).
          Some applications in the router MAY require that the user
          be able to enter a source route.
          A router MUST NOT originate a datagram containing multiple
          source route options.  What a router should do if asked to
          forward a packet containing multiple source route options
          is described in Section [5.2.4.1].
          When a source route option is created, it MUST be
          correctly formed even if it is being created by reversing
          a recorded route that erroneously includes the source host
          (see case (B) in the discussion below).
          DISCUSSION:
             Suppose a source routed datagram is to be routed from
             source S to destination D via routers G1, G2, ... Gn.
             Source S constructs a datagram with G1's IP address as
             its destination address, and a source route option to
             get the datagram the rest of the way to its
             destination.  However, there is an ambiguity in the
             specification over whether the source route option in a
             datagram sent out by S should be (A) or (B):
             (A):  {>>G2, G3, ... Gn, D}     <--- CORRECT
             (B):  {S, >>G2, G3, ... Gn, D}  <---- WRONG
             (where >> represents the pointer).  If (A) is sent, the
             datagram received at D will contain the option: {G1,
             G2, ... Gn >>}, with S and D as the IP source and
             destination addresses.  If (B) were sent, the datagram
             received at D would again contain S and D as the same
             IP source and destination addresses, but the option
             would be: {S, G1, ...Gn >>}; i.e., the originating host
             would be the first hop in the route.
     (d)  Record Route Option
          Routers MAY support the Record Route option in datagrams
          originated by the router.
     (e)  Timestamp Option
          Routers MAY support the timestamp option in datagrams
          originated by the router.  The following rules apply:
          o  When originating a datagram containing a Timestamp
             Option, a router MUST record a timestamp in the option
             if
             - Its Internet address fields are not pre-specified or
             - Its first pre-specified address is the IP address of
                the logical interface over which the datagram is
                being sent (or the router's router-id if the
                datagram is being sent over an unnumbered
                interface).
          o  If the router itself receives a datagram containing a
             Timestamp Option, the router MUST insert the current
             timestamp into the Timestamp Option (if there is space
             in the option to do so) before passing the option to
             the transport layer or to ICMP for processing.
          o  A timestamp value MUST follow the rules given in
             Section [3.2.2.8] of [INTRO:2].
          IMPLEMENTATION:
             To maximize the utility of the timestamps contained in
             the timestamp option, it is suggested that the
             timestamp inserted be, as nearly as practical, the time
             at which the packet arrived at the router.  For
             datagrams originated by the router, the timestamp
             inserted should be, as nearly as practical, the time at
             which the datagram was passed to the Link Layer for
             transmission.
Addresses in Options: RFC-791 Section 3.1
     When a router inserts its address into a Record Route, Strict
     Source and Record Route, Loose Source and Record Route, or
     Timestamp, it MUST use the IP address of the logical interface
     on which the packet is being sent.  Where this rule cannot be
     obeyed because the output interface has no IP address (i.e., is
     an unnumbered interface), the router MUST instead insert its
     router-id.  The router's router-id is one of the router's IP
     addresses.  Which of the router's addresses is used as the
     router-id MUST NOT change (even across reboots) unless changed
     by the network manager or unless the configuration of the
     router is changed such that the IP address used as the router-
     id ceases to be one of the router's IP addresses.  Routers with
     multiple unnumbered interfaces MAY have multiple router-id's.
     Each unnumbered interface MUST be associated with a particular
     router-id.  This association MUST NOT change (even across
     reboots) without reconfiguration of the router.
     DISCUSSION:
        This specification does not allow for routers which do not
        have at least one IP address.  We do not view this as a
        serious limitation, since a router needs an IP address to
        meet the manageability requirements of Chapter [8] even if
        the router is connected only to point-to-point links.
     IMPLEMENTATION:
        One possible method of choosing the router-id that fulfills
        this requirement is to use the numerically smallest (or
        greatest) IP address (treating the address as a 32-bit
        integer) that is assigned to the router.
Unused IP Header Bits: RFC-791 Section 3.1
     The IP header contains two reserved bits: one in the Type of
     Service byte and the other in the Flags field.  A router MUST
     NOT set either of these bits to one in datagrams originated by
     the router.  A router MUST NOT drop (refuse to receive or
     forward) a packet merely because one or more of these reserved
     bits has a non-zero value.
     DISCUSSION:
        Future revisions to the IP protocol may make use of these
        unused bits.  These rules are intended to ensure that these
        revisions can be deployed without having to simultaneously
        upgrade all routers in the Internet.
Type of Service: RFC-791 Section 3.1
     The Type-of-Service byte in the IP header is divided into three
     sections:  the Precedence field (high-order 3 bits), a field
     that is customarily called Type of Service or TOS (next 4
     bits), and a reserved bit (the low order bit).
     Rules governing the reserved bit were described in Section
     [4.2.2.3].
     A more extensive discussion of the TOS field and its use can be
     found in [ROUTE:11].
     The description of the IP Precedence field is superseded by
     Section [5.3.3].  RFC-795, Service Mappings, is obsolete and
     SHOULD NOT be implemented.
Header Checksum: RFC-791 Section 3.1
     As stated in Section [5.2.2], a router MUST verify the IP
     checksum of any packet which is received.  The router MUST NOT
     provide a means to disable this checksum verification.
     IMPLEMENTATION:
        A more extensive description of the IP checksum, including
        extensive implementation hints, can be found in [INTERNET:6]
        and [INTERNET:7].
Unrecognized Header Options: RFC-791 Section 3.1
     A router MUST ignore IP options which it does not recognize.  A
     corollary of this requirement is that a router MUST implement
     the End of Option List option and the No Operation option,
     since neither contains an explicit length.
     DISCUSSION:
        All future IP options will include an explicit length.
Fragmentation: RFC-791 Section 3.2
     Fragmentation, as described in [INTERNET:1], MUST be supported
     by a router.
     When a router fragments an IP datagram, it SHOULD minimize the
     number of fragments.  When a router fragments an IP datagram,
     it MUST send the fragments in order.  A fragmentation method
     which may generate one IP fragment which is significantly
     smaller than the other MAY cause the first IP fragment to be
     the smaller one.
     DISCUSSION:
        There are several fragmentation techniques in common use in
        the Internet.  One involves splitting the IP datagram into
        IP fragments with the first being MTU sized, and the others
        being approximately the same size, smaller than the MTU.
        The reason for this is twofold.  The first IP fragment in
        the sequence will be the effective MTU of the current path
        between the hosts, and the following IP fragments are sized
        to hopefully minimize the further fragmentation of the IP
        datagram.  Another technique is to split the IP datagram
        into MTU sized IP fragments, with the last fragment being
        the only one smaller, as per page 26 of [INTERNET:1].
        A common trick used by some implementations of TCP/IP is to
        fragment an IP datagram into IP fragments that are no larger
        than 576 bytes when the IP datagram is to travel through a
        router.  In general, this allows the resulting IP fragments
        to pass the rest of the path without further fragmentation.
        This would, though, create more of a load on the destination
        host, since it would have a larger number of IP fragments to
        reassemble into one IP datagram.  It would also not be
        efficient on networks where the MTU only changes once, and
        stays much larger than 576 bytes (such as an 802.5 network
        with a MTU of 2048 or an Ethernet network with an MTU of
        1536).
        One other fragmentation technique discussed was splitting
        the IP datagram into approximately equal sized IP fragments,
        with the size being smaller than the next hop network's MTU.
        This is intended to minimize the number of fragments that
        would result from additional fragmentation further down the
        path.
        In most cases, routers should try and create situations that
        will generate the lowest number of IP fragments possible.
        Work with slow machines leads us to believe that if it is
        necessary to send small packets in a fragmentation scheme,
        sending the small IP fragment first maximizes the chance of
        a host with a slow interface of receiving all the fragments.
Reassembly: RFC-791 Section 3.2
     As specified in Section 3.3.2 of [INTRO:2], a router MUST
     support reassembly of datagrams which it delivers to itself.
Time to Live: RFC-791 Section 3.2
     Time to Live (TTL) handling for packets originated or received
     by the router is governed by [INTRO:2].  Note in particular
     that a router MUST NOT check the TTL of a packet except when
     forwarding it.

4.2.2.10 Multi-subnet Broadcasts: RFC-922

     All-subnets broadcasts (called multi-subnet broadcasts in
     [INTERNET:3]) have been deprecated.  See Section [5.3.5.3].

4.2.2.11 Addressing: RFC-791 Section 3.2

     There are now five classes of IP addresses: Class A through
     Class E.  Class D addresses are used for IP multicasting
     [INTERNET:4], while Class E addresses are reserved for
     experimental use.
     A multicast (Class D) address is a 28-bit logical address that
     stands for a group of hosts, and may be either permanent or
     transient.  Permanent multicast addresses are allocated by the
     Internet Assigned Number Authority [INTRO:7], while transient
     addresses may be allocated dynamically to transient groups.
     Group membership is determined dynamically using IGMP
     [INTERNET:4].
     We now summarize the important special cases for Unicast (that
     is class A, B, and C) IP addresses, using the following
     notation for an IP address:
        { <Network-number>, <Host-number> }
     or
        { <Network-number>, <Subnet-number>, <Host-number> }
     and the notation -1 for a field that contains all 1 bits and
     the notation 0 for a field that contains all 0 bits.  This
     notation is not intended to imply that the 1-bits in a subnet
     mask need be contiguous.
     (a)  { 0, 0 }
          This host on this network.  It MUST NOT be used as a
          source address by routers, except the router MAY use this
          as a source address as part of an initialization procedure
          (e.g., if the router is using BOOTP to load its
          configuration information).
          Incoming datagrams with a source address of { 0, 0 } which
          are received for local delivery (see Section [5.2.3]),
          MUST be accepted if the router implements the associated
          protocol and that protocol clearly defines appropriate
          action to be taken.  Otherwise, a router MUST silently
          discard any locally-delivered datagram whose source
          address is { 0, 0 }.
          DISCUSSION:
             Some protocols define specific actions to take in
             response to a received datagram whose source address is
             { 0, 0 }.  Two examples are BOOTP and ICMP Mask
             Request.  The proper operation of these protocols often
             depends on the ability to receive datagrams whose
             source address is { 0, 0 }.  For most protocols,
             however, it is best to ignore datagrams having a source
             address of { 0, 0 } since they were probably generated
             by a misconfigured host or router.  Thus, if a router
             knows how to deal with a given datagram having a { 0, 0
             } source address, the router MUST accept it.
             Otherwise, the router MUST discard it.
          See also Section [4.2.3.1] for a non-standard use of { 0,
          0 }.
     (b)  { 0, <Host-number> }
          Specified host on this network.  It MUST NOT be sent by
          routers except that the router MAY uses this as a source
          address as part of an initialization procedure by which
          the it learns its own IP address.
     (c)  { -1, -1 }
          Limited broadcast.  It MUST NOT be used as a source
          address.
          A datagram with this destination address will be received
          by every host and router on the connected physical
          network, but will not be forwarded outside that network.
     (d)  { <Network-number>, -1 }
          Network Directed Broadcast - a broadcast directed to the
          specified network.  It MUST NOT be used as a source
          address.  A router MAY originate Network Directed
          Broadcast packets.  A router MUST receive Network Directed
          Broadcast packets; however a router MAY have a
          configuration option to prevent reception of these
          packets.  Such an option MUST default to allowing
          reception.
     (e)  { <Network-number>, <Subnet-number>, -1 }
          Subnetwork Directed Broadcast - a broadcast sent to the
          specified subnet.  It MUST NOT be used as a source
          address.  A router MAY originate Network Directed
          Broadcast packets.  A router MUST receive Network Directed
          Broadcast packets; however a router MAY have a
          configuration option to prevent reception of these
          packets.  Such an option MUST default to allowing
          reception.
     (f)  { <Network-number>, -1, -1 }
          All Subnets Directed Broadcast - a broadcast sent to all
          subnets of the specified subnetted network.  It MUST NOT
          be used as a source address.  A router MAY originate
          Network Directed Broadcast packets.  A router MUST receive
          Network Directed Broadcast packets; however a router MAY
          have a configuration option to prevent reception of these
          packets.  Such an option MUST default to allowing
          reception.
     (g)  { 127, <any> }
          Internal host loopback address.  Addresses of this form
          MUST NOT appear outside a host.
     The <Network-number> is administratively assigned so that its
     value will be unique in the entire world.
     IP addresses are not permitted to have the value 0 or -1 for
     any of the <Host-number>, <Network-number>, or <Subnet-number>
     fields (except in the special cases listed above).  This
     implies that each of these fields will be at least two bits
     long.
     For further discussion of broadcast addresses, see Section
     [4.2.3.1].
     Since (as described in Section [4.2.1]) a router must support
     the subnet extensions to IP, there will be a subnet mask of the
     form: { -1, -1, 0 } associated with each of the host's local IP
     addresses; see Sections [4.3.3.9], [5.2.4.2], and [10.2.2].
     When a router originates any datagram, the IP source address
     MUST be one of its own IP addresses (but not a broadcast or
     multicast address).  The only exception is during
     initialization.
     For most purposes, a datagram addressed to a broadcast or
     multicast destination is processed as if it had been addressed
     to one of the router's IP addresses; that is to say:
     o  A router MUST receive and process normally any packets with
        a broadcast destination address.
     o  A router MUST receive and process normally any packets sent
        to a multicast destination address which the router is
        interested in.
     The term specific-destination address means the equivalent
     local IP address of the host.  The specific-destination address
     is defined to be the destination address in the IP header
     unless the header contains a broadcast or multicast address, in
     which case the specific-destination is an IP address assigned
     to the physical interface on which the datagram arrived.
     A router MUST silently discard any received datagram containing
     an IP source address that is invalid by the rules of this
     section.  This validation could be done either by the IP layer
     or by each protocol in the transport layer.
     DISCUSSION:
        A misaddressed datagram might be caused by a Link Layer
        broadcast of a unicast datagram or by another router or host
        that is confused or misconfigured.

SPECIFIC ISSUES

IP Broadcast Addresses
     For historical reasons, there are a number of IP addresses
     (some standard and some not) which are used to indicate that an
     IP packet is an IP broadcast.  A router
     (1)  MUST treat as IP broadcasts packets addressed to
          255.255.255.255, { <Network-number>, -1 }, { <Network-
          number>, <Subnet-number>, -1 }, and { <Network-number>,
          -1, -1 }.
     (2)  SHOULD silently discard on receipt (i.e., don't even
          deliver to applications in the router) any packet
          addressed to 0.0.0.0, { <Network-number>, 0 }, {
          <Network-number>, <Subnet-number>, 0 }, or { <Network-
          number>, 0, 0 }; if these packets are not silently
          discarded, they MUST be treated as IP broadcasts (see
          Section [5.3.5]).  There MAY be a configuration option to
          allow receipt of these packets.  This option SHOULD
          default to discarding them.
     (3)  SHOULD (by default) use the limited broadcast address
          (255.255.255.255) when originating an IP broadcast
          destined for a connected network or subnet (except when
          sending an ICMP Address Mask Reply, as discussed in
          Section [4.3.3.9]).  A router MUST receive limited
          broadcasts.
     (4)  SHOULD NOT originate datagrams addressed to 0.0.0.0, {
          <Network-number>, 0 }, { <Network-number>, <Subnet-
          number>, 0 }, or { <Network-number>, 0, 0 }.  There MAY be
          a configuration option to allow generation of these
          packets (instead of using the relevant 1s format
          broadcast).  This option SHOULD default to not generating
          them.
     DISCUSSION:
        In the second bullet, the router obviously cannot recognize
        addresses of the form { <Network-number>, <Subnet-number>, 0
        } if the router does not know how the particular network is
        subnetted.  In that case, the rules of the second bullet do
        not apply because, from the point of view of the router, the
        packet is not an IP broadcast packet.
IP Multicasting
     An IP router SHOULD satisfy the Host Requirements with respect
     to IP multicasting, as specified in Section 3.3.7 of [INTRO:2].
     An IP router SHOULD support local IP multicasting on all
     connected networks for which a mapping from Class D IP
     addresses to link-layer addresses has been specified (see the
     various IP-over-xxx specifications), and on all connected
     point-to-point links.  Support for local IP multicasting
     includes originating multicast datagrams, joining multicast
     groups and receiving multicast datagrams, and leaving multicast
     groups.  This implies support for all of [INTERNET:4] including
     IGMP (see Section [4.4]).
     DISCUSSION:
        Although [INTERNET:4] is entitled Host Extensions for IP
        Multicasting, it applies to all IP systems, both hosts and
        routers.  In particular, since routers may join multicast
        groups, it is correct for them to perform the host part of
        IGMP, reporting their group memberships to any multicast
        routers that may be present on their attached networks
        (whether or not they themselves are multicast routers).
        Some router protocols may specifically require support for
        IP multicasting (e.g., OSPF [ROUTE:1]), or may recommend it
        (e.g., ICMP Router Discovery [INTERNET:13]).
Path MTU Discovery
     In order to eliminate fragmentation or minimize it, it is
     desirable to know what is the path MTU along the path from the
     source to destination.  The path MTU is the minimum of the MTUs
     of each hop in the path.  [INTERNET:14] describes a technique
     for dynamically discovering the maximum transmission unit (MTU)
     of an arbitrary internet path.  For a path that passes through
     a router that does not support [INTERNET:14], this technique
     might not discover the correct Path MTU, but it will always
     choose a Path MTU as accurate as, and in many cases more
     accurate than, the Path MTU that would be chosen by older
     techniques or the current practice.
     When a router is originating an IP datagram, it SHOULD use the
     scheme described in [INTERNET:14] to limit the datagram's size.
     If the router's route to the datagram's destination was learned
     from a routing protocol that provides Path MTU information, the
     scheme described in [INTERNET:14] is still used, but the Path
     MTU information from the routing protocol SHOULD be used as the
     initial guess as to the Path MTU and also as an upper bound on
     the Path MTU.
Subnetting
     Under certain circumstances, it may be desirable to support
     subnets of a particular network being interconnected only via a
     path which is not part of the subnetted network.  This is known
     as discontiguous subnetwork support.
     Routers MUST support discontiguous subnetworks.
     IMPLEMENTATION:
        In general, a router should not make assumptions about what
        are subnets and what are not, but simply ignore the concept
        of Class in networks, and treat each route as a { network,
        mask }-tuple.
     DISCUSSION:
        The Internet has been growing at a tremendous rate of late.
        This has been placing severe strains on the IP addressing
        technology.  A major factor in this strain is the strict IP
        Address class boundaries.  These make it difficult to
        efficiently size network numbers to their networks and
        aggregate several network numbers into a single route
        advertisement.  By eliminating the strict class boundaries
        of the IP address and treating each route as a {network
        number, mask}-tuple these strains may be greatly reduced.
        The technology for currently doing this is Classless
        Interdomain Routing (CIDR) [INTERNET:15].
     Furthermore, for similar reasons, a subnetted network need not
     have a consistent subnet mask through all parts of the network.
     For example, one subnet may use an 8 bit subnet mask, another
     10 bit, and another 6 bit.  This is known as variable subnet-
     masks.
     Routers MUST support variable subnet-masks.

INTERNET CONTROL MESSAGE PROTOCOL - ICMP

INTRODUCTION

  ICMP is an auxiliary protocol, which provides routing, diagnostic
  and and error functionality for IP. It is described in
  [INTERNET:8].  A router MUST support ICMP.
  ICMP messages are grouped in two classes which are discussed in
  the following sections:
  ICMP error messages:
  Destination Unreachable     Section 4.3.3.1
  Redirect                    Section 4.3.3.2
  Source Quench               Section 4.3.3.3
  Time Exceeded               Section 4.3.3.4
  Parameter Problem           Section 4.3.3.5
  ICMP query messages:
  Echo                        Section 4.3.3.6
  Information                 Section 4.3.3.7
  Timestamp                   Section 4.3.3.8
  Address Mask                Section 4.3.3.9
  Router Discovery            Section 4.3.3.10
  General ICMP requirements and discussion are in the next section.

GENERAL ISSUES

Unknown Message Types
     If an ICMP message of unknown type is received, it MUST be
     passed to the ICMP user interface (if the router has one) or
     silently discarded (if the router doesn't have one).
ICMP Message TTL
     When originating an ICMP message, the router MUST initialize
     the TTL.  The TTL for ICMP responses must not be taken from the
     packet which triggered the response.
Original Message Header
     Every ICMP error message includes the Internet header and at
     least the first 8 data bytes of the datagram that triggered the
     error.  More than 8 bytes MAY be sent, but the resulting ICMP
     datagram SHOULD have a length of less than or equal to 576
     bytes.  The returned IP header (and user data) MUST be
     identical to that which was received, except that the router is
     not required to undo any modifications to the IP header that
     are normally performed in forwarding that were performed before
     the error was detected (e.g., decrementing the TTL, updating
     options).  Note that the requirements of Section [4.3.3.5]
     supersede this requirement in some cases (i.e., for a Parameter
     Problem message, if the problem  is in a modified field, the
     router must undo the modification).  See Section [4.3.3.5])
ICMP Message Source Address
     Except where this document specifies otherwise, the IP source
     address in an ICMP message originated by the router MUST be one
     of the IP addresses associated with the physical interface over
     which the ICMP message is transmitted.  If the interface has no
     IP addresses associated with it, the router's router-id (see
     Section [5.2.5]) is used instead.
TOS and Precedence
     ICMP error messages SHOULD have their TOS bits set to the same
     value as the TOS bits in the packet which provoked the sending
     of the ICMP error message, unless setting them to that value
     would cause the ICMP error message to be immediately discarded
     because it could not be routed to its destination.  Otherwise,
     ICMP error messages MUST be sent with a normal (i.e. zero) TOS.
     An ICMP reply message SHOULD have its TOS bits set to the same
     value as the TOS bits in the ICMP request that provoked the
     reply.
     EDITOR'S COMMENTS:
        The following paragraph originally read:
           ICMP error messages MUST have their IP Precedence field
           set to the same value as the IP Precedence field in the
           packet which provoked the sending of the ICMP error
           message, except that the precedence value MUST be 6
           (INTERNETWORK CONTROL) or 7 (NETWORK CONTROL), SHOULD be
           7, and MAY be settable for the following types of ICMP
           error messages: Unreachable, Redirect, Time Exceeded, and
           Parameter Problem.
        I believe that the following paragraph is equivalent and
        easier for humans to parse (Source Quench is the only other
        ICMP Error message).  Other interpretations of the original
        are sought.
     ICMP Source Quench error messages MUST have their IP Precedence
     field set to the same value as the IP Precedence field in the
     packet which provoked the sending of the ICMP Source Quench
     message.  All other ICMP error messages (Destination
     Unreachable, Redirect, Time Exceeded, and Parameter Problem)
     MUST have their precedence value set to 6 (INTERNETWORK
     CONTROL) or 7 (NETWORK CONTROL), SHOULD be 7.  The IP
     Precedence value for these error messages MAY be settable.
     An ICMP reply message MUST have its IP Precedence field set to
     the same value as the IP Precedence field in the ICMP request
     that provoked the reply.
Source Route
     If the packet which provokes the sending of an ICMP error
     message contains a source route option, the ICMP error message
     SHOULD also contain a source route option of the same type
     (strict or loose), created by reversing the portion before the
     pointer of the route recorded in the source route option of the
     original packet UNLESS the ICMP error message is an ICMP
     Parameter Problem complaining about a source route option in
     the original packet.
     DISCUSSION:
        In environments which use the U.S. Department of Defense
        security option (defined in [INTERNET:5]), ICMP messages may
        need to include a security option.  Detailed information on
        this topic should be available from the Defense
        Communications Agency.
When Not to Send ICMP Errors
     An ICMP error message MUST NOT be sent as the result of
     receiving:
     o  An ICMP error message, or
     o  A packet which fails the IP header validation tests
        described in Section [5.2.2] (except where that section
        specifically permits the sending of an ICMP error message),
        or
     o  A packet destined to an IP broadcast or IP multicast
        address, or
     o  A packet sent as a Link Layer broadcast or multicast, or
     o  A packet whose source address has a network number of zero
        or is an invalid source address (as defined in Section
        [5.3.7]), or
     o  Any fragment of a datagram other then the first fragment
        (i.e., a packet for which the fragment offset in the IP
        header is nonzero).
     Furthermore, an ICMP error message MUST NOT be sent in any case
     where this memo states that a packet is to be silently
     discarded.
     NOTE:  THESE RESTRICTIONS TAKE PRECEDENCE OVER ANY REQUIREMENT
     ELSEWHERE IN THIS DOCUMENT FOR SENDING ICMP ERROR MESSAGES.
     DISCUSSION:
        These rules aim to prevent the broadcast storms that have
        resulted from routers or hosts returning ICMP error messages
        in response to broadcast packets.  For example, a broadcast
        UDP packet to a non-existent port could trigger a flood of
        ICMP Destination Unreachable datagrams from all devices that
        do not have a client for that destination port.  On a large
        Ethernet, the resulting collisions can render the network
        useless for a second or more.
        Every packet that is broadcast on the connected network
        should have a valid IP broadcast address as its IP
        destination (see Section [5.3.4] and [INTRO:2]).  However,
        some devices violate this rule.  To be certain to detect
        broadcast packets, therefore, routers are required to check
        for a link-layer broadcast as well as an IP-layer address.
     IMPLEMENTATION:
        This requires that the link layer inform the IP layer when a
        link-layer broadcast packet has been received; see Section
        [3.1].
Rate Limiting
     A router which sends ICMP Source Quench messages MUST be able
     to limit the rate at which the messages can be generated.  A
     router SHOULD also be able to limit the rate at which it sends
     other sorts of ICMP error messages (Destination Unreachable,
     Redirect, Time Exceeded, Parameter Problem).  The rate limit
     parameters SHOULD be settable as part of the configuration of
     the router.  How the limits are applied (e.g., per router or
     per interface) is left to the implementor's discretion.
     DISCUSSION:
        Two problems for a router sending ICMP error message are:
        (1)  The consumption of bandwidth on the reverse path, and
        (2)  The use of router resources (e.g., memory, CPU time)
        To help solve these problems a router can limit the
        frequency with which it generates ICMP error messages.  For
        similar reasons, a router may limit the frequency at which
        some other sorts of messages, such as ICMP Echo Replies, are
        generated.
     IMPLEMENTATION:
        Various mechanisms have been used or proposed for limiting
        the rate at which ICMP messages are sent:
        (1)  Count-based - for example, send an ICMP error message
             for every N dropped packets overall or per given source
             host.  This mechanism might be appropriate for ICMP
             Source Quench, but probably not for other types of ICMP
             messages.
        (2)  Timer-based - for example, send an ICMP error message
             to a given source host or overall at most once per T
             milliseconds.
        (3)  Bandwidth-based - for example, limit the rate at which
             ICMP messages are sent over a particular interface to
             some fraction of the attached network's bandwidth.

SPECIFIC ISSUES

Destination Unreachable
     If a route can not forward a packet because it has no routes at
     all to the destination network specified in the packet then the
     router MUST generate a Destination Unreachable, Code 0 (Network
     Unreachable) ICMP message.  If the router does have routes to
     the destination network specified in the packet but the TOS
     specified for the routes is neither the default TOS (0000) nor
     the TOS of the packet that the router is attempting to route,
     then the router MUST generate a Destination Unreachable, Code
     11 (Network Unreachable for TOS) ICMP message.
     If a packet is to be forwarded to a host on a network that is
     directly connected to the router (i.e., the router is the
     last-hop router) and the router has ascertained that there is
     no path to the destination host then the router MUST generate a
     Destination Unreachable, Code 1 (Host Unreachable) ICMP
     message.  If a packet is to be forwarded to a host that is on a
     network that is directly connected to the router and the router
     cannot forward the packet because because no route to the
     destination has a TOS that is either equal to the TOS requested
     in the packet or is the default TOS (0000) then the router MUST
     generate a Destination Unreachable, Code 12 (Host Unreachable
     for TOS) ICMP message.
     DISCUSSION:
        The intent is that a router generates the "generic"
        host/network unreachable if it has no path at all (including
        default routes) to the destination.  If the router has one
        or more paths to the destination, but none of those paths
        have an acceptable TOS, then the router generates the
        "unreachable for TOS" message.
Redirect
     The ICMP Redirect message is generated to inform a host on the
     same subnet that the router used by the host to route certain
     packets should be changed.
     Contrary to section 3.2.2.2 of [INTRO:2], a router MAY ignore
     ICMP Redirects when choosing a path for a packet originated by
     the router if the router is running a routing protocol or if
     forwarding is enabled on the router and on the interface over
     which the packet is being sent.
Source Quench
     A router SHOULD NOT originate ICMP Source Quench messages.  As
     specified in Section [4.3.2], a router which does originate
     Source Quench messages MUST be able to limit the rate at which
     they are generated.
     DISCUSSION:
        Research seems to suggest that Source Quench consumes
        network bandwidth but is an ineffective (and unfair)
        antidote to congestion.  See, for example, [INTERNET:9] and
        [INTERNET:10].  Section [5.3.6] discusses the current
        thinking on how routers ought to deal with overload and
        network congestion.
     A router MAY ignore any ICMP Source Quench messages it
     receives.
     DISCUSSION:
        A router itself may receive a Source Quench as the result of
        originating a packet sent to another router or host.  Such
        datagrams might be, e.g., an EGP update sent to another
        router, or a telnet stream sent to a host.  A mechanism has
        been proposed ([INTERNET:11], [INTERNET:12]) to make the IP
        layer respond directly to Source Quench by controlling the
        rate at which packets are sent, however, this proposal is
        currently experimental and not currently recommended.
Time Exceeded
     When a router is forwarding a packet and the TTL field of the
     packet is reduced to 0, the requirements of section [5.2.3.8]
     apply.
     When the router is reassembling a packet that is destined for
     the router, it MUST fulfill requirements of [INTRO:2], section
     [3.3.2] apply.
     When the router receives (i.e., is destined for the router) a
     Time Exceeded message, it MUST comply with section 3.2.2.4 of
     [INTRO:2].
Parameter Problem
     A router MUST generate a Parameter Problem message for any
     error not specifically covered by another ICMP message.  The IP
     header field or IP option including the byte indicated by the
     pointer field MUST be included unchanged in the IP header
     returned with this ICMP message.  Section [4.3.2] defines an
     exception to this requirement.
     A new variant of the Parameter Problem message was defined in
     [INTRO:2]:
          Code 1 = required option is missing.
     DISCUSSION:
        This variant is currently in use in the military community
        for a missing security option.
Echo Request/Reply
     A router MUST implement an ICMP Echo server function that
     receives Echo Requests and sends corresponding Echo Replies.  A
     router MUST be prepared to receive, reassemble and echo an ICMP
     Echo Request datagram at least as large as the maximum of 576
     and the MTUs of all the connected networks.
     The Echo server function MAY choose not to respond to ICMP echo
     requests addressed to IP broadcast or IP multicast addresses.
     A router SHOULD have a configuration option which, if enabled,
     causes the router to silently ignore all ICMP echo requests; if
     provided, this option MUST default to allowing responses.
     DISCUSSION:
        The neutral provision about responding to broadcast and
        multicast Echo Requests results from the conclusions reached
        in section [3.2.2.6] of [INTRO:2].
     As stated in Section [10.3.3], a router MUST also implement an
     user/application-layer interface for sending an Echo Request
     and receiving an Echo Reply, for diagnostic purposes.  All ICMP
     Echo Reply messages MUST be passed to this interface.
     The IP source address in an ICMP Echo Reply MUST be the same as
     the specific-destination address of the corresponding ICMP Echo
     Request message.
     Data received in an ICMP Echo Request MUST be entirely included
     in the resulting Echo Reply.
     If a Record Route and/or Timestamp option is received in an
     ICMP Echo Request, this option (these options) SHOULD be
     updated to include the current router and included in the IP
     header of the Echo Reply message, without truncation.  Thus,
     the recorded route will be for the entire round trip.
     If a Source Route option is received in an ICMP Echo Request,
     the return route MUST be reversed and used as a Source Route
     option for the Echo Reply message.
Information Request/Reply
     A router SHOULD NOT originate or respond to these messages.
     DISCUSSION:
        The Information Request/Reply pair was intended to support
        self-configuring systems such as diskless workstations, to
        allow them to discover their IP network numbers at boot
        time.  However, these messages are now obsolete.  The RARP
        and BOOTP protocols provide better mechanisms for a host to
        discover its own IP address.
Timestamp and Timestamp Reply
     A router MAY implement Timestamp and Timestamp Reply.  If they
     are implemented then:
     o  The ICMP Timestamp server function MUST return a Timestamp
        Reply to every Timestamp message that is received.  It
        SHOULD be designed for minimum variability in delay.
     o  An ICMP Timestamp Request message to an IP broadcast or IP
        multicast address MAY be silently discarded.
     o  The IP source address in an ICMP Timestamp Reply MUST be the
        same as the specific-destination address of the
        corresponding Timestamp Request message.
     o  If a Source Route option is received in an ICMP Timestamp
        Request, the return route MUST be reversed and used as a
        Source Route option for the Timestamp Reply message.
     o  If a Record Route and/or Timestamp option is received in a
        Timestamp Request, this (these) option(s) SHOULD be updated
        to include the current router and included in the IP header
        of the Timestamp Reply message.
     o  If the router provides an application-layer interface for
        sending Timestamp Request messages then incoming Timestamp
        Reply messages MUST be passed up to the ICMP user interface.
     The preferred form for a timestamp value (the standard value)
     is milliseconds since midnight, Universal Time.  However, it
     may be difficult to provide this value with millisecond
     resolution. For example, many systems use clocks that update
     only at line frequency, 50 or 60 times per second.  Therefore,
     some latitude is allowed in a standard value:
     (a)  A standard value MUST be updated at least 16 times per
          second (i.e., at most the six low-order bits of the value
          may be undefined).
     (b)  The accuracy of a standard value MUST approximate that of
          operator-set CPU clocks, i.e., correct within a few
          minutes.
     IMPLEMENTATION:
        To meet the second condition, a router may need to query
        some time server when the router is booted or restarted. It
        is recommended that the UDP Time Server Protocol be used for
        this purpose. A more advanced implementation would use the
        Network Time Protocol (NTP) to achieve nearly millisecond
        clock synchronization; however, this is not required.
Address Mask Request/Reply
     A router MUST implement support for receiving ICMP Address Mask
     Request messages and responding with ICMP Address Mask Reply
     messages.  These messages are defined in [INTERNET:2].
     A router SHOULD have a configuration option for each logical
     interface specifying whether the router is allowed to answer
     Address Mask Requests for that interface; this option MUST
     default to allowing responses.  A router MUST NOT respond to an
     Address Mask Request before the router knows the correct subnet
     mask.
     A router MUST NOT respond to an Address Mask Request which has
     a source address of 0.0.0.0 and which arrives on a physical
     interface which has associated with it multiple logical
     interfaces and the subnet masks for those interfaces are not
     all the same.
     A router SHOULD examine all ICMP Address Mask Replies which it
     receives to determine whether the information it contains
     matches the router's knowledge of the subnet mask.  If the ICMP
     Address Mask Reply appears to be in error, the router SHOULD
     log the subnet mask and the sender's IP address.  A router MUST
     NOT use the contents of an ICMP Address Mask Reply to determine
     the correct subnet mask.
     Because hosts may not be able to learn the subnet mask if a
     router is down when the host boots up, a router MAY broadcast a
     gratuitous ICMP Address Mask Reply on each of its logical
     interfaces after it has configured its own subnet masks.
     However, this feature can be dangerous in environments which
     use variable length subnet masks.  Therefore, if this feature
     is implemented, gratuitous Address Mask Replies MUST NOT be
     broadcast over any logical interface(s) which either:
     o  Are not configured to send gratuitous Address Mask Replies.
        Each logical interface MUST have a configuration parameter
        controlling this, and that parameter MUST default to not
        sending the gratuitous Address Mask Replies.
     o  Share the same IP network number and physical interface but
        have different subnet masks.
     The { <Network-number>, -1, -1 } form (on subnetted networks)
     or the { <Network-number>, -1 } form (on non-subnetted
     networks) of the IP broadcast address MUST be used for
     broadcast Address Mask Replies.
     DISCUSSION:
        The ability to disable sending Address Mask Replies by
        routers is required at a few sites which intentionally lie
        to their hosts about the subnet mask.  The need for this is
        expected to go away as more and more hosts become compliant
        with the Host Requirements standards.
        The reason for both the second bullet above and the
        requirement about which IP broadcast address to use is to
        prevent problems when multiple IP networks or subnets are in
        use on the same physical network.

4.3.3.10 Router Advertisement and Solicitations

     An IP router MUST support the router part of the ICMP Router
     Discovery Protocol [INTERNET:13] on all connected networks on
     which the router supports either IP multicast or IP broadcast
     addressing.  The implementation MUST include all of the
     configuration variables specified for routers, with the
     specified defaults.
     DISCUSSION:
        Routers are not required to implement the host part of the
        ICMP Router Discovery Protocol, but might find it useful for
        operation while IP forwarding is disabled (i.e., when
        operating as a host).
     DISCUSSION:
        We note that it is quite common for hosts to use RIP as the
        router discovery protocol.  Such hosts listen to RIP traffic
        and use and use information extracted from that traffic to
        discover routers and to make decisions as to which router to
        use as a first-hop router for a given destination.  While
        this behavior is discouraged, it is still common and
        implementors should be aware of it.

INTERNET GROUP MANAGEMENT PROTOCOL - IGMP

IGMP [INTERNET:4] is a protocol used between hosts and multicast routers on a single physical network to establish hosts' membership in particular multicast groups. Multicast routers use this information, in conjunction with a multicast routing protocol, to support IP multicast forwarding across the Internet.

A router SHOULD implement the host part of IGMP.

INTERNET LAYER - FORWARDING

INTRODUCTION

This section describes the process of forwarding packets.

FORWARDING WALK-THROUGH

There is no separate specification of the forwarding function in IP. Instead, forwarding is covered by the protocol specifications for the internet layer protocols ([INTERNET:1], [INTERNET:2], [INTERNET:3], [INTERNET:8], and [ROUTE:11]).

Forwarding Algorithm

  Since none of the primary protocol documents describe the
  forwarding algorithm in any detail, we present it here.  This is
  just a general outline, and omits important details, such as
  handling of congestion, that are dealt with in later sections.
  It is not required that an implementation follow exactly the
  algorithms given in sections [5.2.1.1], [5.2.1.2], and [5.2.1.3].
  Much of the challenge of writing router software is to maximize
  the rate at which the router can forward packets while still
  achieving the same effect of the algorithm.  Details of how to do
  that are beyond the scope of this document, in part because they
  are heavily dependent on the architecture of the router.  Instead,
  we merely point out the order dependencies among the steps:
  (1)  A router MUST verify the IP header, as described in section
       [5.2.2], before performing any actions based on the contents
       of the header.  This allows the router to detect and discard
       bad packets before the expenditure of other resources.
  (2)  Processing of certain IP options requires that the router
       insert its IP address into the option.  As noted in Section
       [5.2.4], the address inserted MUST be the address of the
       logical interface on which the packet is sent or the router's
       router-id if the packet is sent over an unnumbered interface.
       Thus, processing of these options cannot be completed until
       after the output interface is chosen.
  (3)  The router cannot check and decrement the TTL before checking
       whether the packet should be delivered to the router itself,
       for reasons mentioned in Section [4.2.2.9].
  (4)  More generally, when a packet is delivered locally to the
       router, its IP header MUST NOT be modified in any way (except
       that a router may be required to insert a timestamp into any
       Timestamp options in the IP header).  Thus, before the router
       determines whether the packet is to be delivered locally to
       the router, it cannot update the IP header in any way that it
       is not prepared to undo.
General
     This section covers the general forwarding algorithm.  This
     algorithm applies to all forms of packets to be forwarded:
     unicast, multicast, and broadcast.
     (1)  The router receives the IP packet (plus additional
          information about it, as described in Section [3.1]) from
          the Link Layer.
     (2)  The router validates the IP header, as described in
          Section [5.2.2].  Note that IP reassembly is not done,
          except on IP fragments to be queued for local delivery in
          step (4).
     (3)  The router performs most of the processing of any IP
          options.  As described in Section [5.2.4], some IP options
          require additional processing after the routing decision
          has been made.
     (4)  The router examines the destination IP address of the IP
          datagram, as described in Section [5.2.3], to determine
          how it should continue to process the IP datagram.  There
          are three possibilities:
          o  The IP datagram is destined for the router, and should
             be queued for local delivery, doing reassembly if
             needed.
          o  The IP datagram is not destined for the router, and
             should be queued for forwarding.
          o  The IP datagram should be queued for forwarding, but (a
             copy) must also be queued for local delivery.
Unicast
     Since the local delivery case is well-covered by [INTRO:2], the
     following assumes that the IP datagram was queued for
     forwarding.  If the destination is an IP unicast address:
     (5)  The forwarder determines the next hop IP address for the
          packet, usually by looking up the packet's destination in
          the router's routing table.  This procedure is described
          in more detail in Section [5.2.4].  This procedure also
          decides which network interface should be used to send the
          packet.
     (6)  The forwarder verifies that forwarding the packet is
          permitted.  The source and destination addresses should be
          valid, as described in Section [5.3.7] and Section [5.3.4]
          If the router supports administrative constraints on
          forwarding, such as those described in Section [5.3.9],
          those constraints must be satisfied.
     (7)  The forwarder decrements (by at least one) and checks the
          packet's TTL, as described in Section [5.3.1].
     (8)  The forwarder performs any IP option processing that could
          not be completed in step 3.
     (9)  The forwarder performs any necessary IP fragmentation, as
          described in Section [4.2.2.7].  Since this step occurs
          after outbound interface selection (step 5), all fragments
          of the same datagram will be transmitted out the same
          interface.
     (10) The forwarder determines the Link Layer address of the
          packet's next hop.  The mechanisms for doing this are Link
          Layer-dependent (see chapter 3).
     (11) The forwarder encapsulates the IP datagram (or each of the
          fragments thereof) in an appropriate Link Layer frame and
          queues it for output on the interface selected in step 5.
     (12) The forwarder sends an ICMP redirect if necessary, as
          described in Section [4.3.3.2].
Multicast
     If the destination is an IP multicast, the following steps are
     taken.
     Note that the main differences between the forwarding of IP
     unicasts and the forwarding of IP multicasts are
     o  IP multicasts are usually forwarded based on both the
        datagram's source and destination IP addresses,
     o  IP multicast uses an expanding ring search,
     o  IP multicasts are forwarded as Link Level multicasts, and
     o  ICMP errors are never sent in response to IP multicast
        datagrams.
     Note that the forwarding of IP multicasts is still somewhat
     experimental. As a result, the algorithm presented below is not
     mandatory, and is provided as an example only.
     (5a) Based on the IP source and destination addresses found in
          the datagram header, the router determines whether the
          datagram has been received on the proper interface for
          forwarding. If not, the datagram is dropped silently.  The
          method for determining the proper receiving interface
          depends on the multicast routing algorithm(s) in use. In
          one of the simplest algorithms, reverse path forwarding
          (RPF), the proper interface is the one that would be used
          to forward unicasts back to the datagram source.
     (6a) Based on the IP source and destination addresses found in
          the datagram header, the router determines the datagram's
          outgoing interfaces. In order to implement IP multicast's
          expanding ring search (see [INTERNET:4]) a minimum TTL
          value is specified for each outgoing interface. A copy of
          the multicast datagram is forwarded out each outgoing
          interface whose minimum TTL value is less than or equal to
          the TTL value in the datagram header, by separately
          applying the remaining steps on each such interface.
     (7a) The router decrements the packet's TTL by one.
     (8a) The forwarder performs any IP option processing that could
          not be completed in step (3).
     (9a) The forwarder performs any necessary IP fragmentation, as
          described in Section [4.2.2.7].
     (10a) The forwarder determines the Link Layer address to use in
          the Link Level encapsulation. The mechanisms for doing
          this are Link Layer-dependent. On LANs a Link Level
          multicast or broadcast is selected, as an algorithmic
          translation of the datagrams' class D destination address.
          See the various IP-over-xxx specifications for more
          details.
     (11a) The forwarder encapsulates the packet (or each of the
          fragments thereof) in an appropriate Link Layer frame and
          queues it for output on the appropriate interface.

IP Header Validation

  Before a router can process any IP packet, it MUST perform a the
  following basic validity checks on the packet's IP header to
  ensure that the header is meaningful.  If the packet fails any of
  the following tests, it MUST be silently discarded, and the error
  SHOULD be logged.
  (1)  The packet length reported by the Link Layer must be large
       enough to hold the minimum length legal IP datagram (20
       bytes).
  (2)  The IP checksum must be correct.
  (3)  The IP version number must be 4.  If the version number is
       not 4 then the packet may well be another version of IP, such
       as ST-II.
  (4)  The IP header length field must be at least 5.
  (5)  The IP total length field must be at least 4 * IP header
       length field.
  A router MUST NOT have a configuration option which allows
  disabling any of these tests.
  If the packet passes the second and third tests, the IP header
  length field is at least 4, and both the IP total length field and
  the packet length reported by the Link Layer are at least 16 then,
  despite the above rule, the router MAY respond with an ICMP
  Parameter Problem message, whose pointer points at the IP header
  length field (if it failed the fourth test) or the IP total length
  field (if it failed the fifth test).  However, it still MUST
  discard the packet and still SHOULD log the error.
  These rules (and this entire document) apply only to version 4 of
  the Internet Protocol.  These rules should not be construed as
  prohibiting routers from supporting other versions of IP.
  Furthermore, if a router can truly classify a packet as being some
  other version of IP then it ought not treat that packet as an
  error packet within the context of this memo.
  IMPLEMENTATION:
     It is desirable for purposes of error reporting, though not
     always entirely possible, to determine why a header was
     invalid.  There are four possible reasons:
     o  The Link Layer truncated the IP header
     o  The datagram is using a version of IP other than the
        standard one (version 4).
     o  The IP header has been corrupted in transit.
     o  The sender generated an illegal IP header.
     It is probably desirable to perform the checks in the order
     listed, since we believe that this ordering is most likely to
     correctly categorize the cause of the error.  For purposes of
     error reporting, it may also be desirable to check if a packet
     which fails these tests has an IP version number equal to 6.
     If it does, the packet is probably an ST-II datagram and should
     be treated as such.  ST-II is described in [FORWARD:1].
  Additionally, the router SHOULD verify that the packet length
  reported by the Link Layer is at least as large as the IP total
  length recorded in the packet's IP header.  If it appears that the
  packet has been truncated, the packet MUST be discarded, the error
  SHOULD be logged, and the router SHOULD respond with an ICMP
  Parameter Problem message whose pointer points at the IP total
  length field.
  DISCUSSION:
     Because any higher layer protocol which concerns itself with
     data corruption will detect truncation of the packet data when
     it reaches its final destination, it is not absolutely
     necessary for routers to perform the check suggested above in
     order to maintain protocol correctness.  However, by making
     this check a router can simplify considerably the task of
     determining which hop in the path is truncating the packets.
     It will also reduce the expenditure of resources down-stream
     from the router in that down-stream systems will not need to
     deal with the packet.
  Finally, if the destination address in the IP header is not one of
  the addresses of the router, the router SHOULD verify that the
  packet does not contain a Strict Source and Record Route option.
  If a packet fails this test, the router SHOULD log the error and
  SHOULD respond with an ICMP Parameter Problem error with the
  pointer pointing at the offending packet's IP destination address.
  DISCUSSION:
     Some people might suggest that the router should respond with a
     Bad Source Route message instead of a Parameter Problem
     message.  However, when a packet fails this test, it usually
     indicates a protocol error by the previous hop router, whereas
     Bad Source Route would suggest that the source host had
     requested a nonexistent or broken path through the network.

Local Delivery Decision

  When a router receives an IP packet, it must decide whether the
  packet is addressed to the router (and should be delivered
  locally) or the packet is addressed to another system (and should
  be handled by the forwarder).  There is also a hybrid case, where
  certain IP broadcasts and IP multicasts are both delivered locally
  and forwarded.  A router MUST determine which of the these three
  cases applies using the following rules:
  o  An unexpired source route option is one whose pointer value
     does not point past the last entry in the source route.  If the
     packet contains an unexpired source route option, the pointer
     in the option is advanced until either the pointer does point
     past the last address in the option or else the next address is
     not one of the router's own addresses.  In the latter (normal)
     case, the  packet is forwarded (and not delivered locally)
     regardless of the rules below.
  o  The packet is delivered locally and not considered for
     forwarding in the following cases:
     - The packet's destination address exactly matches one of the
        router's IP addresses,
     - The packet's destination address is a limited broadcast
        address ({-1, -1}), and
     - The packet's destination is an IP multicast address which is
        limited to a single subnet (such as 224.0.0.1 or 224.0.0.2)
        and (at least) one of the logical interfaces associated with
        the physical interface on which the packet arrived is a
        member of the destination multicast group.
  o  The packet is passed to the forwarder AND delivered locally in
     the following cases:
     - The packet's destination address is an IP broadcast address
        that addresses at least one of the router's logical
        interfaces but does not address any of the logical
        interfaces associated with the physical interface on which
        the packet arrived
     - The packet's destination is an IP multicast address which is
        not limited to a single subnetwork (such as 224.0.0.1 and
        224.0.0.2 are) and (at least) one of the logical interfaces
        associated with the physical interface on which the packet
        arrived is a member of the destination multicast group.
  o  The packet is delivered locally if the packet's destination
     address is an IP broadcast address (other than a limited
     broadcast address) that addresses at least one of the logical
     interfaces associated with the physical interface on which the
     packet arrived.  The packet is ALSO passed to the forwarder
     unless the link on which the packet arrived uses an IP
     encapsulation that does not encapsulate broadcasts differently
     than unicasts (e.g. by using different Link Layer destination
     addresses).
  o  The packet is passed to the forwarder in all other cases.
  DISCUSSION:
     The purpose of the requirement in the last sentence of the
     fourth bullet is to deal with a directed broadcast to another
     net or subnet on the same physical cable.  Normally, this works
     as expected: the sender sends the broadcast to the router as a
     Link Layer unicast.  The router notes that it arrived as a
     unicast, and therefore must be destined for a different logical
     net (or subnet) than the sender sent it on.  Therefore, the
     router can safely send it as a Link Layer broadcast out the
     same (physical) interface over which it arrived.  However, if
     the router can't tell whether the packet was received as a Link
     Layer unicast, the sentence ensures that the router does the
     safe but wrong thing rather than the unsafe but right thing.
  IMPLEMENTATION:
     As described in Section [5.3.4], packets received as Link Layer
     broadcasts are generally not forwarded.  It may be advantageous
     to avoid passing to the forwarder packets it would later
     discard because of the rules in that section.
     Some Link Layers (either because of the hardware or because of
     special code in the drivers) can deliver to the router copies
     of all Link Layer broadcasts and multicasts it transmits.  Use
     of this feature can simplify the implementation of cases where
     a packet has to both be passed to the forwarder and delivered
     locally, since forwarding the packet will automatically cause
     the router to receive a copy of the packet that it can then
     deliver locally.  One must use care in these circumstances in
     order to prevent treating a received loop-back packet as a
     normal packet that was received (and then being subject to the
     rules of forwarding, etc etc).
     Even in the absence of such a Link Layer, it is of course
     hardly necessary to make a copy of an entire packet in order to
     queue it both for forwarding and for local delivery, though
     care must be taken with fragments, since reassembly is
     performed on locally delivered packets but not on forwarded
     packets.  One simple scheme is to associate a flag with each
     packet on the router's output queue which indicates whether it
     should be queued for local delivery after it has been sent.

Determining the Next Hop Address

  When a router is going to forward a packet, it must determine
  whether it can send it directly to its destination, or whether it
  needs to pass it through another router.  If the latter, it needs
  to determine which router to use.  This section explains how these
  determinations are made.
  This section makes use of the following definitions:
  o  LSRR - IP Loose Source and Record Route option
  o  SSRR - IP Strict Source and Record Route option
  o  Source Route Option - an LSRR or an SSRR
  o  Ultimate Destination Address - where the packet is being sent
     to: the last address in the source route of a source-routed
     packet, or the destination address in the IP header of a non-
     source-routed packet
  o  Adjacent - reachable without going through any IP routers
  o  Next Hop Address - the IP address of the adjacent host or
     router to which the packet should be sent next
  o  Immediate Destination Address - the ultimate destination
     address, except in source routed packets, where it is the next
     address specified in the source route
  o  Immediate Destination - the node, system, router, end-system,
     or whatever that is addressed by the Immediate Destination
     Address.
Immediate Destination Address
     If the destination address in the IP header is one of the
     addresses of the router and the packet contains a Source Route
     Option, the Immediate Destination Address is the address
     pointed at by the pointer in that option if the pointer does
     not point past the end of the option.  Otherwise, the Immediate
     Destination Address is the same as the IP destination address
     in the IP header.
     A router MUST use the Immediate Destination Address, not the
     Ultimate Destination Address, when determining how to handle a
     packet.
     It is an error for more than one source route option to appear
     in a datagram.  If it receives one, it SHOULD discard the
     packet and reply with an ICMP Parameter Problem message whose
     pointer points at the beginning of the second source route
     option.
Local/Remote Decision
     After it has been determined that the IP packet needs to be
     forwarded in accordance with the rules specified in Section
     [5.2.3], the following algorithm MUST be used to determine if
     the Immediate Destination is directly accessible (see
     [INTERNET:2]):
     (1)  For each network interface that has not been assigned any
          IP address (the unnumbered lines as described in Section
          [2.2.7]), compare the router-id of the other end of the
          line to the Immediate Destination Address.  If they are
          exactly equal, the packet can be transmitted through this
          interface.
          DISCUSSION:
             In other words, the router or host at the remote end of
             the line is the destination of the packet or is the
             next step in the source route of a source routed
             packet.
     (2)  If no network interface has been selected in the first
          step, for each IP address assigned to the router:
          (a)  Apply the subnet mask associated with the address to
               this IP address.
               IMPLEMENTATION:
                  The result of this operation will usually have
                  been computed and saved during initialization.
          (b)  Apply the same subnet mask to the Immediate
               Destination Address of the packet.
          (c)  Compare the resulting values. If they are equal to
               each other, the packet can be transmitted through the
               corresponding network interface.
     (3)  If an interface has still not been selected, the Immediate
          Destination is accessible only through some other router.
          The selection of the router and the next hop IP address is
          described in Section [5.2.4.3].
Next Hop Address
     EDITOR'S COMMENTS:
        Note that this section has been extensively rewritten.  The
        original document indicated that Phil Almquist wished to
        revise this section to conform to his "Ruminations on the
        Next Hop" document.  I am under the assumption that the
        working group generally agreed with this goal; there was an
        editor's note from Phil that remained in this document to
        that effect, and the RoNH document contains a "mandatory
        RRWG algorithm".
        So, I have taken said algorithm from RoNH and moved it into
        here.
        Additional useful or interesting information from RoNH has
        been extracted and placed into an appendix to this note.
     The router applies the algorithm in the previous section to
     determine if the Immediate Destination Address is adjacent.  If
     so, the next hop address is the same as the Immediate
     Destination Address.  Otherwise, the packet must be forwarded
     through another router to reach its Immediate Destination.  The
     selection of this router is the topic of this section.
     If the packet contains an SSRR, the router MUST discard the
     packet and reply with an ICMP Bad Source Route error.
     Otherwise, the router looks up the Immediate Destination
     Address in its routing table to determine an appropriate next
     hop address.
     DISCUSSION:
        Per the IP specification, a Strict Source Route must specify
        a sequence of nodes through which the packet must traverse;
        the packet must go from one node of the source route to the
        next, traversing intermediate networks only.  Thus, if the
        router is not adjacent to the next step of the source route,
        the source route can not be fulfilled.  Therefore, the ICMP
        Bad Source Route error.
     The goal of the next-hop selection process is to examine the
     entries in the router's Forwarding Information Base (FIB) and
     select the best route (if there is one) for the packet from
     those available in the FIB.
     Conceptually, any route lookup algorithm starts out with a set
     of candidate routes which consists of the entire contents of
     the FIB.  The algorithm consists of a series of steps which
     discard routes from the set.  These steps are referred to as
     Pruning Rules.  Normally, when the algorithm terminates there
     is exactly one route remaining in the set.  If the set ever
     becomes empty, the packet is discarded because the destination
     is unreachable.  It is also possible for the algorithm to
     terminate when more than one route remains in the set.  In this
     case, the router may arbitrarily discard all but one of them,
     or may perform "load-splitting" by choosing whichever of the
     routes has been least recently used.
     With the exception of rule 3 (Weak TOS), a router MUST use the
     following Pruning Rules when selecting a next hop for a packet.
     If a router does consider TOS when making next-hop decisions,
     the Rule 3 must be applied in the order indicated below.  These
     rules MUST be (conceptually) applied to the FIB in the order
     that they are presented.  (For some historical perspective,
     additional pruning rules, and other common algorithms in use,
     see Appendix E).
     DISCUSSION:
        Rule 3 is optional in that Section [5.3.2] says that a
        router only SHOULD consider TOS when making forwarding
        decisions.
     (1)  Basic Match
          This rule discards any routes to destinations other than
          the Immediate Destination Address of the packet.  For
          example, if a packet's Immediate Destination Address is
          36.144.2.5, this step would discard a route to net
          128.12.0.0 but would retain any routes to net 36.0.0.0,
          any routes to subnet 36.144.0.0, and any default routes.
          More precisely, we assume that each route has a
          destination attribute, called route.dest, and a
          corresponding mask, called route.mask, to specify which
          bits of route.dest are significant.  The Immediate
          Destination Address of the packet being forwarded is
          ip.dest.  This rule discards all routes from the set of
          candidate routes except those for which (route.dest &
          route.mask) = (ip.dest & route.mask).
     (2)  Longest Match
          Longest Match is a refinement of Basic Match, described
          above.  After Basic Match pruning is performed, the
          remaining routes are examined to determine the maximum
          number of bits set in any of their route.mask attributes.
          The step then discards from the set of candidate routes
          any routes which have fewer than that maximum number of
          bits set in their route.mask attributes.
          For example, if a packet's Immediate Destination Address
          is 36.144.2.5 and there are  {route.dest, route.mask}
          pairs of {36.144.2.0, 255.255.255.0}, {36.144.0.5,
          255.255.0.255}, {36.144.0.0, 255.255.0.0}, and {36.0.0.0,
          255.0.0.0}, then this rule would keep only the first two
          pairs; {36.144.2.0, 255.255.255.0} and {36.144.0.5,
          255.255.0.255}.
     (3)  Weak TOS
          Each route has a type of service attribute, called
          route.tos, whose possible values are assumed to be
          identical to those used in the TOS field of the IP header.
          Routing protocols which distribute TOS information fill in
          route.tos appropriately in routes they add to the FIB;
          routes from other routing protocols are treated as if they
          have the default TOS (0000).  The TOS field in the IP
          header of the packet being routed is called ip.tos.
          The set of candidate routes is examined to determine if it
          contains any routes for which route.tos = ip.tos.  If so,
          all routes except those for which route.tos = ip.tos are
          discarded.  If not, all routes except those for which
          route.tos = 0000 are discarded from the set of candidate
          routes.
          Additional discussion of routing based on Weak TOS may be
          found in [ROUTE:11].
          DISCUSSION:
             The effect of this rule is to select only those routes
             which have a TOS that matches the TOS requested in the
             packet.  If no such routes exist then routes with the
             default TOS are considered.  Routes with a non-default
             TOS that is not the TOS requested in the packet are
             never used, even if such routes are the only available
             routes that go to the packet's destination.
     (4)  Best Metric
          Each route has a metric attribute, called route.metric,
          and a routing domain identifier, called route.domain.
          Each member of the set of candidate routes is compared
          with each other member of the set.  If route.domain is
          equal for the two routes and route.metric is strictly
          inferior for one when compared with the other, then the
          one with the inferior metric is discarded from the set.
          The determination of inferior is usually by a simple
          arithmetic comparison, though some protocols may have
          structured metrics requiring more complex comparisons.
     (5)  Vendor Policy
          Vendor Policy is sort of a catch-all to make up for the
          fact that the previously listed rules are often inadequate
          to chose from among the possible routes.  Vendor Policy
          pruning rules are extremely vendor-specific.  See section
          [5.2.4.4].
     This algorithm has two distinct disadvantages.  Presumably, a
     router implementor might develop techniques to deal with these
     disadvantages and make them a part of the Vendor Policy pruning
     rule.
     (1)  IS-IS and OSPF route classes are not directly handled.
     (2)  Path properties other than type of service (e.g. MTU) are
          ignored.
     It is also worth noting a deficiency in the way that TOS is
     supported: routing protocols which support TOS are implicitly
     preferred when forwarding packets which have non-zero TOS
     values.
     The Basic Match and Longest Match pruning rules generalize the
     treatment of a number of particular types of routes.  These
     routes are selected in the following, decreasing, order of
     preference:
     (1)  Host Route: This is a route to a specific end system.
     (2)  Subnetwork Route: This is a route to a particular subnet
          of a network.
     (3)  Default Subnetwork Route: This is a route to all subnets
          of a particular net for which there are not (explicit)
          subnet routes.
     (4)  Network Route: This is a route to a particular network.
     (5)  Default Network Route (also known as the default route):
          This is a route to all networks for which there are no
          explicit routes to the net or any of its subnets.
     If, after application of the pruning rules, the set of routes
     is empty (i.e., no routes were found), the packet MUST be
     discarded and an appropriate ICMP error generated (ICMP Bad
     Source Route if the Immediate Destination Address came from a
     source route option; otherwise, whichever of ICMP Destination
     Host Unreachable or Destination Network Unreachable is
     appropriate, as described in Section [4.3.3.1]).
Administrative Preference
     One suggested mechanism for the Vendor Policy Pruning Rule is
     to use administrative preference.
     Each route has associated with it a preference value, based on
     various attributes of the route (specific mechanisms for
     assignment of preference values are suggested below).  This
     preference value is an integer in the range [0..255], with zero
     being the most preferred and 254 being the least preferred.
     255 is a special value that means that the route should never
     be used.  The first step in the Vendor Policy pruning rule
     discards all but the most preferable routes (and always
     discards routes whose preference value is 255).
     This policy is not safe in that it can easily be misused to
     create routing loops.  Since no protocol ensures that the
     preferences configured for a router are consistent with the
     preferences configured in its neighbors, network managers must
     exercise care in configuring preferences.
     o  Address Match
        It is useful to be able to assign a single preference value
        to all routes (learned from the same routing domain) to any
        of a specified set of destinations, where the set of
        destinations is all destinations that match a specified
        address/mask pair.
     o  Route Class
        For routing protocols which maintain the distinction, it is
        useful to be able to assign a single preference value to all
        routes (learned from the same routing domain) which have a
        particular route class (intra-area, inter-area, external
        with internal metrics, or external with external metrics).
     o  Interface
        It is useful to be able to assign a single preference value
        to all routes (learned from a particular routing domain)
        that would cause packets to be routed out a particular
        logical interface on the router (logical interfaces
        generally map one-to-one onto the router's network
        interfaces, except that any network interface which has
        multiple IP addresses will have multiple logical interfaces
        associated with it).
     o  Source router
        It is useful to be able to assign a single preference value
        to all routes (learned from the same routing domain) which
        were learned from any of a set of routers, where the set of
        routers are those whose updates have a source address which
        match a specified address/mask pair.
     o  Originating AS
        For routing protocols which provide the information, it is
        useful to be able to assign a single preference value to all
        routes (learned from a particular routing domain) which
        originated in another particular routing domain.  For BGP
        routes, the originating AS is the first AS listed in the
        route's AS_PATH attribute.  For OSPF external routes, the
        originating AS may be considered to be the low order 16 bits
        of the route's external route tag if the tag's Automatic bit
        is set and the tag's PathLength is not equal to 3.
     o  External route tag
        It is useful to be able to assign a single preference value
        to all OSPF external routes (learned from the same routing
        domain) whose external route tags match any of a list of
        specified values.  Because the external route tag may
        contain a structured value, it may be useful to provide the
        ability to match particular subfields of the tag.
     o  AS path
        It may be useful to be able to assign a single preference
        value to all BGP routes (learned from the same routing
        domain) whose AS path "matches" any of a set of specified
        values.  It is not yet clear exactly what kinds of matches
        are most useful.  A simple option would be to allow matching
        of all routes for which a particular AS number appears (or
        alternatively, does not appear) anywhere in the route's
        AS_PATH attribute.  A more general but somewhat more
        difficult alternative would be to allow matching all routes
        for which the AS path matches a specified regular
        expression.
Load Splitting
     At the end of the Next-hop selection process, multiple routes
     may still remain.  A router has several options when this
     occurs.  It may arbitrarily discard some of the routes.  It may
     reduce the number of candidate routes by comparing metrics of
     routes from routing domains which are not considered
     equivalent.  It may retain more than one route and employ a
     load-splitting mechanism to divide traffic among them.  Perhaps
     the only thing that can be said about the relative merits of
     the options is that load-splitting is useful in some situations
     but not in others, so a wise implementor who implements load-
     splitting will also provide a way for the network manager to
     disable it.

Unused IP Header Bits: RFC-791 Section 3.1

  The IP header contains several reserved bits, in the Type of
  Service field and in the Flags field.  Routers MUST NOT drop
  packets merely because one or more of these reserved bits has a
  non-zero value.
  Routers MUST ignore and MUST pass through unchanged the values of
  these reserved bits.  If a router fragments a packet, it MUST copy
  these bits into each fragment.
  DISCUSSION:
     Future revisions to the IP protocol may make use of these
     unused bits.  These rules are intended to ensure that these
     revisions can be deployed without having to simultaneously
     upgrade all routers in the Internet.

Fragmentation and Reassembly: RFC-791 Section 3.2

  As was discussed in Section [4.2.2.7], a router MUST support IP
  fragmentation.
  A router MUST NOT reassemble any datagram before forwarding it.
  DISCUSSION:
     A few people have suggested that there might be some topologies
     where reassembly of transit datagrams by routers might improve
     performance.  In general, however, the fact that fragments may
     take different paths to the destination precludes safe use of
     such a feature.
     Nothing in this section should be construed to control or limit
     fragmentation or reassembly performed as a link layer function
     by the router.

Internet Control Message Protocol - ICMP

  General requirements for ICMP were discussed in Section [4.3].
  This section discusses ICMP messages which are sent only by
  routers.
Destination Unreachable
     The ICMP Destination Unreachable message is sent by a router in
     response to a packet which it cannot forward because the
     destination (or next hop) is unreachable or a service is
     unavailable
     A router MUST be able to generate ICMP Destination Unreachable
     messages and SHOULD choose a response code that most closely
     matches the reason why the message is being generated.
     The following codes are defined in [INTERNET:8] and [INTRO:2]:
     0 =  Network Unreachable - generated by a router if a
          forwarding path (route) to the destination network is not
          available;
     1 =  Host Unreachable - generated by a router if a forwarding
          path (route) to the destination host on a directly
          connected network is not available;
     2 =  Protocol Unreachable - generated if the transport protocol
          designated in a datagram is not supported in the transport
          layer of the final destination;
     3 =  Port Unreachable -  generated if the designated transport
          protocol (e.g. UDP) is unable to demultiplex the datagram
          in the transport layer of the final destination but has no
          protocol mechanism to inform the sender;
     4 =  Fragmentation Needed and DF Set - generated if a router
          needs to fragment a datagram but cannot since the DF flag
          is set;
     5 =  Source Route Failed - generated if a router cannot forward
          a packet to the next hop in a source route option;
     6 =  Destination Network Unknown - This code SHOULD NOT be
          generated since it would imply on the part of the router
          that the destination network does not exist (net
          unreachable code 0 SHOULD be used in place of code 6);
     7 =  Destination Host Unknown - generated only when a router
          can determine (from link layer advice) that the
          destination host does not exist;
     11 = Network Unreachable For Type Of Service - generated by a
          router if a forwarding path (route) to the destination
          network with the requested or default TOS is not
          available;
     12 = Host Unreachable For Type Of Service - generated if a
          router cannot forward a packet because its route(s) to the
          destination do not match either the TOS requested in the
          datagram or the default TOS (0).
     The following additional codes are hereby defined:
     13 = Communication Administratively Prohibited - generated if a
          router cannot forward a packet due to administrative
          filtering;
     14 = Host Precedence Violation.  Sent by the first hop router
          to a host to indicate that a requested precedence is not
          permitted for the particular combination of
          source/destination host or network, upper layer protocol,
          and source/destination port;
     15 = Precedence cutoff in effect.  The network operators have
          imposed a minimum level of precedence required for
          operation, the datagram was sent with a precedence below
          this level;
     NOTE: [INTRO:2] defined Code 8 for source host isolated.
     Routers SHOULD NOT generate Code 8; whichever of Codes 0
     (Network Unreachable) and 1 (Host Unreachable) is appropriate
     SHOULD be used instead.  [INTRO:2] also defined Code 9 for
     communication with destination network administratively
     prohibited and Code 10 for communication with destination host
     administratively prohibited.  These codes were intended for use
     by end-to-end encryption devices used by U.S military agencies.
     Routers SHOULD use the newly defined Code 13 (Communication
     Administratively Prohibited) if they administratively filter
     packets.
     Routers MAY have a configuration option that causes Code 13
     (Communication Administratively Prohibited) messages not to be
     generated.  When this option is enabled, no ICMP error message
     is sent in response to a packet which is dropped because its
     forwarding is administratively prohibited.
     Similarly, routers MAY have a configuration option that causes
     Code 14 (Host Precedence Violation) and Code 15 (Precedence
     Cutoff in Effect) messages not to be generated.  When this
     option is enabled, no ICMP error message is sent in response to
     a packet which is dropped  because of a precedence violation.
     Routers MUST use Host Unreachable or Destination Host Unknown
     codes whenever other hosts on the same destination network
     might be reachable; otherwise, the source host may erroneously
     conclude that all hosts on the network are unreachable, and
     that may not be the case.
     [INTERNET:14] describes a slight modification the form of
     Destination Unreachable messages containing Code 4
     (Fragmentation needed and DF set).  A router MUST use this
     modified form when originating Code 4 Destination Unreachable
     messages.
Redirect
     The ICMP Redirect message is generated to inform a host on the
     same subnet that the router used by the host to route certain
     packets should be changed.
     Routers MUST NOT generate the Redirect for Network or Redirect
     for Network and Type of Service messages (Codes 0 and 2)
     specified in [INTERNET:8].  Routers MUST be able to generate
     the Redirect for Host message (Code 1) and SHOULD be able to
     generate the Redirect for Type of Service and Host message
     (Code 3) specified in [INTERNET:8].
     DISCUSSION:
        If the directly-connected network is not subnetted, a router
        can normally generate a network Redirect which applies to
        all hosts on a specified remote network.  Using a network
        rather than a host Redirect may economize slightly on
        network traffic and on host routing table storage.  However,
        the savings are not significant, and subnets create an
        ambiguity about the subnet mask to be used to interpret a
        network Redirect.  In a general subnet environment, it is
        difficult to specify precisely the cases in which network
        Redirects can be used.  Therefore, routers must send only
        host (or host and type of service) Redirects.
     A Code 3 (Redirect for Host and Type of Service) message is
     generated when the packet provoking the redirect has a
     destination for which the path chosen by the router would
     depend (in part) on the TOS requested.
     Routers which can generate Code 3 redirects (Host and Type of
     Service) MUST have a configuration option (which defaults to
     on) to enable Code 1 (Host) redirects to be substituted for
     Code 3 redirects.  A router MUST send a Code 1 Redirect in
     place of a Code 3 Redirect if it has been configured to do so.
     If a router is not able to generate Code 3 Redirects then it
     MUST generate Code 1 Redirects in situations where a Code 3
     Redirect is called for.
     Routers MUST NOT generate a Redirect Message unless all of the
     following conditions are met:
     o  The packet is being forwarded out the same physical
        interface that it was received from,
     o  The IP source address in the packet is on the same Logical
        IP (sub)network as the next-hop IP address, and
     o  The packet does not contain an IP source route option.
     The source address used in the ICMP Redirect MUST belong to the
     same logical (sub)net as the destination address.
     A router using a routing protocol (other than static routes)
     MUST NOT consider paths learned from ICMP Redirects when
     forwarding a packet.  If a router is not using a routing
     protocol, a router MAY have a configuration which, if set,
     allows the router to consider routes learned via ICMP Redirects
     when forwarding packets.
     DISCUSSION:
        ICMP Redirect is a mechanism for routers to convey routing
        information to hosts.  Routers use other mechanisms to learn
        routing information, and therefore have no reason to obey
        redirects.  Believing a redirect which contradicted the
        router's other information would likely create routing
        loops.
        On the other hand, when a router is not acting as a router,
        it MUST comply with the behavior required of a host.
Time Exceeded
     A router MUST generate a Time Exceeded message Code 0 (In
     Transit) when it discards a packet due to an expired TTL field.
     A router MAY have a per-interface option to disable origination
     of these messages on that interface, but that option MUST
     default to allowing the messages to be originated.

INTERNET GROUP MANAGEMENT PROTOCOL - IGMP

  IGMP [INTERNET:4] is a protocol used between hosts and multicast
  routers on a single physical network to establish hosts'
  membership in particular multicast groups.  Multicast routers use
  this information, in conjunction with a multicast routing
  protocol, to support IP multicast forwarding across the Internet.
  A router SHOULD implement the multicast router part of IGMP.

SPECIFIC ISSUES

Time to Live (TTL)

  The Time-to-Live (TTL) field of the IP header is defined to be a
  timer limiting the lifetime of a datagram.  It is an 8-bit field
  and the units are seconds.  Each router (or other module) that
  handles a packet MUST decrement the TTL by at least one, even if
  the elapsed time was much less than a second.  Since this is very
  often the case, the TTL is effectively a hop count limit on how
  far a datagram can propagate through the Internet.
  When a router forwards a packet, it MUST reduce the TTL by at
  least one.  If it holds a packet for more than one second, it MAY
  decrement the TTL by one for each second.
  If the TTL is reduced to zero (or less), the packet MUST be
  discarded, and if the destination is not a multicast address the
  router MUST send an ICMP Time Exceeded message, Code 0 (TTL
  Exceeded in Transit) message to the source.  Note that a router
  MUST NOT discard an IP unicast or broadcast packet with a non-zero
  TTL merely because it can predict that another router on the path
  to the packet's final destination will decrement the TTL to zero.
  However, a router MAY do so for IP multicasts, in order to more
  efficiently implement IP multicast's expanding ring search
  algorithm (see [INTERNET:4]).
  DISCUSSION:
     The IP TTL is used, somewhat schizophrenically, as both a hop
     count limit and a time limit.  Its hop count function is
     critical to ensuring that routing problems can't melt down the
     network by causing packets to loop infinitely in the network.
     The time limit function is used by transport protocols such as
     TCP to ensure reliable data transfer.  Many current
     implementations treat TTL as a pure hop count, and in parts of
     the Internet community there is a strong sentiment that the
     time limit function should instead be performed by the
     transport protocols that need it.
     In this specification, we have reluctantly decided to follow
     the strong belief among the router vendors that the time limit
     function should be optional.  They argued that implementation
     of the time limit function is difficult enough that it is
     currently not generally done.  They further pointed to the lack
     of documented cases where this shortcut has caused TCP to
     corrupt data (of course, we would expect the problems created
     to be rare and difficult to reproduce, so the lack of
     documented cases provides little reassurance that there haven't
     been a number of undocumented cases).
     IP multicast notions such as the expanding ring search may not
     work as expected unless the TTL is treated as a pure hop count.
     The same thing is somewhat true of traceroute.
     ICMP Time Exceeded messages are required because the traceroute
     diagnostic tool depends on them.
     Thus, the tradeoff is between severely crippling, if not
     eliminating, two very useful tools vs. a very rare and
     transient data transport problem (which may not occur at all).

Type of Service (TOS)

  The Type-of-Service byte in the IP header is divided into three
  sections:  the Precedence field (high-order 3 bits), a field that
  is customarily called Type of Service or "TOS (next 4 bits), and a
  reserved bit (the low order bit).  Rules governing the reserved
  bit were described in Section [4.2.2.3].  The Precedence field
  will be discussed in Section [5.3.3].  A more extensive discussion
  of the TOS field and its use can be found in [ROUTE:11].
  A router SHOULD consider the TOS field in a packet's IP header
  when deciding how to forward it.  The remainder of this section
  describes the rules that apply to routers that conform to this
  requirement.
  A router MUST maintain a TOS value for each route in its routing
  table.  Routes learned via a routing protocol which does not
  support TOS MUST be assigned a TOS of zero (the default TOS).
  To choose a route to a destination, a router MUST use an algorithm
  equivalent to the following:
  (1)  The router locates in its routing table all available routes
       to the destination (see Section [5.2.4]).
  (2)  If there are none, the router drops the packet because the
       destination is unreachable.  See section [5.2.4].
  (3)  If one or more of those routes have a TOS that exactly
       matches the TOS specified in the packet, the router chooses
       the route with the best metric.
  (4)  Otherwise, the router repeats the above step, except looking
       at routes whose TOS is zero.
  (5)  If no route was chosen above, the router drops the packet
       because the destination is unreachable.  The router returns
       an ICMP Destination Unreachable error specifying the
       appropriate code: either Network Unreachable with Type of
       Service (code 11) or Host Unreachable with Type of Service
       (code 12).
  DISCUSSION:
     Although TOS has been little used in the past, its use by hosts
     is now mandated by the Requirements for Internet Hosts RFCs
     ([INTRO:2] and [INTRO:3]).  Support for TOS in routers may
     become a MUST in the future, but is a SHOULD for now until we
     get more experience with it and can better judge both its
     benefits and its costs.
     Various people have proposed that TOS should affect other
     aspects of the forwarding function.  For example:
     (1)  A router could place packets which have the Low Delay bit
          set ahead of other packets in its output queues.
     (2)  a router is forced to discard packets, it could try to
          avoid discarding those which have the High Reliability bit
          set.
     These ideas have been explored in more detail in [INTERNET:17]
     but we don't yet have enough experience with such schemes to
     make requirements in this area.

IP Precedence

  This section specifies requirements and guidelines for appropriate
  processing of the IP Precedence field in routers.  Precedence is a
  scheme for allocating resources in the network based on the
  relative importance of different traffic flows.  The IP
  specification defines specific values to be used in this field for
  various types of traffic.
  The basic mechanisms for precedence processing in a router are
  preferential resource allocation, including both precedence-
  ordered queue service and precedence-based congestion control, and
  selection of Link Layer priority features.  The router also
  selects the IP precedence for routing, management and control
  traffic it originates.  For a more extensive discussion of IP
  Precedence and its implementation see [FORWARD:6].
  Precedence-ordered queue service, as discussed in this section,
  includes but is not limited to the queue for the forwarding
  process and queues for outgoing links.  It is intended that a
  router supporting precedence should also use the precedence
  indication at whatever points in its processing are concerned with
  allocation of finite resources, such as packet buffers or Link
  Layer connections.  The set of such points is implementation-
  dependent.
  DISCUSSION:
     Although the Precedence field was originally provided for use
     in DOD systems where large traffic surges or major damage to
     the network are viewed as inherent threats, it has useful
     applications for many non-military IP networks.  Although the
     traffic handling capacity of networks has grown greatly in
     recent years, the traffic generating ability of the users has
     also grown, and network overload conditions still occur at
     times.  Since IP-based routing and management protocols have
     become more critical to the successful operation of the
     Internet, overloads present two additional risks to the
     network:
     (1)  High delays may result in routing protocol packets being
          lost.  This may cause the routing protocol to falsely
          deduce a topology change and propagate this false
          information to other routers.  Not only can this cause
          routes to oscillate, but an extra processing burden may be
          placed on other routers.
     (2)  High delays may interfere with the use of network
          management tools to analyze and perhaps correct or relieve
          the problem in the network that caused the overload
          condition to occur.
     Implementation and appropriate use of the Precedence mechanism
     alleviates both of these problems.
Precedence-Ordered Queue Service
     Routers SHOULD implement precedence-ordered queue service.
     Precedence-ordered queue service means that when a packet is
     selected for output on a (logical) link, the packet of highest
     precedence that has been queued for that link is sent.  Routers
     that implement precedence-ordered queue service MUST also have
     a configuration option to suppress precedence-ordered queue
     service in the Internet Layer.
     Any router MAY implement other policy-based throughput
     management procedures that result in other than strict
     precedence ordering, but it MUST be configurable to suppress
     them (i.e., use strict ordering).
     As detailed in Section [5.3.6], routers that implement
     precedence-ordered queue service discard low precedence packets
     before discarding high precedence packets for congestion
     control purposes.
     Preemption (interruption of processing or transmission of a
     packet) is not envisioned as a function of the Internet Layer.
     Some protocols at other layers may provide preemption features.
Lower Layer Precedence Mappings
     Routers that implement precedence-ordered queueing MUST
     IMPLEMENT, and other routers SHOULD IMPLEMENT, Lower Layer
     Precedence Mapping.
     A router which implements Lower Layer Precedence Mapping:
     o  MUST be able to map IP Precedence to Link Layer priority
        mechanisms for link layers that have such a feature defined.
     o  MUST have a configuration option to select the Link Layer's
        default priority treatment for all IP traffic
     o  SHOULD be able to configure specific nonstandard mappings of
        IP precedence values to Link Layer priority values for each
        interface.
     DISCUSSION:
        Some research questions the workability of the priority
        features of some Link Layer protocols, and some networks may
        have faulty implementations of the link layer priority
        mechanism.  It seems prudent to provide an escape mechanism
        in case such problems show up in a network.
        On the other hand, there are proposals to use novel queueing
        strategies to implement special services such as low-delay
        service.  Special services and queueing strategies to
        support them need further research and experimentation
        before they are put into widespread use in the Internet.
        Since these requirements are intended to encourage (but not
        force) the use of precedence features in the hope of
        providing better Internet service to all users, routers
        supporting precedence-ordered queue service should default
        to maintaining strict precedence ordering regardless of the
        type of service requested.
        Implementors may wish to consider that correct link layer
        mapping of IP precedence is required by DOD policy for
        TCP/IP systems used on DOD networks.
Precedence Handling For All Routers
     A router (whether or not it employs precedence-ordered queue
     service):
     (1)  MUST accept and process incoming traffic of all precedence
          levels normally, unless it has been administratively
          configured to do otherwise.
     (2)  MAY implement a validation filter to administratively
          restrict the use of precedence levels by particular
          traffic sources.  If provided, this filter MUST NOT filter
          out or cut off the following sorts of ICMP error messages:
          Destination Unreachable, Redirect, Time Exceeded, and
          Parameter Problem.  If this filter is provided, the
          procedures required for packet filtering by addresses are
          required for this filter also.
          DISCUSSION:
             Precedence filtering should be applicable to specific
             source/destination IP Address pairs, specific
             protocols, specific ports, and so on.
          An ICMP Destination Unreachable message with code 14
          SHOULD be sent when a packet is dropped by the validation
          filter, unless this has been suppressed by configuration
          choice.
     (3)  MAY implement a cutoff function which allows the router to
          be set to refuse or drop traffic with precedence below a
          specified level.  This function may be activated by
          management actions or by some implementation dependent
          heuristics, but there MUST be a configuration option to
          disable any heuristic mechanism that operates without
          human intervention.  An ICMP Destination Unreachable
          message with code 15 SHOULD be sent when a packet is
          dropped by the cutoff function, unless this has been
          suppressed by configuration choice.
          A router MUST NOT refuse to forward datagrams with IP
          precedence of 6 (Internetwork Control) or 7 (Network
          Control) solely due to precedence cutoff.  However, other
          criteria may be used in conjunction with precedence cutoff
          to filter high precedence traffic.
          DISCUSSION:
             Unrestricted precedence cutoff could result in an
             unintentional cutoff of routing and control traffic.
             In general, host traffic should be restricted to a
             value of 5 (CRITIC/ECP) or below although this is not a
             requirement and may not be valid in certain systems.
     (4)  MUST NOT change precedence settings on packets it did not
          originate.
     (5)  SHOULD be able to configure distinct precedence values to
          be used for each routing or management protocol supported
          (except for those protocols, such as OSPF, which specify
          which precedence value must be used).
     (6)  MAY be able to configure routing or management traffic
          precedence values independently for each peer address.
     (7)  MUST respond appropriately to Link Layer precedence-
          related error indications where provided.  An ICMP
          Destination Unreachable message with code 15 SHOULD be
          sent when a packet is dropped because a link cannot accept
          it due to a precedence-related condition, unless this has
          been suppressed by configuration choice.
          DISCUSSION:
             The precedence cutoff mechanism described in (3) is
             somewhat controversial.  Depending on the topological
             location of the area affected by the cutoff, transit
             traffic may be directed by routing protocols into the
             area of the cutoff, where it will be dropped.  This is
             only a problem if another path which is unaffected by
             the cutoff exists between the communicating points.
             Proposed ways of avoiding this problem include
             providing some minimum bandwidth to all precedence
             levels even under overload conditions, or propagating
             cutoff information in routing protocols.  In the
             absence of a widely accepted (and implemented) solution
             to this problem, great caution is recommended in
             activating cutoff mechanisms in transit networks.
             A transport layer relay could legitimately provide the
             function prohibited by (4) above.  Changing precedence
             levels may cause subtle interactions with TCP and
             perhaps other protocols; a correct design is a non-
             trivial task.
             The intent of (5) and (6) (and the discussion of IP
             Precedence in ICMP messages in Section [4.3.2]) is that
             the IP precedence bits should be appropriately set,
             whether or not this router acts upon those bits in any
             other way.  We expect that in the future specifications
             for routing protocols and network management protocols
             will specify how the IP Precedence should be set for
             messages sent by those protocols.
             The appropriate response for (7) depends on the link
             layer protocol in use.  Typically, the router should
             stop trying to send offensive traffic to that
             destination for some period of time, and should return
             an ICMP Destination Unreachable message with code 15
             (service not available for precedence requested) to the
             traffic source.  It also should not try to reestablish
             a preempted Link Layer connection for some period of
             time.

Forwarding of Link Layer Broadcasts

  The encapsulation of IP packets in most Link Layer protocols
  (except PPP) allows a receiver to distinguish broadcasts and
  multicasts from unicasts simply by examining the Link Layer
  protocol headers (most commonly, the Link Layer destination
  address).  The rules in this section which refer to Link Layer
  broadcasts apply only to Link Layer protocols which allow
  broadcasts to be distinguished; likewise, the rules which refer to
  Link Layer multicasts apply only to Link Layer protocols which
  allow multicasts to be distinguished.
  A router MUST NOT forward any packet which the router received as
  a Link Layer broadcast (even if the IP destination address is also
  some form of broadcast address) unless the packet is an all-
  subnets-directed broadcast being forwarded as specified in
  [INTERNET:3].
  DISCUSSION:
     As noted in Section [5.3.5.3], forwarding of all-subnets-
     directed broadcasts in accordance with [INTERNET:3] is optional
     and is not something that routers do by default.
  A router MUST NOT forward any packet which the router received as
  a Link Layer multicast unless the packet's destination address is
  an IP multicast address.
  A router SHOULD silently discard a packet that is received via a
  Link Layer broadcast but does not specify an IP multicast or IP
  broadcast destination address.
  When a router sends a packet as a Link Layer broadcast, the IP
  destination address MUST be a legal IP broadcast or IP multicast
  address.

Forwarding of Internet Layer Broadcasts

  There are two major types of IP broadcast addresses; limited
  broadcast and directed broadcast.  In addition, there are three
  subtypes of directed broadcast; a broadcast directed to a
  specified network, a broadcast directed to a specified subnetwork,
  and a broadcast directed to all subnets of a specified network.
  Classification by a router of a broadcast into one of these
  categories depends on the broadcast address and on the router's
  understanding (if any) of the subnet structure of the destination
  network.  The same broadcast will be classified differently by
  different routers.
  A limited IP broadcast address is defined to be all-ones: { -1, -1
  } or 255.255.255.255.
  A net-directed broadcast is composed of the network portion of the
  IP address with a local part of all-ones, { <Network-number>, -1
  }.  For example, a Class A net broadcast address is
  net.255.255.255, a Class B net broadcast address is
  net.net.255.255 and a Class C net broadcast address is
  net.net.net.255 where net is a byte of the network address.
  An all-subnets-directed broadcast is composed of the network part
  of the IP address with a subnet and a host part of all-ones, {
  <Network-number>, -1, -1 }.  For example, an all-subnets broadcast
  on a subnetted class B network is net.net.255.255.  A network must
  be known to be subnetted and the subnet part must be all-ones
  before a broadcast can be classified as all-subnets-directed.
  A subnet-directed broadcast address is composed of the network and
  subnet part of the IP address with a host part of all-ones, {
  <Network-number>, <Subnet-number>, -1 }.  For example, a subnet-
  directed broadcast to subnet 2 of a class B network might be
  net.net.2.255 (if the subnet mask was 255.255.255.0) or
  net.net.1.127 (if the subnet mask was 255.255.255.128).  A network
  must be known to be subnetted and the net and subnet part must not
  be all-ones before an IP broadcast can be classified as subnet-
  directed.
  As was described in Section [4.2.3.1], a router may encounter
  certain non-standard IP broadcast addresses:
  o  0.0.0.0 is an obsolete form of the limited broadcast address
  o  { broadcast address.
  o  { broadcast address.
  o  { form of a subnet-directed broadcast address.
  As was described in that section, packets addressed to any of
  these addresses SHOULD be silently discarded, but if they are not,
  they MUST be treated in accordance with the same rules that apply
  to packets addressed to the non-obsolete forms of the broadcast
  addresses described above.  These rules are described in the next
  few sections.
Limited Broadcasts
     Limited broadcasts MUST NOT be forwarded.  Limited broadcasts
     MUST NOT be discarded.  Limited broadcasts MAY be sent and
     SHOULD be sent instead of directed broadcasts where limited
     broadcasts will suffice.
     DISCUSSION:
        Some routers contain UDP servers which function by resending
        the requests (as unicasts or directed broadcasts) to other
        servers.  This requirement should not be interpreted as
        prohibiting such servers.  Note, however, that such servers
        can easily cause packet looping if misconfigured.  Thus,
        providers of such servers would probably be well-advised to
        document their setup carefully and to consider carefully the
        TTL on packets which are sent.
Net-directed Broadcasts
     A router MUST classify as net-directed broadcasts all valid,
     directed broadcasts destined for a remote network or an
     attached nonsubnetted network.  A router MUST forward net-
     directed broadcasts.  Net-directed broadcasts MAY be sent.
     A router MAY have an option to disable receiving net-directed
     broadcasts on an interface and MUST have an option to disable
     forwarding net-directed broadcasts.  These options MUST default
     to permit receiving and forwarding net-directed broadcasts.
     DISCUSSION:
        There has been some debate about forwarding or not
        forwarding directed broadcasts.  In this memo we have made
        the forwarding decision depend on the router's knowledge of
        the subnet mask for the destination network.  Forwarding
        decisions for subnetted networks should be made by routers
        with an understanding of the subnet structure.  Therefore,
        in general, routers must forward directed broadcasts for
        networks they are not attached to and for which they do not
        understand the subnet structure.  One router may interpret
        and handle the same IP broadcast packet differently than
        another, depending on its own understanding of the structure
        of the destination (sub)network.
All-subnets-directed Broadcasts
     A router MUST classify as all-subnets-directed broadcasts all
     valid directed broadcasts destined for a directly attached
     subnetted network which have all-ones in the subnet part of the
     address.  If the destination network is not subnetted, the
     broadcast MUST be treated as a net-directed broadcast.
     A router MUST forward an all-subnets-directed broadcast as a
     link level broadcast out all physical interfaces connected to
     the IP network addressed by the broadcast, except that:
     o  A router MUST NOT forward an all-subnet-directed broadcast
        that was received by the router as a Link Layer broadcast,
        unless the router is forwarding the broadcast in accordance
        with [INTERNET:3] (see below).
     o  If a router receives an all-subnets-directed broadcast over
        a network which does not indicate via Link Layer framing
        whether the frame is a broadcast or a unicast, the packet
        MUST NOT be forwarded to any network which likewise does not
        indicate whether a frame is a broadcast.
     o  A router MUST NOT forward an all-subnets-directed broadcast
        if the router is configured not to forward such broadcasts.
        A router MUST have a configuration option to deny forwarding
        of all-subnets-directed broadcasts.  The configuration
        option MUST default to permit forwarding of all-subnets-
        directed broadcasts.
     EDITOR'S COMMENTS:
        The algorithm presented here is broken.  The working group
        explicitly desired this algorithm, knowing its failures.
        The second bullet, above, prevents All Subnets Directed
        Broadcasts from traversing more than one PPP (or other
        serial) link in a row.  Such a topology is easily conceived.
        Suppose that some corporation builds its corporate backbone
        out of PPP links, connecting routers at geographically
        dispersed locations.  Suppose that this corporation has 3
        sites (S1, S2, and S3) and there is a router at each site
        (R1, R2, and R3).  At each site there are also several LANs
        connected to the local router.  Let there be a PPP link
        connecting S1 to S2 and one connecting S2 to S3 (i.e. the
        links are R1-R2 and R2-R3).  So, if a host on a LAN at S1
        sends a All Subnets Directed Broadcast, R1 will forward the
        broadcast over the R1-R2 link to R2.  R2 will forward the
        broadcast to the LAN(s) connected to R2.  Since the PPP does
        not differentiate broadcast from non-broadcast frames, R2
        will NOT forward the broadcast onto the R2-R3 link.
        Therefore, the broadcast will not reach S3.
     [INTERNET:3] describes an alternative set of rules for
     forwarding of all-subnets-directed broadcasts (called multi-
     subnet-broadcasts in that document).  A router MAY IMPLEMENT
     that alternative set of rules, but MUST use the set of rules
     described above unless explicitly configured to use the
     [INTERNET:3] rules.  If routers will do [INTERNET:3]-style
     forwarding, then the router MUST have a configuration option
     which MUST default to doing the rules presented in this
     document.
     DISCUSSION:
        As far as we know, the rules for multi-subnet broadcasts
        described in [INTERNET:3] have never been implemented,
        suggesting that either they are too complex or the utility
        of multi-subnet broadcasts is low.  The rules described in
        this section match current practice.  In the future, we
        expect that IP multicast (see [INTERNET:4]) will be used to
        better solve the sorts of problems that multi-subnets
        broadcasts were intended to address.
        We were also concerned that hosts whose system managers
        neglected to configure with a subnet mask could
        unintentionally send multi-subnet broadcasts.
     A router SHOULD NOT originate all-subnets broadcasts, except as
     required by Section [4.3.3.9] when sending ICMP Address Mask
     Replies on subnetted networks.
     DISCUSSION:
        The current intention is to decree that (like 0-filled IP
        broadcasts) the notion of the all-subnets broadcast is
        obsolete.  It should be treated as a directed broadcast to
        the first subnet of the net in question that it appears on.
        Routers may implement a switch (default off) which if turned
        on enables the [INTERNET:3] behavior for all-subnets
        broadcasts.
        If a router has a configuration option to allow for
        forwarding all-subnet broadcasts, it should use a spanning
        tree, RPF, or other multicast forwarding algorithm (which
        may be computed for other purposes such as bridging or OSPF)
        to distribute the all-subnets broadcast efficiently.  In
        general, it is better to use an IP multicast address rather
        than an all-subnets broadcast.
Subnet-directed Broadcasts
     A router MUST classify as subnet-directed broadcasts all valid
     directed broadcasts destined for a directly attached subnetted
     network in which the subnet part is not all-ones.  If the
     destination network is not subnetted, the broadcast MUST be
     treated as a net-directed broadcast.
     A router MUST forward subnet-directed broadcasts.
     A router MUST have a configuration option to prohibit
     forwarding of subnet-directed broadcasts.  Its default setting
     MUST permit forwarding of subnet-directed broadcasts.
     A router MAY have a configuration option to prohibit forwarding
     of subnet-directed broadcasts from a source on a network on
     which the router has an interface.  If such an option is
     provided, its default setting MUST permit forwarding of
     subnet-directed broadcasts.

Congestion Control

  Congestion in a network is loosely defined as a condition where
  demand for resources (usually bandwidth or CPU time) exceeds
  capacity.  Congestion avoidance tries to prevent demand from
  exceeding capacity, while congestion recovery tries to restore an
  operative state.  It is possible for a router to contribute to
  both of these mechanisms.  A great deal of effort has been spent
  studying the problem.  The reader is encouraged to read
  [FORWARD:2] for a survey of the work.  Important papers on the
  subject include [FORWARD:3], [FORWARD:4], [FORWARD:5], and
  [INTERNET:10], among others.
  The amount of storage that router should have available to handle
  peak instantaneous demand when hosts use reasonable congestion
  policies, such as described in [FORWARD:5], is a function of the
  product of the bandwidth of the link times the path delay of the
  flows using the link, and therefore storage should increase as
  this Bandwidth*Delay product increases.  The exact function
  relating storage capacity to probability of discard is not known.
  When a router receives a packet beyond its storage capacity it
  must (by definition, not by decree) discard it or some other
  packet or packets.  Which packet to discard is the subject of much
  study but, unfortunately, little agreement so far.
  A router MAY discard the packet it has just received; this is the
  simplest but not the best policy.  It is considered better policy
  to randomly pick some transit packet on the queue and discard it
  (see [FORWARD:2]).  A router MAY use this Random Drop algorithm to
  determine which packet to discard.
  If a router implements a discard policy (such as Random Drop)
  under which it chooses a packet to discard from among a pool of
  eligible packets:
  o  If precedence-ordered queue service (described in Section
     [5.3.3.1]) is implemented and enabled, the router MUST NOT
     discard a packet whose IP precedence is higher than that of a
     packet which is not discarded.
  o  A router MAY protect packets whose IP headers request the
     maximize reliability TOS, except where doing so would be in
     violation of the previous rule.
  o  A router MAY protect fragmented IP packets, on the theory that
     dropping a fragment of a datagram may increase congestion by
     causing all fragments of the datagram to be retransmitted by
     the source.
  o  To help prevent routing perturbations or disruption of
     management functions, the router MAY protect packets used for
     routing control, link control, or network management from being
     discarded.  Dedicated routers (i.e.. routers which are not also
     general purpose hosts, terminal servers, etc.) can achieve an
     approximation of this rule by protecting packets whose source
     or destination is the router itself.
  Advanced methods of congestion control include a notion of
  fairness, so that the 'user' that is penalized by losing a packet
  is the one that contributed the most to the congestion.  No matter
  what mechanism is implemented to deal with bandwidth congestion
  control, it is important that the CPU effort expended be
  sufficiently small that the router is not driven into CPU
  congestion also.
  As described in Section [4.3.3.3], this document recommends that a
  router should not send a Source Quench to the sender of the packet
  that it is discarding.  ICMP Source Quench is a very weak
  mechanism, so it is not necessary for a router to send it, and
  host software should not use it exclusively as an indicator of
  congestion.

Martian Address Filtering

  An IP source address is invalid if it is an IP broadcast address
  or is not a class A, B, or C address.
  An IP destination address is invalid if it is not a class A, B, C,
  or D address.
  A router SHOULD NOT forward any packet which has an invalid IP
  source address or a source address on network 0.  A router SHOULD
  NOT forward, except over a loopback interface, any packet which
  has a source address on network 127.  A router MAY have a switch
  which allows the network manager to disable these checks.  If such
  a switch is provided, it MUST default to performing the checks.
  A router SHOULD NOT forward any packet which has an invalid IP
  destination address or a destination address on network 0.  A
  router SHOULD NOT forward, except over a loopback interface, any
  packet which has a destination address on network 127.  A router
  MAY have a switch which allows the network manager to disable
  these checks.  If such a switch is provided, it MUST default to
  performing the checks.
  If a router discards a packet because of these rules, it SHOULD
  log at least the IP source address, the IP destination address,
  and, if the problem was with the source address, the physical
  interface on which the packet was received and the Link Layer
  address of the host or router from which the packet was received.

Source Address Validation

  A router SHOULD IMPLEMENT the ability to filter traffic based on a
  comparison of the source address of a packet and the forwarding
  table for a logical interface on which the packet was received.
  If this filtering is enabled, the router MUST silently discard a
  packet if the interface on which the packet was received is not
  the interface on which a packet would be forwarded to reach the
  address contained in the source address.  In simpler terms, if a
  router wouldn't route a packet containing this address through a
  particular interface, it shouldn't believe the address if it
  appears as a source address in a packet read from this interface.
  If this feature is implemented, it MUST be disabled by default.
  DISCUSSION:
     This feature can provide useful security improvements in some
     situations, but can erroneously discard valid packets in
     situations where paths are asymmetric.

Packet Filtering and Access Lists

  As a means of providing security and/or limiting traffic through
  portions of a network a router SHOULD provide the ability to
  selectively forward (or filter) packets.  If this capability is
  provided, filtering of packets MUST be configurable either to
  forward all packets or to selectively forward them based upon the
  source and destination addresses.  Each source and destination
  address SHOULD allow specification of an arbitrary mask.
  If supported, a router MUST be configurable to allow one of an
  o  Include list -  specification of a list of address pairs to be
     forwarded, or an
  o  Exclude list -  specification of a list of address pairs NOT to
     be forwarded.
  A router MAY provide a configuration switch which allows a choice
  between specifying an include or an exclude list.
  A value matching any address (e.g. a keyword any or an address
  with a mask of all 0's) MUST be allowed as a source and/or
  destination address.
  In addition to address pairs, the router MAY allow any combination
  of transport and/or application protocol and source and
  destination ports to be specified.
  The router MUST allow packets to be silently discarded (i.e..
  discarded without an ICMP error message being sent).
  The router SHOULD allow an appropriate ICMP unreachable message to
  be sent when a packet is discarded. The ICMP message SHOULD
  specify Communication Administratively Prohibited (code 13) as the
  reason for the destination being unreachable.
  The router SHOULD allow the sending of ICMP destination
  unreachable messages (code 13) to be configured for each
  combination of address pairs, protocol types, and ports it allows
  to be specified.
  The router SHOULD count and SHOULD allow selective logging of
  packets not forwarded.

5.3.10 Multicast Routing

  An IP router SHOULD support forwarding of IP multicast packets,
  based either on static multicast routes or on routes dynamically
  determined by a multicast routing protocol (e.g., DVMRP
  [ROUTE:9]).  A router that forwards IP multicast packets is called
  a multicast router.

5.3.11 Controls on Forwarding

  For each physical interface, a router SHOULD have a configuration
  option which specifies whether forwarding is enabled on that
  interface.  When forwarding on an interface is disabled, the
  router:
  o  MUST silently discard any packets which are received on that
     interface but are not addressed to the router
  o  MUST NOT send packets out that interface, except for datagrams
     originated by the router
  o  MUST NOT announce via any routing protocols the availability of
     paths through the interface
  DISCUSSION:
     This feature allows the network manager to essentially turn off
     an interface but leaves it accessible for network management.
     Ideally, this control would apply to logical rather than
     physical interfaces, but cannot because there is no known way
     for a router to determine which logical interface a packet
     arrived on when there is not a one-to-one correspondence
     between logical and physical interfaces.

5.3.12 State Changes

  During the course of router operation, interfaces may fail or be
  manually disabled, or may become available for use by the router.
  Similarly, forwarding may be disabled for a particular interface
  or for the entire router or may be (re)enabled.  While such
  transitions are (usually) uncommon, it is important that routers
  handle them correctly.

5.3.12.1 When a Router Ceases Forwarding

     When a router ceases forwarding it MUST stop advertising all
     routes, except for third party routes.  It MAY continue to
     receive and use routes from other routers in its routing
     domains.  If the forwarding database is retained, the router
     MUST NOT cease timing the routes in the forwarding database.
     If routes that have been received from other routers are
     remembered, the router MUST NOT cease timing the routes which
     it has remembered.  It MUST discard any routes whose timers
     expire while forwarding is disabled, just as it would do if
     forwarding were enabled.
     DISCUSSION:
        When a router ceases forwarding, it essentially ceases being
        a router.  It is still a host, and must follow all of the
        requirements of Host Requirements [INTRO: 2].  The router
        may still be a passive member of one or more routing
        domains, however.  As such, it is allowed to maintain its
        forwarding database by listening to other routers in its
        routing domain.  It may not, however, advertise any of the
        routes in its forwarding database, since it itself is doing
        no forwarding.  The only exception to this rule is when the
        router is advertising a route which uses only some other
        router, but which this router has been asked to advertise.
     A router MAY send ICMP destination unreachable (host
     unreachable) messages to the senders of packets that it is
     unable to forward. It SHOULD NOT send ICMP redirect messages.
     DISCUSSION:
        Note that sending an ICMP destination unreachable (host
        unreachable) is a router action.  This message should not be
        sent by hosts.   This exception to the rules for hosts is
        allowed so that packets may be rerouted in the shortest
        possible time, and so that black holes are avoided.

5.3.12.2 When a Router Starts Forwarding

     When a router begins forwarding, it SHOULD expedite the sending
     of new routing information to all routers with which it
     normally exchanges routing information.

5.3.12.3 When an Interface Fails or is Disabled

     If an interface fails or is disabled a router MUST remove and
     stop advertising all routes in its forwarding database which
     make use of that interface.  It MUST disable all static routes
     which make use of that interface.  If other routes to the same
     destination and TOS are learned or remembered by the router,
     the router MUST choose the best alternate, and add it to its
     forwarding database.  The router SHOULD send ICMP destination
     unreachable or ICMP redirect messages, as appropriate, in reply
     to all packets which it is unable to forward due to the
     interface being unavailable.

5.3.12.4 When an Interface is Enabled

     If an interface which had not been available becomes available,
     a router MUST reenable any static routes which use that
     interface.  If routes which would use that interface are
     learned by the router,  then these routes MUST be evaluated
     along with all of the other learned routes, and the router MUST
     make a decision as to which routes should be placed in the
     forwarding database.  The implementor is referred to Chapter
     [7], Application Layer - Routing Protocols for further
     information on how this decision is made.
     A router SHOULD expedite the sending of new routing information
     to all routers with which it normally exchanges routing
     information.

5.3.13 IP Options

  Several options, such as Record Route and Timestamp, contain slots
  into which a router inserts its address when forwarding the
  packet.  However, each such option has a finite number of slots,
  and therefore a router may find that there is not free slot into
  which it can insert its address.  No requirement listed below
  should be construed as requiring a router to insert its address
  into an option that has no remaining slot to insert it into.
  Section [5.2.5] discusses how a router must choose which of its
  addresses to insert into an option.

5.3.13.1 Unrecognized Options

     Unrecognized IP options in forwarded packets MUST be passed
     through unchanged.

5.3.13.2 Security Option

     Some environments require the Security option in every packet;
     such a requirement is outside the scope of this document and
     the IP standard specification.  Note, however, that the
     security options described in [INTERNET:1] and [INTERNET:16]
     are obsolete.  Routers SHOULD IMPLEMENT the revised security
     option described in [INTERNET:5].

5.3.13.3 Stream Identifier Option

     This option is obsolete.  If the Stream Identifier option is
     present in a packet forwarded by the router, the option MUST be
     ignored and passed through unchanged.

5.3.13.4 Source Route Options

     A router MUST implement support for source route options in
     forwarded packets.  A router MAY implement a configuration
     option which, when enabled, causes all source-routed packets to
     be discarded.  However, such an option MUST NOT be enabled by
     default.
     DISCUSSION:
        The ability to source route datagrams through the Internet
        is important to various network diagnostic tools.  However,
        in a few rare cases, source routing may be used to bypass
        administrative and security controls within a network.
        Specifically, those cases where manipulation of routing
        tables is used to provide administrative separation in lieu
        of other methods such as packet filtering may be vulnerable
        through source routed packets.

5.3.13.5 Record Route Option

     Routers MUST support the Record Route option in forwarded
     packets.
     A router MAY provide a configuration option which, if enabled,
     will cause the router to ignore (i.e. pass through unchanged)
     Record Route options in forwarded packets.  If provided, such
     an option MUST default to enabling the record-route.  This
     option does not affect the processing of Record Route options
     in datagrams received by the router itself (in particular,
     Record Route options in ICMP echo requests will still be
     processed in accordance with Section [4.3.3.6]).
     DISCUSSION:
        There are some people who believe that Record Route is a
        security problem because it discloses information about the
        topology of the network.  Thus, this document allows it to
        be disabled.

5.3.13.6 Timestamp Option

     Routers MUST support the timestamp option in forwarded packets.
     A timestamp value MUST follow the rules given in Section
     [3.2.2.8] of [INTRO:2].
     If the flags field = 3 (timestamp and prespecified address),
     the router MUST add its timestamp if the next prespecified
     address matches any of the router's IP addresses.  It is not
     necessary that the prespecified address be either the address
     of the interface on which the packet arrived or the address of
     the interface over which it will be sent.
     IMPLEMENTATION:
        To maximize the utility of the timestamps contained in the
        timestamp option, it is suggested that the timestamp
        inserted be, as nearly as practical, the time at which the
        packet arrived at the router.  For datagrams originated by
        the router, the timestamp inserted should be, as nearly as
        practical, the time at which the datagram was passed to the
        network layer for transmission.
     A router MAY provide a configuration option which, if enabled,
     will cause the router to ignore (i.e. pass through unchanged)
     Timestamp options in forwarded datagrams when the flag word is
     set to zero (timestamps only) or one (timestamp and registering
     IP address).  If provided, such an option MUST default to off
     (that is, the router does not ignore the timestamp).  This
     option does not affect the processing of Timestamp options in
     datagrams received by the router itself (in particular, a
     router will insert timestamps into Timestamp options in
     datagrams received by the router, and Timestamp options in ICMP
     echo requests will still be processed in accordance with
     Section [4.3.3.6]).
     DISCUSSION:
        Like the Record Route option, the Timestamp option can
        reveal information about a network's topology.  Some people
        consider this to be a security concern.

TRANSPORT LAYER

A router is not required to implement any Transport Layer protocols except those required to support Application Layer protocols supported by the router. In practice, this means that most routers implement both the Transmission Control Protocol (TCP) and the User Datagram Protocol (UDP).

USER DATAGRAM PROTOCOL - UDP

The User Datagram Protocol (UDP) is specified in [TRANS:1].

A router which implements UDP MUST be compliant, and SHOULD be unconditionally compliant, with the requirements of section 4.1.3 of [INTRO:2], except that:

o This specification does not specify the interfaces between the

  various protocol layers.  Thus, a router need not comply with
  sections 4.1.3.2, 4.1.3.3, and 4.1.3.5 of [INTRO:2] (except of
  course where compliance is required for proper functioning of
  Application Layer protocols supported by the router).

o Contrary to section 4.1.3.4 of [INTRO:2], an application MUST NOT

  be able to disable to generation of UDP checksums.

DISCUSSION:

  Although a particular application protocol may require that UDP
  datagrams it receives must contain a UDP checksum, there is no
  general requirement that received UDP datagrams contain UDP
  checksums.  Of course, if a UDP checksum is present in a received
  datagram, the checksum must be verified and the datagram discarded
  if the checksum is incorrect.

TRANSMISSION CONTROL PROTOCOL - TCP

The Transmission Control Protocol (TCP) is specified in [TRANS:2].

A router which implements TCP MUST be compliant, and SHOULD be unconditionally compliant, with the requirements of section 4.2 of [INTRO:2], except that:

o This specification does not specify the interfaces between the

  various protocol layers.  Thus, a router need not comply with the
  following requirements of [INTRO:2] (except of course where
  compliance is required for proper functioning of Application Layer
  protocols supported by the router):
  Section 4.2.2.2:
       Passing a received PSH flag to the application layer is now
       OPTIONAL.
  Section 4.2.2.4:
       A TCP MUST inform the application layer asynchronously
       whenever it receives an Urgent pointer and there was
       previously no pending urgent data, or whenever the Urgent
       pointer advances in the data stream.  There MUST be a way for
       the application to learn how much urgent data remains to be
       read from the connection, or at least to determine whether or
       not more urgent data remains to be read.
  Section 4.2.3.5:
       An application MUST be able to set the value for R2 for a
       particular connection.  For example, an interactive
       application might set R2 to ``infinity, giving the user
       control over when to disconnect.
  Section 4.2.3.7:
       If an application on a multihomed host does not specify the
       local IP address when actively opening a TCP connection, then
       the TCP MUST ask the IP layer to select a local IP address
       before sending the (first) SYN.  See the function
       GET_SRCADDR() in Section 3.4.
  Section 4.2.3.8:
       An application MUST be able to specify a source route when it
       actively opens a TCP connection, and this MUST take
       precedence over a source route received in a datagram.

o For similar reasons, a router need not comply with any of the

  requirements of section 4.2.4 of [INTRO:2].

o The requirements of section 4.2.2.6 of [INTRO:2] are amended as

  follows: a router which implements the host portion of MTU
  discovery (discussed in Section [4.2.3.3] of this memo) uses 536
  as the default value of SendMSS only if the path MTU is unknown;
  if the path MTU is known, the default value for SendMSS is the
  path MTU - 40.

o The requirements of section 4.2.2.6 of [INTRO:2] are amended as

  follows: ICMP Destination Unreachable codes 11 and 12 are
  additional soft error conditions.  Therefore, these message MUST
  NOT cause TCP to abort a connection.

DISCUSSION:

  It should particularly be noted that a TCP implementation in a
  router must conform to the following requirements of [INTRO:2]:
  o  Providing a configurable TTL. [4.2.2.1]
  o  Providing an interface to configure keep-alive behavior, if
     keep-alives are used at all. [4.2.3.6]
  o  Providing an error reporting mechanism, and the ability to
     manage it.  [4.2.4.1]
  o  Specifying type of service. [4.2.4.2]
  The general paradigm applied is that if a particular interface is
  visible outside the router, then all requirements for the
  interface must be followed.  For example, if a router provides a
  telnet function, then it will be generating traffic, likely to be
  routed in the external networks.  Therefore, it must be able to
  set the type of service correctly or else the telnet traffic may
  not get through.

APPLICATION LAYER - ROUTING PROTOCOLS

INTRODUCTION

An Autonomous System (AS) is defined as a set of routers all belonging under the same authority and all subject to a consistent set of routing policies. Interior gateway protocols (IGPs) are used to distribute routing information inside of an AS (i.e. intra-AS routing). Exterior gateway protocols are used to exchange routing information between ASs (i.e. inter-AS routing).

Routing Security Considerations

  Routing is one of the few places where the Robustness Principle
  (be liberal in what you accept) does not apply.  Routers should be
  relatively suspicious in accepting routing data from other routing
  systems.
  A router SHOULD provide the ability to rank routing information
  sources from most trustworthy to least trustworthy and to accept
  routing information about any particular destination from the most
  trustworthy sources first.  This was implicit in the original
  core/stub autonomous system routing model using EGP and various
  interior routing protocols.  It is even more important with the
  demise of a central, trusted core.
  A router SHOULD provide a mechanism to filter out obviously
  invalid routes (such as those for net 127).
  Routers MUST NOT by default redistribute routing data they do not
  themselves use, trust or otherwise consider invalid.  In rare
  cases, it may be necessary to redistribute suspicious information,
  but this should only happen under direct intercession by some
  human agency.
  In general, routers must be at least a little paranoid about
  accepting routing data from anyone, and must be especially careful
  when they distribute routing information provided to them by
  another party.  See below for specific guidelines.
  Routers SHOULD IMPLEMENT peer-to-peer authentication for those
  routing protocols that support them.

Precedence

  Except where the specification for a particular routing protocol
  specifies otherwise, a router SHOULD set the IP Precedence value
  for IP datagrams carrying routing traffic it originates to 6
  (INTERNETWORK CONTROL).
  DISCUSSION:
     Routing traffic with VERY FEW exceptions should be the highest
     precedence traffic on any network.  If a system's routing
     traffic can't get through, chances are nothing else will.

INTERIOR GATEWAY PROTOCOLS

INTRODUCTION

  An Interior Gateway Protocol (IGP) is used to distribute routing
  information between the various routers in a particular AS.
  Independent of the algorithm used to implement a particular IGP,
  it should perform the following functions:
  (1)  Respond quickly to changes in the internal topology of an AS
  (2)  Provide a mechanism such that circuit flapping does not cause
       continuous routing updates
  (3)  Provide quick convergence to loop-free routing
  (4)  Utilize minimal bandwidth
  (5)  Provide equal cost routes to enable load-splitting
  (6)  Provide a means for authentication of routing updates
  Current IGPs used in the internet today are characterized as
  either being being based on a distance-vector or a link-state
  algorithm.
  Several IGPs are detailed in this section, including those most
  commonly used and some recently developed protocols which may be
  widely used in the future.  Numerous other protocols intended for
  use in intra-AS routing exist in the Internet community.
  A router which implements any routing protocol (other than static
  routes) MUST IMPLEMENT OSPF (see Section [7.2.2]) and MUST
  IMPLEMENT RIP (see Section [7.2.4]).  A router MAY implement
  additional IGPs.

OPEN SHORTEST PATH FIRST - OSPF

Introduction
     Shortest Path First (SPF) based routing protocols are a class
     of link-state algorithms which are based on the shortest-path
     algorithm of Dijkstra.  Although SPF based algorithms have been
     around since the inception of the ARPANet, it is only recently
     that they have achieved popularity both inside both the IP and
     the OSI communities.  In an SPF based system, each router
     obtains an exact replica of the entire topology database via a
     process known as flooding.  Flooding insures a reliable
     transfer of the information. Each individual router then runs
     the SPF algorithm on its database to build the IP routing
     table.  The OSPF routing protocol is an implementation of an
     SPF algorithm.  The current version, OSPF version 2, is
     specified in [ROUTE:1].  Note that RFC-1131, which describes
     OSPF version 1, is obsolete.
     Note that to comply with Section [8.3] of this memo, a router
     which implements OSPF MUST implement the OSPF MIB [MGT:14].
Specific Issues
     Virtual Links
          There is a minor error in the specification that can cause
          routing loops when all of the following conditions are
          simultaneously true:
          (1)  A virtual link is configured through a transit area,
          (2)  Two separate paths exist, each having the same
               endpoints, but one utilizing only non-virtual
               backbone links, and the other using links in the
               transit area, and
          (3)  The latter path is part of the (underlying physical
               representation of the) configured virtual link,
               routing loops may occur.
          To prevent this, an implementation of OSPF SHOULD invoke
          the calculation in Section 16.3 of [ROUTE:1] whenever any
          part of the path to the destination is a virtual link (the
          specification only says this is necessary when the first
          hop is a virtual link).
New Version of OSPF
     As of this writing (4/4/94) there is a new version of the OSPF
     specification that is winding its way through the Internet
     standardization process.  A prudent implementor will be aware
     of this and develop an implementation accordingly.
     The new version fixes several errors in the current
     specification [ROUTE:1].  For this reason, implementors and
     vendors ought to expect to upgrade to the new version
     relatively soon.  In particular, the following problems exist
     in [ROUTE:1] that the new version fixes:
     o  In [ROUTE:1], certain configurations of virtual links can
        lead to incorrect routing and/or routing loops. A fix for
        this is specified in the new specification.
     o  In [ROUTE:1], OSPF external routes to For example, a router
        cannot import into an OSPF domain external routes both for
        192.2.0.0, 255.255.0.0 and 192.2.0.0, 255.255.255.0.  Routes
        such as these may become common with the deployment of CIDR
        [INTERNET:15].  This has been addressed in the new OSPF
        specification.
     o  In [ROUTE:1], OSPF Network-LSAs originated before a router
        changes its OSPF Router ID can confuse the Dijkstra
        calculation if the router again becomes Designated Router
        for the network. This has been fixed.

INTERMEDIATE SYSTEM TO INTERMEDIATE SYSTEM - DUAL IS-IS

  The American National Standards Institute (ANSI) X3S3.3 committee
  has defined an intra-domain routing protocol.  This protocol is
  titled Intermediate System to Intermediate System Routeing
  Exchange Protocol.
  Its application to an IP network has been defined in [ROUTE:2],
  and is referred to as Dual IS-IS (or sometimes as Integrated IS-
  IS).  IS-IS is based on a link-state (SPF) routing algorithm and
  shares all the advantages for this class of protocols.

ROUTING INFORMATION PROTOCOL - RIP

Introduction
     RIP is specified in [ROUTE:3].  Although RIP is still quite
     important in the Internet, it is being replaced in
     sophisticated applications by more modern IGPs such as the ones
     described above.
     Another common use for RIP is as a router discovery protocol.
     Section [4.3.3.10] briefly touches upon this subject.
Protocol Walk-Through
     Dealing with changes in topology: [ROUTE:3], pp. 11
          An implementation of RIP MUST provide a means for timing
          out routes.  Since messages are occasionally lost,
          implementations MUST NOT invalidate a route based on a
          single missed update.
          Implementations MUST by default wait six times the update
          interval before invalidating a route.  A router MAY have
          configuration options to alter this value.
          DISCUSSION:
             It is important to routing stability that all routers
             in a RIP autonomous system use similar timeout value
             for invalidating routes, and therefore it is important
             that an implementation default to the timeout value
             specified in the RIP specification.  However, that
             timeout value is overly conservative in environments
             where packet loss is reasonably rare.  In such an
             environment, a network manager may wish to be able to
             decrease the timeout period in order to promote faster
             recovery from failures.
          IMPLEMENTATION:
             There is a very simple mechanism which a router may use
             to meet the requirement to invalidate routes promptly
             after they time out.  Whenever the router scans the
             routing table to see if any routes have timed out, it
             also notes the age of the least recently updated route
             which has not yet timed out.  Subtracting this age from
             the timeout period gives the amount of time until the
             router again needs to scan the table for timed out
             routes.
     Split Horizon: [ROUTE:3], pp. 14-15
          An implementation of RIP MUST implement split horizon, a
          scheme used for avoiding problems caused by including
          routes in updates sent to the router from which they were
          learned.
          An implementation of RIP SHOULD implement Split horizon
          with poisoned reverse, a variant of split horizon which
          includes routes learned from a router sent to that router,
          but sets their metric to infinity.  Because of the routing
          overhead which may be incurred by implementing split
          horizon with poisoned reverse, implementations MAY include
          an option to select whether poisoned reverse is in effect.
          An implementation SHOULD limit the period of time in which
          it sends reverse routes at an infinite metric.
          IMPLEMENTATION:
             Each of the following algorithms can be used to limit
             the period of time for which poisoned reverse is
             applied to a route.  The first algorithm is more
             complex but does a more complete job of limiting
             poisoned reverse to only those cases where it is
             necessary.
             The goal of both algorithms is to ensure that poison
             reverse is done for any destination whose route has
             changed in the last Route Lifetime (typically 180
             seconds), unless it can be sure that the previous route
             used the same output interface.  The Route Lifetime is
             used because that is the amount of time RIP will keep
             around an old route before declaring it stale.
             The time intervals (and derived variables) used in the
             following algorithms are as follows:
             Tu   The Update Timer; the number of seconds between
                  RIP updates.  This typically defaults to 30
                  seconds.
             Rl   The Route Lifetime, in seconds.  This is the
                  amount of time that a route is presumed to be
                  good, without requiring an update.  This typically
                  defaults to 180 seconds.
             Ul   The Update Loss; the number of consecutive updates
                  that have to be lost or fail to mention a route
                  before RIP deletes the route.  Ul is calculated to
                  be (Rl/Tu)+1.  The +1 is to account for the fact
                  that the first time the ifcounter is decremented
                  will be less than Tu seconds after it is
                  initialized.  Typically, Ul will be 7: (180/30)+1.
             In   The value to set ifcounter to when a destination
                  is newly learned.  This value is Ul-4, where the 4
                  is RIP's garbage collection timer/30
             The first algorithm is:
             - Associated with each destination is a counter, called
                the ifcounter below.  Poison reverse is done for any
                route whose destination's ifcounter is greater than
                zero.
             - After a regular (not triggered or in response to a
                request) update is sent, all of the non-zero
                ifcounters are decremented by one.
             - When a route to a destination is created, its
                ifcounter is set as follows:
                - If the new route is superseding a valid route, and
                   the old route used a different (logical) output
                   interface, then the ifcounter is set to Ul.
                - If the new route is superseding a stale route, and
                   the old route used a different (logical) output
                   interface, then the ifcounter is set to MAX(0, Ul
                   - INT(seconds that the route has been stale/Ut).
                - If there was no previous route to the destination,
                   the ifcounter is set to In.
                - Otherwise, the ifcounter is set to zero
             - RIP also maintains a timer, called the resettimer
                below.  Poison reverse is done on all routes
                whenever resettimer has not expired (regardless of
                the ifcounter values).
             - When RIP is started, restarted, reset, or otherwise
                has its routing table cleared, it sets the
                resettimer to go off in Rl seconds.
             The second algorithm is identical to the first except
             that:
             - The rules which set the ifcounter to non-zero values
                are changed to always set it to Rl/Tu, and
             - The resettimer is eliminated.
        Triggered updates: [ROUTE:3], pp. 15-16; pp. 29
             Triggered updates (also called flash updates) are a
             mechanism for immediately notifying a router's
             neighbors when the router adds or deletes routes or
             changes their metrics.  A router MUST send a triggered
             update when routes are deleted or their metrics are
             increased.  A router MAY send a triggered update when
             routes are added or their metrics decreased.
             Since triggered updates can cause excessive routing
             overhead, implementations MUST use the following
             mechanism to limit the frequency of triggered updates:
             (1)  When a router sends a triggered update, it sets a
                  timer to a random time between one and five
                  seconds in the future.  The router must not
                  generate additional triggered updates before this
                  timer expires.
             (2)  If the router would generate a triggered update
                  during this interval it sets a flag indicating
                  that a triggered update is desired.  The router
                  also logs the desired triggered update.
             (3)  When the triggered update timer expires, the
                  router checks the triggered update flag. If the
                  flag is set then the router sends a single
                  triggered update which includes all of the changes
                  that were logged.  The router then clears the flag
                  and, since a triggered update was sent, restarts
                  this algorithm.
             (4)  The flag is also cleared whenever a regular update
                  is sent.
             Triggered updates SHOULD include all routes that have
             changed since the most recent regular (non-triggered)
             update.  Triggered updates MUST NOT include routes that
             have not changed since the most recent regular update.
             DISCUSSION:
                Sending all routes, whether they have changed
                recently or not, is unacceptable in triggered
                updates because the tremendous size of many Internet
                routing tables could otherwise result in
                considerable bandwidth being wasted on triggered
                updates.
        Use of UDP: [ROUTE:3], pp. 18-19.
             RIP packets sent to an IP broadcast address SHOULD have
             their initial TTL set to one.
             Note that to comply with Section [6.1] of this memo, a
             router MUST use UDP checksums in RIP packets which it
             originates, MUST discard RIP packets received with
             invalid UDP checksums, but MUST not discard received
             RIP packets simply because they do not contain UDP
             checksums.
        Addressing Considerations: [ROUTE:3], pp. 22
             A RIP implementation SHOULD support host routes.  If it
             does not, it MUST (as described on page 27 of
             [ROUTE:3]) ignore host routes in received updates.  A
             router MAY log ignored hosts routes.
             The special address 0.0.0.0 is used to describe a
             default route. A default route is used as the route of
             last resort (i.e. when a route to the specific net does
             not exist in the routing table). The router MUST be
             able to create a RIP entry for the address 0.0.0.0.
        Input Processing - Response: [ROUTE:3], pp. 26
             When processing an update, the following validity
             checks MUST be performed:
             o  The response MUST be from UDP port 520.
             o  The source address MUST be on a directly connected
                subnet (or on a directly connected, non-subnetted
                network) to be considered valid.
             o  The source address MUST NOT be one of the router's
                addresses.
                DISCUSSION:
                   Some networks, media, and interfaces allow a
                   sending node to receive packets that it
                   broadcasts.  A router must not accept its own
                   packets as valid routing updates and process
                   them.  The last requirement prevents a router
                   from accepting its own routing updates and
                   processing them (on the assumption that they were
                   sent by some other router on the network).
             An implementation MUST NOT replace an existing route if
             the metric received is equal to the existing metric
             except in accordance with the following heuristic.
             An implementation MAY choose to implement the following
             heuristic to deal with the above situation. Normally,
             it is useless to change the route to a network from one
             router to another if both are advertised at the same
             metric. However, the route being advertised by one of
             the routers may be in the process of timing out.
             Instead of waiting for the route to timeout, the new
             route can be used after a specified amount of time has
             elapsed. If this heuristic is implemented, it MUST wait
             at least halfway to the expiration point before the new
             route is installed.
Specific Issues
     RIP Shutdown
          An implementation of RIP SHOULD provide for a graceful
          shutdown using the following steps:
          (1)  Input processing is terminated,
          (2)  Four updates are generated at random intervals of
               between two and four seconds, These updates contain
               all routes that were previously announced, but with
               some metric changes.  Routes that were being
               announced at a metric of infinity should continue to
               use this metric.  Routes that had been announced with
               a non-infinite metric should be announced with a
               metric of 15 (infinity - 1).
               DISCUSSION:
                  The metric used for the above really ought to be
                  16 (infinity); setting it to 15 is a kludge to
                  avoid breaking certain old hosts which wiretap the
                  RIP protocol.  Such a host will (erroneously)
                  abort a TCP connection if it tries to send a
                  datagram on the connection while the host has no
                  route to the destination (even if the period when
                  the host has no route lasts only a few seconds
                  while RIP chooses an alternate path to the
                  destination).
     RIP Split Horizon and Static Routes
          Split horizon SHOULD be applied to static routes by
          default.  An implementation SHOULD provide a way to
          specify, per static route, that split horizon should not
          be applied to this route.

GATEWAY TO GATEWAY PROTOCOL - GGP

  The Gateway to Gateway protocol is considered obsolete and SHOULD
  NOT be implemented.

EXTERIOR GATEWAY PROTOCOLS

INTRODUCTION

  Exterior Gateway Protocols are utilized for inter-Autonomous
  System routing to exchange reachability information for a set of
  networks internal to a particular autonomous system to a
  neighboring autonomous system.
  The area of inter-AS routing is a current topic of research inside
  the Internet Engineering Task Force.  The Exterior Gateway
  Protocol (EGP) described in Section [7.3.3] has traditionally been
  the inter-AS protocol of choice.  The Border Gateway Protocol
  (BGP) eliminates many of the restrictions and limitations of EGP,
  and is therefore growing rapidly in popularity.  A router is not
  required to implement any inter-AS routing protocol.  However, if
  a router does implement EGP it also MUST IMPLEMENT BGP.
  Although it was not designed as an exterior gateway protocol, RIP
  (described in Section [7.2.4]) is sometimes used for inter-AS
  routing.

BORDER GATEWAY PROTOCOL - BGP

Introduction
     The Border Gateway Protocol (BGP) is an inter-AS routing
     protocol which exchanges network reachability information with
     other BGP speakers. The information for a network includes the
     complete list of ASs that traffic must transit to reach that
     network. This information can then be used to insure loop-free
     paths.  This information is sufficient to construct a graph of
     AS connectivity from which routing loops may be pruned and some
     policy decisions at the AS level may be enforced.
     BGP is defined by [ROUTE:4].  [ROUTE:5] specifies the proper
     usage of BGP in the Internet, and provides some useful
     implementation hints and guidelines.  [ROUTE:12] and [ROUTE:13]
     provide additional useful information.
     To comply with Section [8.3] of this memo, a router which
     implements BGP MUST also implement the BGP MIB [MGT:15].
     To characterize the set of policy decisions that can be
     enforced using BGP, one must focus on the rule that an AS
     advertises to its neighbor ASs only those routes that it itself
     uses.  This rule reflects the hop-by-hop routing paradigm
     generally used throughout the current Internet.  Note that some
     policies cannot be supported by the hop-by-hop routing paradigm
     and thus require techniques such as source routing to enforce.
     For example, BGP does not enable one AS to send traffic to a
     neighbor AS intending that that traffic take a different route
     from that taken by traffic originating in the neighbor AS.  On
     the other hand, BGP can support any policy conforming to the
     hop-by-hop routing paradigm.
     Implementors of BGP are strongly encouraged to follow the
     recommendations outlined in Section 6 of [ROUTE:5].
Protocol Walk-through
     While BGP provides support for quite complex routing policies
     (as an example see Section 4.2 in [ROUTE:5]), it is not
     required for all BGP implementors to support such policies.  At
     a minimum, however, a BGP implementation:
     (1)  SHOULD allow an AS to control announcements of the BGP
          learned routes to adjacent AS's. Implementations SHOULD
          support such control with at least the granularity of a
          single network. Implementations SHOULD also support such
          control with the granularity of an autonomous system,
          where the autonomous system may be either the autonomous
          system that originated the route, or the autonomous system
          that advertised the route to the local system (adjacent
          autonomous system).
     (2)  SHOULD allow an AS to prefer a particular path to a
          destination (when more than one path is available).  Such
          function SHOULD be implemented by allowing system
          administrator to assign weights to Autonomous Systems, and
          making route selection process to select a route with the
          lowest weight (where weight of a route is defined as a sum
          of weights of all AS's in the AS_PATH path attribute
          associated with that route).
     (3)  SHOULD allow an AS to ignore routes with certain AS's in
          the AS_PATH path attribute. Such function can be
          implemented by using technique outlined in (2), and by
          assigning infinity as weights for such AS's. The route
          selection process must ignore routes that have weight
          equal to infinity.

EXTERIOR GATEWAY PROTOCOL - EGP

Introduction
     The Exterior Gateway Protocol (EGP) specifies an EGP which is
     used to exchange reachability information between routers of
     the same or differing autonomous systems. EGP is not considered
     a routing protocol since there is no standard interpretation
     (i.e. metric) for the distance fields in the EGP update
     message, so distances are comparable only among routers of the
     same AS.  It is however designed to provide high-quality
     reachability information, both about neighbor routers and about
     routes to non-neighbor routers.
     EGP is defined by [ROUTE:6].  An implementor almost certainly
     wants to read [ROUTE:7] and [ROUTE:8] as well, for they contain
     useful explanations and background material.
     DISCUSSION:
        The present EGP specification has serious limitations, most
        importantly a restriction which limits routers to
        advertising only those networks which are reachable from
        within the router's autonomous system.  This restriction
        against propagating third party EGP information is to
        prevent long-lived routing loops.  This effectively limits
        EGP to a two-level hierarchy.
        RFC-975 is not a part of the EGP specification, and should
        be ignored.
Protocol Walk-through
     Indirect Neighbors: RFC-888, pp. 26
        An implementation of EGP MUST include indirect neighbor
        support.
     Polling Intervals: RFC-904, pp. 10
        The interval between Hello command retransmissions and the
        interval between Poll retransmissions SHOULD be configurable
        but there MUST be a minimum value defined.
        The interval at which an implementation will respond to
        Hello commands and Poll commands SHOULD be configurable but
        there MUST be a minimum value defined.
     Network Reachability: RFC-904, pp. 15
        An implementation MUST default to not providing the external
        list of routers in other autonomous systems; only the
        internal list of routers together with the nets which are
        reachable via those routers should be included in an Update
        Response/Indication packet.  However, an implementation MAY
        elect to provide a configuration option enabling the
        external list to be provided.  An implementation MUST NOT
        include in the external list routers which were learned via
        the external list provided by a router in another autonomous
        system. An implementation MUST NOT send a network back to
        the autonomous system from which it is learned, i.e. it MUST
        do split-horizon on an autonomous system level.
        If more than 255 internal or 255 external routers need to be
        specified in a Network Reachability update, the networks
        reachable from routers that can not be listed MUST be merged
        into the list for one of the listed routers.  Which of the
        listed routers is chosen for this purpose SHOULD be user
        configurable, but SHOULD default to the source address of
        the EGP update being generated.
        An EGP update contains a series of blocks of network
        numbers, where each block contains a list of network numbers
        reachable at a particular distance via a particular router.
        If more than 255 networks are reachable at a particular
        distance via a particular router, they are split into
        multiple blocks (all of which have the same distance).
        Similarly, if more than 255 blocks are required to list the
        networks reachable via a particular router, the router's
        address is listed as many times as necessary to include all
        of the blocks in the update.
     Unsolicited Updates: RFC-904, pp. 16
        If a network is shared with the peer, an implementation MUST
        send an unsolicited update upon entry to the Up state
        assuming that the source network is the shared network.
     Neighbor Reachability: RFC-904, pp. 6, 13-15
        The table on page 6 which describes the values of j and k
        (the neighbor up and down thresholds) is incorrect.  It is
        reproduced correctly here:
           Name    Active  Passive Description
           -----------------------------------------------
            j         3       1    neighbor-up threshold
            k         1       0    neighbor-down threshold
        The value for k in passive mode also specified incorrectly
        in RFC-904, pp. 14 The values in parenthesis should read:
           (j = 1, k = 0, and T3/T1 = 4)
        As an optimization, an implementation can refrain from
        sending a Hello command when a Poll is due.  If an
        implementation does so, it SHOULD provide a user
        configurable option to disable this optimization.
     Abort timer: RFC-904, pp. 6, 12, 13
        An EGP implementation MUST include support for the abort
        timer (as documented in section 4.1.4 of RFC-904).  An
        implementation SHOULD use the abort timer in the Idle state
        to automatically issue a Start event to restart the protocol
        machine.  Recommended values are P4 for a critical error
        (Administratively prohibited, Protocol Violation and
        Parameter Problem) and P5 for all others.  The abort timer
        SHOULD NOT be started when a Stop event was manually
        initiated (such as via a network management protocol).
     Cease command received in Idle state: RFC-904, pp. 13
        When the EGP state machine is in the Idle state, it MUST
        reply to Cease commands with a Cease-ack response.
     Hello Polling Mode: RFC-904, pp. 11
        An EGP implementation MUST include support for both active
        and passive polling modes.
     Neighbor Acquisition Messages: RFC-904, pp. 18
        As noted the Hello and Poll Intervals should only be present
        in Request and Confirm messages.  Therefore the length of an
        EGP Neighbor Acquisition Message is 14 bytes for a Request
        or Confirm message and 10 bytes for a Refuse, Cease or
        Cease-ack message.  Implementations MUST NOT send 14 bytes
        for Refuse, Cease or Cease-ack messages but MUST allow for
        implementations that send 14 bytes for these messages.
     Sequence Numbers: RFC-904, pp. 10
        Response or indication packets received with a sequence
        number not equal to S MUST be discarded.  The send sequence
        number S MUST be incremented just before the time a Poll
        command is sent and at no other times.

INTER-AS ROUTING WITHOUT AN EXTERIOR PROTOCOL

  It is possible to exchange routing information between two
  autonomous systems or routing domains without using a standard
  exterior routing protocol between two separate, standard interior
  routing protocols.  The most common way of doing this is to run
  both interior protocols independently in one of the border routers
  with an exchange of route information between the two processes.
  As with the exchange of information from an EGP to an IGP, without
  appropriate controls these exchanges of routing information
  between two IGPs in a single router are subject to creation of
  routing loops.

STATIC ROUTING

Static routing provides a means of explicitly defining the next hop from a router for a particular destination. A router SHOULD provide a means for defining a static route to a destination, where the destination is defined by an address and an address mask. The mechanism SHOULD also allow for a metric to be specified for each static route.

A router which supports a dynamic routing protocol MUST allow static routes to be defined with any metric valid for the routing protocol used. The router MUST provide the ability for the user to specify a list of static routes which may or may not be propagated via the routing protocol. In addition, a router SHOULD support the following additional information if it supports a routing protocol that could make use of the information. They are:

o TOS,

o Subnet mask, or

o A metric specific to a given routing protocol that can import the

  route.

DISCUSSION:

  We intend that one needs to support only the things useful to the
  given routing protocol.  The need for TOS should not require the
  vendor to implement the other parts if they are not used.

Whether a router prefers a static route over a dynamic route (or vice versa) or whether the associated metrics are used to choose between conflicting static and dynamic routes SHOULD be configurable for each static route.

A router MUST allow a metric to be assigned to a static route for each routing domain that it supports. Each such metric MUST be explicitly assigned to a specific routing domain. For example:

    route 36.0.0.0 255.0.0.0 via 192.19.200.3 rip metric 3
    route 36.21.0.0 255.255.0.0 via 192.19.200.4 ospf inter-area
    metric 27
    route 36.22.0.0 255.255.0.0 via 192.19.200.5 egp 123 metric 99
    route 36.23.0.0 255.255.0.0 via 192.19.200.6 igrp 47 metric 1 2
    3 4 5

DISCUSSION:

  It has been suggested that, ideally, static routes should have
  preference values rather than metrics (since metrics can only be
  compared with metrics of other routes in the same routing domain,
  the metric of a static route could only be compared with metrics
  of other static routes).  This is contrary to some current
  implementations, where static routes really do have metrics, and
  those metrics are used to determine whether a particular dynamic
  route overrides the static route to the same destination.  Thus,
  this document uses the term metric rather than preference.
  This technique essentially makes the static route into a RIP
  route, or an OSPF route (or whatever, depending on the domain of
  the metric).  Thus, the route lookup algorithm of that domain
  applies.  However, this is NOT route leaking, in that coercing a
  static route into a dynamic routing domain does not authorize the
  router to redistribute the route into the dynamic routing domain.
  For static routes not put into a specific routing domain, the
  route lookup algorithm is:
  (1)  Basic match
  (2)  Longest match
  (3)  Weak TOS (if TOS supported)
  (4)  Best metric (where metric are implementation-defined)
  The last step may not be necessary, but it's useful in the case
  where you want to have a primary static route over one interface
  and a secondary static route over an alternate interface, with
  failover to the alternate path if the interface for the primary
  route fails.

FILTERING OF ROUTING INFORMATION

Each router within a network makes forwarding decisions based upon information contained within its forwarding database. In a simple network the contents of the database may be statically configured. As the network grows more complex, the need for dynamic updating of the forwarding database becomes critical to the efficient operation of the network.

If the data flow through a network is to be as efficient as possible, it is necessary to provide a mechanism for controlling the propagation of the information a router uses to build its forwarding database. This control takes the form of choosing which sources of routing information should be trusted and selecting which pieces of the information to believe. The resulting forwarding database is a filtered version of the available routing information.

In addition to efficiency, controlling the propagation of routing information can reduce instability by preventing the spread of incorrect or bad routing information.

In some cases local policy may require that complete routing information not be widely propagated.

These filtering requirements apply only to non-SPF-based protocols (and therefore not at all to routers which don't implement any distance vector protocols).

Route Validation

  A router SHOULD log as an error any routing update advertising a
  route to network zero, subnet zero, or subnet -1, unless the
  routing protocol from which the update was received uses those
  values to encode special routes (such as default routes).

Basic Route Filtering

  Filtering of routing information allows control of paths used by a
  router to forward packets it receives.  A router should be
  selective in which sources of routing information it listens to
  and what routes it believes.  Therefore, a router MUST provide the
  ability to specify:
  o  On which logical interfaces routing information will be
     accepted and which routes will be accepted from each logical
     interface.
  o  Whether all routes or only a default route is advertised on a
     logical interface.
  Some routing protocols do not recognize logical interfaces as a
  source of routing information.  In such cases the router MUST
  provide the ability to specify
  o  from which other routers routing information will be accepted.
  For example, assume a router connecting one or more leaf networks
  to the main portion or backbone of a larger network.  Since each
  of the leaf networks has only one path in and out, the router can
  simply send a default route to them.  It advertises the leaf
  networks to the main network.

Advanced Route Filtering

  As the topology of a network grows more complex, the need for more
  complex route filtering arises.  Therefore, a router SHOULD
  provide the ability to specify independently for each routing
  protocol:
  o  Which logical interfaces or routers routing information
     (routes) will be accepted from and which routes will be
     believed from each other router or logical interface,
  o  Which routes will be sent via which logical interface(s), and
  o  Which routers routing information will be sent to, if this is
     supported by the routing protocol in use.
  In many situations it is desirable to assign a reliability
  ordering to routing information received from another router
  instead of the simple believe or don't believe choice listed in
  the first bullet above.  A router MAY provide the ability to
  specify:
  o  A reliability or preference to be assigned to each route
     received.  A route with higher reliability will be chosen over
     one with lower reliability regardless of the routing metric
     associated with each route.
  If a router supports assignment of preferences, the router MUST
  NOT propagate any routes it does not prefer as first party
  information.  If the routing protocol being used to propagate the
  routes does not support distinguishing between first and third
  party information, the router MUST NOT propagate any routes it
  does not prefer.
  DISCUSSION:
     For example, assume a router receives a route to network C from
     router R and a route to the same network from router S.  If
     router R is considered more reliable than router S traffic
     destined for network C will be forwarded to router R regardless
     of the route received from router S.
  Routing information for routes which the router does not use
  (router S in the above example) MUST NOT be passed to any other
  router.

INTER-ROUTING-PROTOCOL INFORMATION EXCHANGE

Routers MUST be able to exchange routing information between separate IP interior routing protocols, if independent IP routing processes can run in the same router. Routers MUST provide some mechanism for avoiding routing loops when routers are configured for bi-directional exchange of routing information between two separate interior routing processes. Routers MUST provide some priority mechanism for choosing routes from among independent routing processes. Routers SHOULD provide administrative control of IGP-IGP exchange when used across administrative boundaries.

Routers SHOULD provide some mechanism for translating or transforming metrics on a per network basis. Routers (or routing protocols) MAY allow for global preference of exterior routes imported into an IGP.

DISCUSSION:

  Different IGPs use different metrics, requiring some translation
  technique when introducing information from one protocol into
  another protocol with a different form of metric.  Some IGPs can
  run multiple instances within the same router or set of routers.
  In this case metric information can be preserved exactly or
  translated.
  There are at least two techniques for translation between
  different routing processes.  The static (or reachability)
  approach uses the existence of a route advertisement in one IGP to
  generate a route advertisement in the other IGP with a given
  metric.  The translation or tabular approach uses the metric in
  one IGP to create a metric in the other IGP through use of either
  a function (such as adding a constant) or a table lookup.
  Bi-directional exchange of routing information is dangerous
  without control mechanisms to limit feedback.  This is the same
  problem that distance vector routing protocols must address with
  the split horizon technique and that EGP addresses with the
  third-party rule.  Routing loops can be avoided explicitly through
  use of tables or lists of permitted/denied routes or implicitly
  through use of a split horizon rule, a no-third-party rule, or a
  route tagging mechanism.  Vendors are encouraged to use implicit
  techniques where possible to make administration easier for
  network operators.

APPLICATION LAYER - NETWORK MANAGEMENT PROTOCOLS

Note that this chapter supersedes any requirements stated in section 6.3 of [INTRO:3].

The Simple Network Management Protocol - SNMP

SNMP Protocol Elements

  Routers MUST be manageable by SNMP [MGT:3].  The SNMP MUST operate
  using UDP/IP as its transport and network protocols.  Others MAY
  be supported (e.g., see [MGT:25, MGT:26, MGT:27, and MGT:28]).
  SNMP management operations MUST operate as if the SNMP was
  implemented on the router itself. Specifically, management
  operations MUST be effected by sending SNMP management requests to
  any of the IP addresses assigned to any of the router's
  interfaces. The actual management operation may be performed
  either by the router or by a proxy for the router.
  DISCUSSION:
     This wording is intended to allow management either by proxy,
     where the proxy device responds to SNMP packets which have one
     of the router's IP addresses in the packets destination address
     field, or the SNMP is implemented directly in the router itself
     and receives packets and responds to them in the proper manner.
     It is important that management operations can be sent to one
     of the router's IP Addresses.  In diagnosing network problems
     the only thing identifying the router that is available may be
     one of the router's IP address; obtained perhaps by looking
     through another router's routing table.
  All SNMP operations (get, get-next, get-response, set, and trap)
  MUST be implemented.
  Routers MUST provide a mechanism for rate-limiting the generation
  of SNMP trap messages.  Routers MAY provide this mechanism via the
  algorithms for asynchronous alert management described in [MGT:5].
  DISCUSSION:
     Although there is general agreement about the need to rate-
     limit traps, there is not yet consensus on how this is best
     achieved.  The reference cited is considered experimental.

Community Table

For the purposes of this specification, we assume that there is an abstract `community table' in the router. This table contains several entries, each entry for a specific community and containing the parameters necessary to completely define the attributes of that community. The actual implementation method of the abstract community table is, of course, implementation specific.

A router's community table MUST allow for at least one entry and SHOULD allow for at least two entries.

DISCUSSION:

  A community table with zero capacity is useless.  It means that
  the router will not recognize any communities and, therefore, all
  SNMP operations will be rejected.
  Therefore, one entry is the minimal useful size of the table.
  Having two entries allows one entry to be limited to read-only
  access while the other would have write capabilities.

Routers MUST allow the user to manually (i.e., without using SNMP) examine, add, delete and change entries in the SNMP community table. The user MUST be able to set the community name. The user MUST be able to configure communities as read-only (i.e., they do not allow SETs) or read-write (i.e., they do allow SETs).

The user MUST be able to define at least one IP address to which traps are sent for each community. These addresses MUST be definable on a per-community basis. Traps MUST be enablable or disablable on a per-community basis.

A router SHOULD provide the ability to specify a list of valid network managers for any particular community. If enabled, a router MUST validate the source address of the SNMP datagram against the list and MUST discard the datagram if its address does not appear. If the datagram is discarded the router MUST take all actions appropriate to an SNMP authentication failure.

DISCUSSION:

  This is a rather limited authentication system, but coupled with
  various forms of packet filtering may provide some small measure
  of increased security.

The community table MUST be saved in non-volatile storage.

The initial state of the community table SHOULD contain one entry,

with the community name string public and read-only access. The default state of this entry MUST NOT send traps. If it is implemented, then this entry MUST remain in the community table until the administrator changes it or deletes it.

DISCUSSION:

  By default, traps are not sent to this community.  Trap PDUs are
  sent to unicast IP addresses. This address must be configured into
  the router in some manner. Before the configuration occurs, there
  is no such address, so to whom should the trap be sent? Therefore
  trap sending to the public community defaults to be disabled. This
  can, of course, be changed by an administrative operation once the
  router is operational.

Standard MIBS

All MIBS relevant to a router's configuration are to be implemented. To wit:

o The System, Interface, IP, ICMP, and UDP groups of MIB-II [MGT:2]

  MUST be implemented.

o The Interface Extensions MIB [MGT:18] MUST be implemented.

o The IP Forwarding Table MIB [MGT:20] MUST be implemented.

o If the router implements TCP (e.g. for Telnet) then the TCP group

  of MIB-II [MGT:2] MUST be implemented.

o If the router implements EGP then the EGP group of MIB-II [MGT:2]

  MUST be implemented.

o If the router supports OSPF then the OSPF MIB [MGT:14] MUST be

  implemented.

o If the router supports BGP then the BGP MIB [MGT:15] MUST be

  implemented.

o If the router has Ethernet, 802.3, or StarLan interfaces then the

  Ethernet-Like MIB [MGT:6] MUST be implemented.

o If the router has 802.4 interfaces then the 802.4 MIB [MGT:7] MAY

  be implemented.

o If the router has 802.5 interfaces then the 802.5 MIB [MGT:8] MUST

  be implemented.

o If the router has FDDI interfaces that implement ANSI SMT 7.3 then

  the FDDI MIB [MGT:9] MUST be implemented.

o If the router has FDDI interfaces that implement ANSI SMT 6.2 then

  the FDDI MIB [MGT:29] MUST be implemented.

o If the router has RS-232 interfaces then the RS-232 [MGT:10] MIB

  MUST be implemented.

o If the router has T1/DS1 interfaces then the T1/DS1 MIB [MGT:16]

  MUST be implemented.

o If the router has T3/DS3 interfaces then the T3/DS3 MIB [MGT:17]

  MUST be implemented.

o If the router has SMDS interfaces then the SMDS Interface Protocol

  MIB [MGT:19] MUST be implemented.

o If the router supports PPP over any of its interfaces then the PPP

  MIBs [MGT:11], [MGT:12], and [MGT:13] MUST be implemented.

o If the router supports RIP Version 2 then the RIP Version 2 MIB

  [MGT:21] MUST be implemented.

o If the router supports X.25 over any of its interfaces then the

  X.25 MIBs [MGT:22, MGT:23 and MGT:24] MUST be implemented.

Vendor Specific MIBS

The Internet Standard and Experimental MIBs do not cover the entire range of statistical, state, configuration and control information that may be available in a network element. This information is, never the less, extremely useful. Vendors of routers (and other network devices) generally have developed MIB extensions that cover this information. These MIB extensions are called Vendor Specific MIBs.

The Vendor Specific MIB for the router MUST provide access to all statistical, state, configuration, and control information that is not available through the Standard and Experimental MIBs that have been implemented. This information MUST be available for both monitoring and control operations.

DISCUSSION:

  The intent of this requirement is to provide the ability to do
  anything on the router via SNMP that can be done via a console.  A
  certain minimal amount of configuration is necessary before SNMP
  can operate (e.g., the router must have an IP address). This
  initial configuration can not be done via SNMP. However, once the
  initial configuration is done, full capabilities ought to be
  available via network management.

The vendor SHOULD make available the specifications for all Vendor Specific MIB variables. These specifications MUST conform to the SMI [MGT:1] and the descriptions MUST be in the form specified in [MGT:4].

DISCUSSION:

  Making the Vendor Specific MIB available to the user is necessary.
  Without this information the users would not be able to configure
  their network management systems to be able to access the Vendor
  Specific parameters.  These parameters would then be useless.
  The format of the MIB specification is also specified.  Parsers
  which read MIB specifications and generate the needed tables for
  the network management station are available.  These parsers
  generally understand only the standard MIB specification format.

Saving Changes

Parameters altered by SNMP MAY be saved to non-volatile storage.

DISCUSSION:

  Reasons why this requirement is a MAY:
  o  The exact physical nature of non-volatile storage is not
     specified in this document.  Hence, parameters may be saved in
     NVRAM/EEPROM, local floppy or hard disk, or in some TFTP file
     server or BOOTP server, etc. Suppose that that this information
     is in a file that is retrieved via TFTP. In that case, a change
     made to a configuration parameter on the router would need to
     be propagated back to the file server holding the configuration
     file.  Alternatively, the SNMP operation would need to be
     directed to the file server, and then the change somehow
     propagated to the router.  The answer to this problem does not
     seem obvious.
     This also places more requirements on the host holding the
     configuration information than just having an available tftp
     server, so much more that its probably unsafe for a vendor to
     assume that any potential customer will have a suitable host
     available.
  o  The timing of committing changed parameters to non-volatile
     storage is still an issue for debate. Some prefer to commit all
     changes immediately. Others prefer to commit changes to non-
     volatile storage only upon an explicit command.

APPLICATION LAYER - MISCELLANEOUS PROTOCOLS

For all additional application protocols that a router implements, the router MUST be compliant and SHOULD be unconditionally compliant with the relevant requirements of [INTRO:3].

BOOTP

Introduction

  The Bootstrap Protocol (BOOTP) is a UDP/IP-based protocol which
  allows a booting host to configure itself dynamically and without
  user supervision.  BOOTP provides a means to notify a host of its
  assigned IP address, the IP address of a boot server host, and the
  name of a file to be loaded into memory and executed ([APPL:1]).
  Other configuration information such as the local subnet mask, the
  local time offset, the addresses of default routers, and the
  addresses of various Internet servers can also be communicated to
  a host using BOOTP ([APPL:2]).

BOOTP Relay Agents

  In many cases, BOOTP clients and their associated BOOTP server(s)
  do not reside on the same IP network or subnet.  In such cases, a
  third-party agent is required to transfer BOOTP messages between
  clients and servers.  Such an agent was originally referred to as
  a BOOTP forwarding agent.  However, in order to avoid confusion
  with the IP forwarding function of a router, the name BOOTP relay
  agent has been adopted instead.
  DISCUSSION:
     A BOOTP relay agent performs a task which is distinct from a
     router's normal IP forwarding function.  While a router
     normally switches IP datagrams between networks more-or-less
     transparently, a BOOTP relay agent may more properly be thought
     to receive BOOTP messages as a final destination and then
     generate new BOOTP messages as a result.  One should resist the
     notion of simply forwarding a BOOTP message straight through
     like a regular packet.
  This relay-agent functionality is most conveniently located in the
  routers which interconnect the clients and servers (although it
  may alternatively be located in a host which is directly connected
  to the client subnet).
  A router MAY provide BOOTP relay-agent capability.  If it does, it
  MUST conform to the specifications in [APPL:3].
  Section [5.2.3] discussed the circumstances under which a packet
  is delivered locally (to the router).  All locally delivered UDP
  messages whose UDP destination port number is BOOTPS (67) are
  considered for special processing by the router's logical BOOTP
  relay agent.
  Sections [4.2.2.11] and [5.3.7] discussed invalid IP source
  addresses.  According to these rules, a router must not forward
  any received datagram whose IP source address is 0.0.0.0.
  However, routers which support a BOOTP relay agent MUST accept for
  local delivery to the relay agent BOOTREQUEST messages whose IP
  source address is 0.0.0.0.

10. OPERATIONS AND MAINTENANCE

This chapter supersedes any requirements stated in section 6.2 of [INTRO:3].

Facilities to support operation and maintenance (O&M) activities form an essential part of any router implementation. Although these functions do not seem to relate directly to interoperability, they are essential to the network manager who must make the router interoperate and must track down problems when it doesn't. This chapter also includes some discussion of router initialization and of facilities to assist network managers in securing and accounting for their networks.

10.1 Introduction

The following kinds of activities are included under router O&M:

o Diagnosing hardware problems in the router's processor, in its

  network interfaces, or in its connected networks, modems, or
  communication lines.

o Installing new hardware

o Installing new software.

o Restarting or rebooting the router after a crash.

o Configuring (or reconfiguring) the router.

o Detecting and diagnosing Internet problems such as congestion,

  routing loops, bad IP addresses, black holes, packet avalanches,
  and misbehaved hosts.

o Changing network topology, either temporarily (e.g., to bypass a

  communication line problem) or permanently.

o Monitoring the status and performance of the routers and the

  connected networks.

o Collecting traffic statistics for use in (Inter-)network planning.

o Coordinating the above activities with appropriate vendors and

  telecommunications specialists.

Routers and their connected communication lines are often operated as a system by a centralized O&M organization. This organization may maintain a (Inter-)network operation center, or NOC, to carry out its

O&M functions. It is essential that routers support remote control and monitoring from such a NOC through an Internet path, since routers might not be connected to the same network as their NOC. Since a network failure may temporarily preclude network access, many NOCs insist that routers be accessible for network management via an alternative means, often dialup modems attached to console ports on the routers.

Since an IP packet traversing an internet will often use routers under the control of more than one NOC, Internet problem diagnosis will often involve cooperation of personnel of more than one NOC. In some cases, the same router may need to be monitored by more than one NOC, but only if necessary, because excessive monitoring could impact a router's performance.

The tools available for monitoring at a NOC may cover a wide range of sophistication. Current implementations include multi-window, dynamic displays of the entire router system. The use of AI techniques for automatic problem diagnosis is proposed for the future.

Router O&M facilities discussed here are only a part of the large and difficult problem of Internet management. These problems encompass not only multiple management organizations, but also multiple protocol layers. For example, at the current stage of evolution of the Internet architecture, there is a strong coupling between host TCP implementations and eventual IP-level congestion in the router system [OPER:1]. Therefore, diagnosis of congestion problems will sometimes require the monitoring of TCP statistics in hosts. There are currently a number of R&D efforts in progress in the area of Internet management and more specifically router O&M. These R&D efforts have already produced standards for router O&M. This is also an area in which vendor creativity can make a significant contribution.

10.2 Router Initialization

10.2.1 Minimum Router Configuration

  There exists a minimum set of conditions that must be satisfied
  before a router may forward packets.  A router MUST NOT enable
  forwarding on any physical interface unless either:
  (1)  The router knows the IP address and associated subnet mask of
       at least one logical interface associated with that physical
       interface, or
  (2)  The router knows that the interface is an unnumbered
       interface and also knows its router-id.
  These parameters MUST be explicitly configured:
  o  A router MUST NOT use factory-configured default values for its
     IP addresses, subnet masks, or router-id, and
  o  A router MUST NOT assume that an unconfigured interface is an
     unnumbered interface.
  DISCUSSION:
     There have been instances in which routers have been shipped
     with vendor-installed default addresses for interfaces.  In a
     few cases, this has resulted in routers advertising these
     default addresses into active networks.

10.2.2 Address and Address Mask Initialization

  A router MUST allow its IP addresses and their subnet masks to be
  statically configured and saved in permanent storage.
  A router MAY obtain its IP addresses and their corresponding
  subnet masks dynamically as a side effect of the system
  initialization process (see Section 10.2.3]);
  If the dynamic method is provided, the choice of method to be used
  in a particular router MUST be configurable.
  As was described in Section [4.2.2.11], IP addresses are not
  permitted to have the value 0 or -1 for any of the <Host-number>,
  <Network-number>, or <Subnet-number> fields.  Therefore, a router
  SHOULD NOT allow an IP address or subnet mask to be set to a value
  which would make any of the the three fields above have the value
  zero or -1.
  DISCUSSION:
     It is possible using variable length subnet masks to create
     situations in which routing is ambiguous (i.e., two routes with
     different but equally-specific subnet masks match a particular
     destination address).  We suspect that a router could, when
     setting a subnet mask, check whether the mask would cause
     routing to be ambiguous, and that implementors might be able to
     decrease their customer support costs by having routers
     prohibit or log such erroneous configurations.  However, at
     this time we do not require routers to make such checks because
     we know of no published method for accurately making this
     check.
  A router SHOULD make the following checks on any subnet mask it
  installs:
  o  The mask is not all 1-bits.
  o  The bits which correspond to the network number part of the
     address are all set to 1.
  DISCUSSION:
     The masks associated with routes are also sometimes called
     subnet masks, this test should not be applied to them.

10.2.3 Network Booting using BOOTP and TFTP

  There has been a lot of discussion on how routers can and should
  be booted from the network.  In general, these discussions have
  centered around BOOTP and TFTP.  Currently, there are routers that
  boot with TFTP from the network.  There is no reason that BOOTP
  could not be used for locating the server that the boot image
  should be loaded from.
  In general, BOOTP is a protocol used to boot end systems, and
  requires some stretching to accommodate its use with routers.  If
  a router is using BOOTP to locate the current boot host, it should
  send a BOOTP Request with its hardware address for its first
  interface, or, if it has been previously configured otherwise,
  with either another interface's hardware address, or another
  number to put in the hardware address field of the BOOTP packet.
  This is to allow routers without hardware addresses (like sync
  line only routers) to use BOOTP for bootload discovery.  TFTP can
  then be used to retrieve the image found in the BOOTP Reply.  If
  there are no configured interfaces or numbers to use, a router MAY
  cycle through the interface hardware addresses it has until a
  match is found by the BOOTP server.
  A router SHOULD IMPLEMENT the ability to store parameters learned
  via BOOTP into local stable storage.  A router MAY implement the
  ability to store a system image loaded over the network into local
  stable storage.
  A router MAY have a facility to allow a remote user to request
  that the router get a new boot image.  Differentiation should be
  made between getting the new boot image from one of three
  locations: the one included in the request, from the last boot
  image server, and using BOOTP to locate a server.

10.3 Operation and Maintenance

10.3.1 Introduction

  There is a range of possible models for performing O&M functions
  on a router.  At one extreme is the local-only model, under which
  the O&M functions can only be executed locally (e.g., from a
  terminal plugged into the router machine).  At the other extreme,
  the fully-remote model allows only an absolute minimum of
  functions to be performed locally (e.g., forcing a boot), with
  most O&M being done remotely from the NOC.  There are intermediate
  models, such as one in which NOC personnel can log into the router
  as a host, using the Telnet protocol, to perform functions which
  can also be invoked locally.  The local-only model may be adequate
  in a few router installations, but in general remote operation
  from a NOC will be required, and therefore remote O&M provisions
  are required for most routers.
  Remote O&M functions may be exercised through a control agent
  (program).  In the direct approach, the router would support
  remote O&M functions directly from the NOC using standard Internet
  protocols (e.g., SNMP, UDP or TCP); in the indirect approach, the
  control agent would support these protocols and control the router
  itself using proprietary protocols.  The direct approach is
  preferred, although either approach is acceptable.  The use of
  specialized host hardware and/or software requiring significant
  additional investment is discouraged; nevertheless, some vendors
  may elect to provide the control agent as an integrated part of
  the network in which the routers are a part.  If this is the case,
  it is required that a means be available to operate the control
  agent from a remote site using Internet protocols and paths and
  with equivalent functionality with respect to a local agent
  terminal.
  It is desirable that a control agent and any other NOC software
  tools which a vendor provides operate as user programs in a
  standard operating system.  The use of the standard Internet
  protocols UDP and TCP for communicating with the routers should
  facilitate this.
  Remote router monitoring and (especially) remote router control
  present important access control problems which must be addressed.
  Care must also be taken to ensure control of the use of router
  resources for these functions.  It is not desirable to let router
  monitoring take more than some limited fraction of the router CPU
  time, for example.  On the other hand, O&M functions must receive
  priority so they can be exercised when the router is congested,
  since often that is when O&M is most needed.

10.3.2 Out Of Band Access

  Routers MUST support Out-Of-Band (OOB) access.  OOB access SHOULD
  provide the same functionality as in-band access.
  DISCUSSION:
     This Out-Of-Band access will allow the NOC a way to access
     isolated routers during times when network access is not
     available.
     Out-Of-Band access is an important management tool for the
     network administrator.  It allows the access of equipment
     independent of the network connections.  There are many ways to
     achieve this access.  Whichever one is used it is important
     that the access is independent of the network connections.  An
     example of Out-Of-Band access would be a serial port connected
     to a modem that provides dial up access to the router.
     It is important that the OOB access provides the same
     functionality as in-band access.  In-band access, or accessing
     equipment through the existing network connection, is limiting,
     because most of the time, administrators need to reach
     equipment to figure out why it is unreachable.  In band access
     is still very important for configuring a router, and for
     troubleshooting more subtle problems.

10.3.2 Router O&M Functions

10.3.2.1 Maintenance - Hardware Diagnosis

     Each router SHOULD operate as a stand-alone device for the
     purposes of local hardware maintenance.  Means SHOULD be
     available to run diagnostic programs at the router site using
     only on-site tools.  A router SHOULD be able to run diagnostics
     in case of a fault.  For suggested hardware and software
     diagnostics see Section [10.3.3].

10.3.2.2 Control - Dumping and Rebooting

     A router MUST include both in-band and out-of-band mechanisms
     to allow the network manager to reload, stop, and restart the
     router.  A router SHOULD also contain a mechanism (such as a
     watchdog timer) which will reboot the router automatically if
     it hangs due to a software or hardware fault.
     A router SHOULD IMPLEMENT a mechanism for dumping the contents
     of a router's memory (and/or other state useful for vendor
     debugging after a crash), and either saving them on a stable
     storage device local to the router or saving them on another
     host via an up-line dump mechanism such as TFTP (see [OPER:2],
     [INTRO:3]).

10.3.2.3 Control - Configuring the Router

     Every router has configuration parameters which may need to be
     set.  It SHOULD be possible to update the parameters without
     rebooting the router; at worst, a restart MAY be required.
     There may be cases when it is not possible to change parameters
     without rebooting the router (for instance, changing the IP
     address of an interface).  In these cases, care should be taken
     to minimize disruption to the router and the surrounding
     network.
     There SHOULD be a way to configure the router over the network
     either manually or automatically.  A router SHOULD be able to
     upload or download its parameters from a host or another
     router, and these parameters SHOULD be convertible into some
     sort of text format for making changes and then back to the
     form the router can read.  A router SHOULD have some sort of
     stable storage for its configuration. A router SHOULD NOT
     believe protocols such as RARP, ICMP Address Mask Reply, and
     MAY not believe BOOTP.
     DISCUSSION:
        It is necessary to note here that in the future RARP, ICMP
        Address Mask Reply, BOOTP and other mechanisms may be needed
        to allow a router to auto-configure.  Although routers may
        in the future be able to configure automatically, the intent
        here is to discourage this practice in a production
        environment until such time as auto-configuration has been
        tested more thoroughly. The intent is NOT to discourage
        auto-configuration all together.  In cases where a router is
        expected to get its configuration automatically it may be
        wise to allow the router to believe these things as it comes
        up and then ignore them after it has gotten its
        configuration.

10.3.2.4 Netbooting of System Software

     A router SHOULD keep its system image in local non-volatile
     storage such as PROM, NVRAM, or disk. It MAY also be able to
     load its system software over the network from a host or
     another router.
     A router which can keep its system image in local non-volatile
     storage MAY be configurable to boot its system image over the
     network.  A router which offers this option SHOULD be
     configurable to boot the system image in its non-volatile local
     storage if it is unable to boot its system image over the
     network.
     DISCUSSION:
        It is important that the router be able to come up and run
        on its own.  NVRAM may be a particular solution for routers
        used in large networks, since changing PROMs can be quite
        time consuming for a network manager responsible for
        numerous or geographically dispersed routers.  It is
        important to be able to netboot the system image because
        there should be an easy way for a router to get a bug fix or
        new feature more quickly than getting PROMS installed.  Also
        if the router has NVRAM instead of PROMs, it will netboot
        the image and then put it in NVRAM.
     A router MAY also be able to distinguish between different
     configurations based on which software it is running. If
     configuration commands change from one software version to
     another, it would be helpful if the router could use the
     configuration that was compatible with the software.

10.3.2.5 Detecting and responding to misconfiguration

     There MUST be mechanisms for detecting and responding to
     misconfigurations.  If a command is executed incorrectly, the
     router SHOULD give an error message.  The router SHOULD NOT
     accept a poorly formed command as if it were correct.
     DISCUSSION:
        There are cases where it is not possible to detect errors:
        the command is correctly formed, but incorrect with respect
        to the network.  This may be detected by the router, but may
        not be possible.
     Another form of misconfiguration is misconfiguration of the
     network to which the router is attached.  A router MAY detect
     misconfigurations in the network.  The router MAY log these
     findings to a file, either on the router or a host, so that the
     network manager will see that there are possible problems on
     the network.
     DISCUSSION:
        Examples of such misconfigurations might be another router
        with the same address as the one in question or a router
        with the wrong subnet mask.  If a router detects such
        problems it is probably not the best idea for the router to
        try to fix the situation.  That could cause more harm than
        good.

10.3.2.6 Minimizing Disruption

     Changing the configuration of a router SHOULD have minimal
     affect on the network.   Routing tables SHOULD NOT be
     unnecessarily flushed when a simple change is made to the
     router.  If a router is running several routing protocols,
     stopping one routing protocol SHOULD NOT disrupt other routing
     protocols, except in the case where one network is learned by
     more than one routing protocol.
     DISCUSSION:
        It is the goal of a network manager to run a network so that
        users of the network get the best connectivity possible.
        Reloading a router for simple configuration changes can
        cause disruptions in routing and ultimately cause
        disruptions to the network and its users.  If routing tables
        are unnecessarily flushed, for instance, the default route
        will be lost as well as specific routes to sites within the
        network.  This sort of disruption will cause significant
        downtime for the users. It is the purpose of this section to
        point out that whenever possible, these disruptions should
        be avoided.

10.3.2.7 Control - Troubleshooting Problems

     (1)  A router MUST provide in-band network access, but (except
          as required by Section [8.2]) for security considerations
          this access SHOULD be disabled by default.  Vendors MUST
          document the default state of any in-band access.
          DISCUSSION:
             In-band access primarily refers to access via the
             normal network protocols which may or may not affect
             the permanent operational state of the router.  This
             includes, but is not limited to Telnet/RLOGIN console
             access and SNMP operations.
             This was a point of contention between the operational
             out of the box and secure out of the box contingents.
             Any automagic access to the router may introduce
             insecurities, but it may be more important for the
             customer to have a router which is accessible over the
             network as soon as it is plugged in.  At least one
             vendor supplies routers without any external console
             access and depends on being able to access the router
             via the network to complete its configuration.
             Basically, it is the vendors call whether or not in-
             band access is enabled by default; but it is also the
             vendors responsibility to make its customers aware of
             possible insecurities.
     (2)  A router MUST provide the ability to initiate an ICMP
          echo.  The following options SHOULD be implemented:
          o  Choice of data patterns
          o  Choice of packet size
          o  Record route
          and the following additional options MAY be implemented:
          o  Loose source route
          o  Strict source route
          o  Timestamps
     (3)  A router SHOULD provide the ability to initiate a
          traceroute.  If traceroute is provided, then the 3rd party
          traceroute SHOULD be implemented.
     Each of the above three facilities (if implemented) SHOULD have
     access restrictions placed on it to prevent its abuse by
     unauthorized persons.

10.4 Security Considerations

10.4.1 Auditing and Audit Trails

  Auditing and billing are the bane of the network operator, but are
  the two features most requested by those in charge of network
  security and those who are responsible for paying the bills.  In
  the context of security, auditing is desirable if it helps you
  keep your network working and protects your resources from abuse,
  without costing you more than those resources are worth.
  (1)  Configuration Changes
       Router SHOULD provide a method for auditing a configuration
       change of a router, even if it's something as simple as
       recording the operator's initials and time of change.
       DISCUSSION:
          Having the ability to track who made changes and when is
          highly desirable, especially if your packets suddenly
          start getting routed through Alaska on their way across
          town.
  (2)  Packet Accounting
       Vendors should strongly consider providing a system for
       tracking traffic levels between pairs of hosts or networks.
       A mechanism for limiting the collection of this information
       to specific pairs of hosts or networks is also strongly
       encouraged.
       DISCUSSION:
          A host traffic matrix as described above can give the
          network operator a glimpse of traffic trends not apparent
          from other statistics.  It can also identify hosts or
          networks which are probing the structure of the attached
          networks - e.g., a single external host which tries to
          send packets to every IP address in the network address
          range for a connected network.
  (3)  Security Auditing
       Routers MUST provide a method for auditing security related
       failures or violations to include:
       o  Authorization Failures:  bad passwords, invalid SNMP
          communities, invalid authorization tokens,
       o  Violations of Policy Controls:  Prohibited Source Routes,
          Filtered Destinations, and
       o  Authorization Approvals:  good passwords - Telnet in-band
          access, console access.
       Routers MUST provide a method of limiting or disabling such
       auditing but auditing SHOULD be on by default.  Possible
       methods for auditing include listing violations to a console
       if present, logging or counting them internally, or logging
       them to a remote security server via the SNMP trap mechanism
       or the Unix logging mechanism as appropriate.  A router MUST
       implement at least one of these reporting mechanisms - it MAY
       implement more than one.

10.4.2 Configuration Control

  A vendor has a responsibility to use good configuration control
  practices in the creation of the software/firmware loads for their
  routers.  In particular, if a vendor makes updates and loads
  available for retrieval over the Internet, the vendor should also
  provide a way for the customer to confirm the load is a valid one,
  perhaps by the verification of a checksum over the load.
  DISCUSSION:
     Many vendors currently provide short notice updates of their
     software products via the Internet.  This a good trend and
     should be encouraged, but provides a point of vulnerability in
     the configuration control process.
  If a vendor provides the ability for the customer to change the
  configuration parameters of a router remotely, for example via a
  Telnet session, the ability to do so SHOULD be configurable and
  SHOULD default to off.  The router SHOULD require a password or
  other valid authentication before permitting remote
  reconfiguration.
  DISCUSSION:
     Allowing your properly identified network operator to twiddle
     with your routers is necessary; allowing anyone else to do so
     is foolhardy.
  A router MUST NOT have undocumented back door access and master
  passwords.  A vendor MUST ensure any such access added for
  purposes of debugging or product development are deleted before
  the product is distributed to its customers.
  DISCUSSION:
     A vendor has a responsibility to its customers to ensure they
     are aware of the vulnerabilities present in its code by
     intention - e.g.  in-band access.  Trap doors, back doors and
     master passwords intentional or unintentional can turn a
     relatively secure router into a major problem on an operational
     network.  The supposed operational benefits are not matched by
     the potential problems.

11. REFERENCES

Implementors should be aware that Internet protocol standards are occasionally updated. These references are current as of this writing, but a cautious implementor will always check a recent version of the RFC index to ensure that an RFC has not been updated or superseded by another, more recent RFC. Reference [INTRO:6] explains various ways to obtain a current RFC index.

APPL:1.

 B. Croft and J. Gilmore, Bootstrap Protocol (BOOTP), Request For
 Comments (RFC) 951, Stanford and SUN Microsystems, September 1985.

APPL:2.

 S. Alexander and R. Droms, DHCP Options and BOOTP Vendor
 Extensions, Request For Comments (RFC) 1533, Lachman Technology,
 Inc., Bucknell University, October 1993.

APPL:3.

 W. Wimer, Clarifications and Extensions for the Bootstrap Protocol,
 Request For Comments (RFC) 1542, Carnegie Mellon University,
 October 1993.

ARCH:1.

 DDN Protocol Handbook, NIC-50004, NIC-50005, NIC-50006 (three
 volumes), DDN Network Information Center, SRI International, Menlo
 Park, California, USA, December 1985.

ARCH:2.

 V. Cerf and R. Kahn, A Protocol for Packet Network
 Intercommunication," IEEE Transactions on Communication, May 1974.
 Also included in [ARCH:1].

ARCH:3.

 J. Postel, C. Sunshine, and D. Cohen, The ARPA Internet Protocol,"
 Computer Networks, vol. 5, no. 4, July 1981.  Also included in
 [ARCH:1].

ARCH:4.

 B. Leiner, J. Postel, R. Cole, and D. Mills, The DARPA Internet
 Protocol Suite, Proceedings of INFOCOM '85, IEEE, Washington, DC,
 March 1985.  Also in: IEEE Communications Magazine, March 1985.
 Also available from the Information Sciences Institute, University
 of Southern California as Technical Report ISI-RS-85-153.

ARCH:5.

 D. Comer, Internetworking With TCP/IP Volume 1: Principles,
 Protocols, and Architecture, Prentice Hall, Englewood Cliffs, NJ,
 1991.

ARCH:6.

 W. Stallings, Handbook of Computer-Communications Standards Volume
 3: The TCP/IP Protocol Suite, Macmillan, New York, NY, 1990.

ARCH:7.

 J. Postel, Internet Official Protocol Standards, Request For
 Comments (RFC) 1610, STD 1, USC/Information Sciences Institute,
 July 1994.

ARCH:8.

 Information processing systems - Open Systems Interconnection -
 Basic Reference Model, ISO 7489, International Standards
 Organization, 1984.

FORWARD:1.

 IETF CIP Working Group (C. Topolcic, Editor), Experimental Internet
 Stream Protocol, Version 2 (ST-II), Request For Comments (RFC)
 1190, CIP Working Group, October 1990.

FORWARD:2.

 A. Mankin and K. Ramakrishnan, Editors, Gateway Congestion Control
 Survey, Request For Comments (RFC) 1254, MITRE, Digital Equipment
 Corporation, August 1991.

FORWARD:3.

 J. Nagle, On Packet Switches with Infinite Storage, IEEE
 Transactions on Communications, vol. COM-35, no. 4, April 1987.

FORWARD:4.

 R. Jain, K. Ramakrishnan, and D. Chiu, Congestion Avoidance in
 Computer Networks With a Connectionless Network Layer, Technical
 Report DEC-TR-506, Digital Equipment Corporation.

FORWARD:5.

 V. Jacobson, Congestion Avoidance and Control, Proceedings of
 SIGCOMM '88, Association for Computing Machinery, August 1988.

FORWARD:6.

 W. Barns, Precedence and Priority Access Implementation for
 Department of Defense Data Networks, Technical Report MTR-91W00029,
 The Mitre Corporation, McLean, Virginia, USA, July 1991.

INTERNET:1.

 J. Postel, Internet Protocol, Request For Comments (RFC) 791, STD
 5, USC/Information Sciences Institute, September 1981.

INTERNET:2.

 J. Mogul and J. Postel, Internet Standard Subnetting Procedure,
 Request For Comments (RFC) 950, STD 5, USC/Information Sciences
 Institute, August 1985.

INTERNET:3.

 J. Mogul, Broadcasting Internet Datagrams in the Presence of
 Subnets, Request For Comments (RFC) 922, STD 5, Stanford, October
 1984.

INTERNET:4.

 S. Deering, Host Extensions for IP Multicasting, Request For
 Comments (RFC) 1112, STD 5, Stanford University, August 1989.

INTERNET:5.

 S. Kent, U.S. Department of Defense Security Options for the
 Internet Protocol, Request for Comments (RFC) 1108, BBN
 Communications, November 1991.

INTERNET:6.

 R. Braden, D. Borman, and C. Partridge, Computing the Internet
 Checksum, Request For Comments (RFC) 1071, USC/Information Sciences
 Institute, Cray Researc, BBN, September 1988.

INTERNET:7.

 T. Mallory and A. Kullberg, Incremental Updating of the Internet
 Checksum, Request For Comments (RFC) 1141, BBN, January 1990.

INTERNET:8.

 J. Postel, Internet Control Message Protocol, Request For Comments
 (RFC) 792, STD 5, USC/Information Sciences Institute, September
 1981.

INTERNET:9.

 A. Mankin, G. Hollingsworth, G. Reichlen, K. Thompson, R.  Wilder,
 and R. Zahavi, Evaluation of Internet Performance - FY89, Technical
 Report MTR-89W00216, MITRE Corporation, February, 1990.

INTERNET:10.

 G. Finn, A Connectionless Congestion Control Algorithm, Computer
 Communications Review, vol. 19, no. 5, Association for Computing
 Machinery, October 1989.

INTERNET:11.

 W. Prue, J. Postel, The Source Quench Introduced Delay (SQuID),
 Request For Comments (RFC) 1016, USC/Information Sciences
 Institute, August 1987.

INTERNET:12.

 A. McKenzie, Some comments on SQuID, Request For Comments (RFC)
 1018, BBN, August 1987.

INTERNET:13.

 S. Deering, ICMP Router Discovery Messages, Request For Comments
 (RFC) 1256, Xerox PARC, September 1991.

INTERNET:14.

 J. Mogul and S. Deering, Path MTU Discovery, Request For Comments
 (RFC) 1191, DECWRL, Stanford University, November 1990.

INTERNET:15

 V. Fuller, T. Li, J. Yi, and K. Varadhan, Classless Inter-Domain
 Routing (CIDR): an Address Assignment and Aggregation Strategy
 Request For Comments (RFC) 1519, BARRNet, cisco, Merit, OARnet,
 September 1993.

INTERNET:16

 M. St. Johns, Draft Revised IP Security Option, Request for
 Comments 1038, IETF, January 1988.

INTERNET:17

 W. Prue and J. Postel, Queuing Algorithm to Provide Type-of-service
 For IP Links, Request for Comments 1046, USC/Information Sciences
 Institute, February 1988.

INTRO:1.

 R. Braden and J. Postel, Requirements for Internet Gateways,
 Request For Comments (RFC) 1009, STD 4, USC/Information Sciences
 Institute, June 1987.

INTRO:2.

 Internet Engineering Task Force (R. Braden, Editor), Requirements
 for Internet Hosts - Communication Layers, Request For Comments
 (RFC) 1122, STD 3, USC/Information Sciences Institute, October
 1989.

INTRO:3.

 Internet Engineering Task Force (R. Braden, Editor), Requirements
 for Internet Hosts - Application and Support, Request For Comments
 (RFC) 1123, STD 3, USC/Information Sciences Institute, October
 1989.

INTRO:4.

 D. Clark, Modularity and Efficiency in Protocol Implementations,
 Request For Comments (RFC) 817, MIT, July 1982.

INTRO:5.

 D. Clark, The Structuring of Systems Using Upcalls, Proceedings of
 10th ACM SOSP, December 1985.

INTRO:6.

 O. Jacobsen and J. Postel, Protocol Document Order Information,
 Request For Comments (RFC) 980, SRI, USC/Information Sciences
 Institute, March 1986.

INTRO:7.

 J. Reynolds and J. Postel, Assigned Numbers, Request For Comments
 (RFC) 1700, STD 2, USC/Information Sciences Institute, October
 1994.  This document is periodically updated and reissued with a
 new number.  It is wise to verify occasionally that the version you
 have is still current.

INTRO:8.

 DoD Trusted Computer System Evaluation Criteria, DoD publication
 5200.28-STD, U.S. Department of Defense, December 1985.

INTRO:9

 G. Malkin and T. LaQuey Parker, Internet Users' Glossary, Request
 for Comments (RFC) 1392 (also FYI 0018), Xylogics, Inc., UTexas,
 January 1993.

LINK:1.

 S. Leffler and M. Karels, Trailer Encapsulations, Request For
 Comments (RFC) 893, U. C. Berkeley, April 1984.

LINK:2

 W. Simpson, The Point-to-Point Protocol (PPP) for the Transmission
 of Multi-protocol Datagrams over Point-to-Point Links, Daydreamer,
 Request For Comments (RFC) 1331, May 1992.

LINK:3

 G. McGregor, The PPP Internet Protocol Control Protocol (IPCP),
 Request For Comments (RFC) 1332, Merit, May 1992.

LINK:4

 B. Lloyd, W. Simpson, PPP Authentication Protocols, Request For
 Comments (RFC) 1334, Daydreamer, May 1992.

LINK:5

 W. Simpson, PPP Link Quality Monitoring, Daydreamer, Request For
 Comments (RFC) 1333, May 1992.

MGT:1.

 M. Rose and K. McCloghrie, Structure and Identification of
 Management Information of TCP/IP-based Internets, Request For
 Comments (RFC) 1155, STD 16, Performance Systems International,
 Hughes LAN Systems, May 1990.

MGT:2.

 K. McCloghrie and M. Rose (Editors), Management Information Base of
 TCP/IP-Based Internets: MIB-II, Request For Comments (RFC) 1213,
 STD 16, Hughes LAN Systems, Performance Systems International,
 March 1991.

MGT:3.

 J. Case, M. Fedor, M. Schoffstall, and J. Davin, Simple Network
 Management Protocol, Request For Comments (RFC) 1157, STD 15, SNMP
 Research, Performance Systems International, MIT Laboratory for
 Computer Science, May 1990.

MGT:4.

 M. Rose and K. McCloghrie (Editors), Towards Concise MIB
 Definitions, Request For Comments (RFC) 1212, STD 16, Performance
 Systems International, Hughes LAN Systems, March 1991.

MGT:5.

 L. Steinberg, Techniques for Managing Asynchronously Generated
 Alerts, Request for Comments (RFC) 1224, IBM, May 1991.

MGT:6.

 F. Kastenholz, Definitions of Managed Objects for the Ethernet-like
 Interface Types, Request for Comments (RFC) 1398, FTP Software
 January 1993.

MGT:7.

 R. Fox and K. McCloghrie, IEEE 802.4 Token Bus MIB, Request for
 Comments (RFC) 1230, Hughes LAN Systems, Synoptics, Inc., May 1991.

MGT:8.

 K. McCloghrie, R. Fox and E. Decker, IEEE 802.5 Token Ring MIB,
 Request for Comments (RFC) 1231, Hughes LAN Systems, Synoptics,
 Inc., cisco Systems, Inc., February 1993.

MGT:9.

 J. Case and A. Rijsinghani, FDDI Management Information Base,
 Request for Comments (RFC) 1512, SNMP Research, Digital Equipment
 Corporation, September 1993.

MGT:10.

 B. Stewart, Definitions of Managed Objects for RS-232-like Hardware
 Devices, Request for Comments (RFC) 1317, Xyplex, Inc., April 1992.

MGT:11.

 F. Kastenholz, Definitions of Managed Objects for the Link Control
 Protocol of the Point-to-Point Protocol, Request For Comments (RFC)
 1471, FTP Software, June 1992.

MGT:12.

 F. Kastenholz, The Definitions of Managed Objects for the Security
 Protocols of the Point-to-Point Protocol, Request For Comments
 (RFC) 1472, FTP Software, June 1992.

MGT:13.

 F. Kastenholz, The Definitions of Managed Objects for the IP
 Network Control Protocol of the Point-to-Point Protocol, Request
 For Comments (RFC) 1473, FTP Software, June 1992.

MGT:14.

 F. Baker and R. Coltun, OSPF Version 2 Management Information Base,
 Request For Comments (RFC) 1253, ACC, Computer Science Center,
 August 1991.

MGT:15.

 S. Willis and J. Burruss, Definitions of Managed Objects for the
 Border Gateway Protocol (Version 3), Request For Comments (RFC)
 1269, Wellfleet Communications Inc., October 1991.

MGT:16.

 F. Baker, J. Watt, Definitions of Managed Objects for the DS1 and
 E1 Interface Types, Request For Comments (RFC) 1406, Advanced
 Computer Communications, Newbridge Networks Corporation, January
 1993.

MGT:17.

 T. Cox and K. Tesink, Definitions of Managed Objects for the DS3/E3
 Interface Types, Request For Comments (RFC) 1407, Bell
 Communications Research, January 1993.

MGT:18.

 K. McCloghrie, Extensions to the Generic-Interface MIB, Request For
 Comments (RFC) 1229,  Hughes LAN Systems, August 1992.

MGT:19.

 T. Cox and K. Tesink, Definitions of Managed Objects for the SIP
 Interface Type, Request For Comments (RFC) 1304, Bell
 Communications Research, February 1992.

MGT:20

 F. Baker, IP Forwarding Table MIB, Request For Comments (RFC) 1354,
 ACC, July 1992.

MGT:21.

 G. Malkin and F. Baker, RIP Version 2 MIB Extension, Request For
 Comments (RFC) 1389, Xylogics, Inc., Advanced Computer
 Communications, January 1993.

MGT:22.

 D. Throop, SNMP MIB Extension for the X.25 Packet Layer, Request
 For Comments (RFC) 1382, Data General Corporation, November 1992.

MGT:23.

 D. Throop and F. Baker, SNMP MIB Extension for X.25 LAPB, Request
 For Comments (RFC) 1381, Data General Corporation, Advanced
 Computer Communications, November 1992.

MGT:24.

 D. Throop and F. Baker, SNMP MIB Extension for MultiProtocol
 Interconnect over X.25, Request For Comments (RFC) 1461, Data
 General Corporation, May 1993.

MGT:25.

 M. Rose, SNMP over OSI, Request For Comments (RFC) 1418, Dover
 Beach Consulting, Inc., March 1993.

MGT:26.

 G. Minshall and M. Ritter, SNMP over AppleTalk, Request For
 Comments (RFC) 1419, Novell, Inc., Apple Computer, Inc., March
 1993.

MGT:27.

 S. Bostock, SNMP over IPX, Request For Comments (RFC) 1420, Novell,
 Inc., March 1993.

MGT:28.

 M. Schoffstall, C. Davin, M. Fedor, J. Case, SNMP over Ethernet,
 Request For Comments (RFC) 1089, Rensselaer Polytechnic Institute,
 MIT Laboratory for Computer Science, NYSERNet, Inc., University of
 Tennessee at Knoxville, February 1989.

MGT:29.

 J. Case, FDDI Management Information Base, Request For Comments
 (RFC) 1285, SNMP Research, Incorporated, January 1992.

OPER:1.

 J. Nagle, Congestion Control in IP/TCP Internetworks, Request For
 Comments (RFC) 896, FACC, January 1984.

OPER:2.

 K.R. Sollins, TFTP Protocol (revision 2), Request For Comments
 (RFC) 1350, MIT, July 1992.

ROUTE:1.

 J. Moy, OSPF Version 2, Request For Comments (RFC) 1247, Proteon,
 July 1991.

ROUTE:2.

 R. Callon, Use of OSI IS-IS for Routing in TCP/IP and Dual
 Environments, Request For Comments (RFC) 1195, DEC, December 1990.

ROUTE:3.

 C. L. Hedrick, Routing Information Protocol, Request For Comments
 (RFC) 1058, Rutgers University, June 1988.

ROUTE:4.

 K. Lougheed and Y. Rekhter, A Border Gateway Protocol 3 (BGP-3),
 Request For Comments (RFC) 1267, cisco, T.J. Watson Research
 Center, IBM Corp., October 1991.

ROUTE:5.

 Y. Rekhter and P. Gross Application of the Border Gateway Protocol
 in the Internet, Request For Comments (RFC) 1268, T.J. Watson
 Research Center, IBM Corp., ANS, October 1991.

ROUTE:6.

 D. Mills, Exterior Gateway Protocol Formal Specification, Request
 For Comments (RFC) 904, UDEL, April 1984.

ROUTE:7.

 E. Rosen, Exterior Gateway Protocol (EGP), Request For Comments
 (RFC) 827, BBN, October 1982.

ROUTE:8.

 L. Seamonson and E. Rosen, "STUB" Exterior Gateway Protocol,
 Request For Comments (RFC) 888, BBN, January 1984.

ROUTE:9.

 D. Waitzman, C. Partridge, and S. Deering, Distance Vector
 Multicast Routing Protocol, Request For Comments (RFC) 1075, BBN,
 Stanford, November 1988.

ROUTE:10.

 S. Deering, Multicast Routing in Internetworks and Extended LANs,
 Proceedings of SIGCOMM '88, Association for Computing Machinery,
 August 1988.

ROUTE:11.

 P. Almquist, Type of Service in the Internet Protocol Suite,
 Request for Comments (RFC) 1349, Consultant, July 1992.

ROUTE:12.

 Y. Rekhter, Experience with the BGP Protocol, Request For Comments
 (RFC) 1266, T.J. Watson Research Center, IBM Corp., October 1991.

ROUTE:13.

 Y. Rekhter, BGP Protocol Analysis, Request For Comments (RFC) 1265,
 T.J. Watson Research Center, IBM Corp., October 1991.

TRANS:1.

 J. Postel, User Datagram Protocol, Request For Comments (RFC) 768,
 STD 6, USC/Information Sciences Institute, August 1980.

TRANS:2.

 J. Postel, Transmission Control Protocol, Request For Comments
 (RFC) 793, STD 7, T.J. Watson Research Center, IBM Corp., September
 1981.

APPENDIX A. REQUIREMENTS FOR SOURCE-ROUTING HOSTS

Subject to restrictions given below, a host MAY be able to act as an intermediate hop in a source route, forwarding a source-routed datagram to the next specified hop.

However, in performing this router-like function, the host MUST obey all the relevant rules for a router forwarding source-routed datagrams [INTRO:2]. This includes the following specific provisions:

(A) TTL

 The TTL field MUST be decremented and the datagram perhaps
 discarded as specified for a router in [INTRO:2].

(B) ICMP Destination Unreachable

 A host MUST be able to generate Destination Unreachable messages
 with the following codes:
 4 (Fragmentation Required but DF Set) when a source-routed datagram
   cannot be fragmented to fit into the target network;
 5 (Source Route Failed) when a source-routed datagram cannot be
   forwarded, e.g., because of a routing problem or because the next
   hop of a strict source route is not on a connected network.

(C) IP Source Address

 A source-routed datagram being forwarded MAY (and normally will)
 have a source address that is not one of the IP addresses of the
 forwarding host.

(D) Record Route Option

 A host that is forwarding a source-routed datagram containing a
 Record Route option MUST update that option, if it has room.

(E) Timestamp Option

 A host that is forwarding a source-routed datagram containing a
 Timestamp Option MUST add the current timestamp to that option,
 according to the rules for this option.

To define the rules restricting host forwarding of source-routed datagrams, we use the term local source-routing if the next hop will be through the same physical interface through which the datagram arrived; otherwise, it is non-local source-routing.

A host is permitted to perform local source-routing without restriction.

A host that supports non-local source-routing MUST have a configurable switch to disable forwarding, and this switch MUST default to disabled.

The host MUST satisfy all router requirements for configurable policy filters [INTRO:2] restricting non-local forwarding.

If a host receives a datagram with an incomplete source route but does not forward it for some reason, the host SHOULD return an ICMP Destination Unreachable (code 5, Source Route Failed) message, unless the datagram was itself an ICMP error message.

APPENDIX B. GLOSSARY

This Appendix defines specific terms used in this memo. It also defines some general purpose terms that may be of interest. See also [INTRO:9] for a more general set of definitions.

AS

 Autonomous System A collection of routers under a single
 administrative authority using a common Interior Gateway Protocol
 for routing packets.

Connected Network

 A network to which a router is interfaced is often known as the
 local network or the subnetwork relative to that router. However,
 these terms can cause confusion, and therefore we use the term
 Connected Network in this memo.

Connected (Sub)Network

 A Connected (Sub)Network is an IP subnetwork to which a router is
 interfaced, or a connected network if the connected network is not
 subnetted.  See also Connected Network.

Datagram

 The unit transmitted between a pair of internet modules.  data,
 called datagrams, from sources to destinations.  The Internet
 Protocol does not provide a reliable communication facility.  There
 are no acknowledgments either end-to-end or hop-by-hop.  There is
 no error no retransmissions.  There is no flow control.  See IP.

Default Route

 A routing table entry which is used to direct any data addressed to
 any network numbers not explicitly listed in the routing table.

EGP

 Exterior Gateway Protocol A protocol which distributes routing
 information to the gateways (routers) which connect autonomous
 systems.  See IGP.

EGP-2

 Exterior Gateway Protocol version 2 This is an EGP routing protocol
 developed to handle traffic between AS's in the Internet.

Forwarder

 The logical entity within a router that is responsible for
 switching packets among the router's interfaces.  The Forwarder
 also makes the decisions to queue a packet for local delivery, to
 queue a packet for transmission out another interface, or both.

Forwarding

 Forwarding is the process a router goes through for each packet
 received by the router.  The packet may be consumed by the router,
 it may be output on one or more interfaces of the router, or both.
 Forwarding includes the process of deciding what to do with the
 packet as well as queuing it up for (possible) output or internal
 consumption.

Fragment

 An IP datagram which represents a portion of a higher layer's
 packet which was too large to be sent in its entirety over the
 output network.

IGP

 Interior Gateway Protocol A protocol which distributes routing
 information with an Autonomous System (AS).  See EGP.

Interface IP Address

 The IP Address and subnet mask that is assigned to a specific
 interface of a router.

Internet Address

 An assigned number which identifies a host in an internet.  It has
 two or three parts: network number, optional subnet number, and
 host number.

IP

 Internet Protocol The network layer protocol for the Internet.  It
 is a packet switching, datagram protocol defined in RFC 791.  IP
 does not provide a reliable communications facility; that is, there
 are no end-to-end of hop-by-hop acknowledgments.

IP Datagram

 An IP Datagram is the unit of end-to-end transmission in the
 Internet Protocol.  An IP Datagram consists of an IP header
 followed by all of higher-layer data (such as TCP, UDP, ICMP, and
 the like).  An IP Datagram is an IP header followed by a message.
 An IP Datagram is a complete IP end-to-end transmission unit.  An
 IP Datagram is composed of one or more IP Fragments.
 In this memo, the unqualified term Datagram should be understood to
 refer to an IP Datagram.

IP Fragment

 An IP Fragment is a component of an IP Datagram.  An IP Fragment
 consists of an IP header followed by all or part of the higher-
 layer of the original IP Datagram.
 One or more IP Fragments comprises a single IP Datagram.
 In this memo, the unqualified term Fragment should be understood to
 refer to an IP Fragment.

IP Packet

 An IP Datagram or an IP Fragment.
 In this memo, the unqualified term Packet should generally be
 understood to refer to an IP Packet.

Logical [network] interface

 We define a logical [network] interface to be a logical path,
 distinguished by a unique IP address, to a connected network.

Martian Filtering

 A packet which contains an invalid source or destination address is
 considered to be martian and discarded.

MTU (Maximum Transmission Unit)

 The size of the largest packet that can be transmitted or received
 through a logical interface.  This size includes the IP header but
 does not include the size of any Link Layer headers or framing.

Multicast

 A packet which is destined for multiple hosts.  See broadcast.

Multicast Address

 A special type of address which is recognized by multiple hosts.
 A Multicast Address is sometimes known as a Functional Address or a
 Group Address.

Originate

 Packets can be transmitted by a router for one of two reasons: 1)
 the packet was received and is being forwarded or 2) the router
 itself created the packet for transmission (such as route
 advertisements).  Packets that the router creates for transmission
 are said to originate at the router.

Packet

 A packet is the unit of data passed across the interface between
 the Internet Layer and the Link Layer.  It includes an IP header
 and data.  A packet may be a complete IP datagram or a fragment of
 an IP datagram.

Path

 The sequence of routers and (sub-)networks which a packet traverses
 from a particular router to a particular destination host.  Note
 that a path is uni-directional; it is not unusual to have different
 paths in the two directions between a given host pair.

Physical Network

 A Physical Network is a network (or a piece of an internet) which
 is contiguous at the Link Layer.  Its internal structure (if any)
 is transparent to the Internet Layer.
 In this memo, several media components that are connected together
 via devices such as bridges or repeaters are considered to be a
 single Physical Network since such devices are transparent to the
 IP.

Physical Network Interface

 This is a physical interface to a Connected Network and has a
 (possibly unique) Link-Layer address.  Multiple Physical Network
 Interfaces on a single router may share the same Link-Layer
 address, but the address must be unique for different routers on
 the same Physical Network.

router

 A special-purpose dedicated computer that attaches several networks
 together.  Routers switch packets between these networks in a
 process known as forwarding.  This process may be repeated several
 times on a single packet by multiple routers until the packet can
 be delivered to the final destination - switching the packet from
 router to router to router... until the packet gets to its
 destination.

RPF

 Reverse Path Forwarding A method used to deduce the next hops for
 broadcast and multicast packets.

serial line

 A physical medium which we cannot define, but we recognize one when
 we see one.  See the U.S. Supreme Court's definitions on
 pornography.

Silently Discard

 This memo specifies several cases where a router is to Silently
 Discard a received packet (or datagram).  This means that the
 router should discard the packet without further processing, and
 that the router will not send any ICMP error message (see Section
 [4.3.2]) as a result.  However, for diagnosis of problems, the
 router should provide the capability of logging the error (see
 Section [1.3.3]), including the contents of the silently-discarded
 packet, and should record the event in a statistics counter.

Silently Ignore

 A router is said to Silently Ignore an error or condition if it
 takes no action other than possibly generating an error report in
 an error log or via some network management protocol, and
 discarding, or ignoring, the source of the error.  In particular,
 the router does NOT generate an ICMP error message.

Specific-destination address

 This is defined to be the destination address in the IP header
 unless the header contains an IP broadcast or IP multicast address,
 in which case the specific-destination is an IP address assigned to
 the physical interface on which the packet arrived.

subnet

 A portion of a network, which may be a physically independent
 network, which shares a network address with other portions of the
 network and is distinguished by a subnet number.  A subnet is to a
 network what a network is to an internet.

subnet number

 A part of the internet address which designates a subnet.  It is
 ignored for the purposes internet routing, but is used for intranet
 routing.

TOS

 Type Of Service A field in the IP header which represents the
 degree of reliability expected from the network layer by the
 transport layer or application.

TTL

 Time To Live A field in the IP header which represents how long a
 packet is considered valid.  It is a combination hop count and
 timer value.

APPENDIX C. FUTURE DIRECTIONS

This appendix lists work that future revisions of this document may wish to address.

In the preparation of Router Requirements, we stumbled across several other architectural issues. Each of these is dealt with somewhat in the document, but still ought to be classified as an open issue in the IP architecture.

Most of the he topics presented here generally indicate areas where the technology is still relatively new and it is not appropriate to develop specific requirements since the community is still gaining operational experience.

Other topics represent areas of ongoing research and indicate areas that the prudent developer would closely monitor.

(1) SNMP Version 2

(2) Additional SNMP MIBs

(3) IDPR

(4) CIPSO

(5) IP Next Generation research

(6) More detailed requirements for next-hop selection

(7) More detailed requirements for leaking routes between routing

 protocols

(8) Router system security

(9) Routing protocol security

(10) Internetwork Protocol layer security. There has been extensive

 work refining the security of IP since the original work writing
 this document.  This security work should be included in here.

(11) Route caching

(12) Load Splitting

(13) Sending fragments along different paths

(14) Variable width subnet masks (i.e., not all subnets of a particular

 net use the same subnet mask).  Routers are required (MUST) support
 them, but are not required to detect ambiguous configurations.

(15) Multiple logical (sub)nets on the same wire. Router Requirements

 does not require support for this.  We made some attempt to
 identify pieces of the architecture (e.g. forwarding of directed
 broadcasts and issuing of Redirects) where the wording of the rules
 has to be done carefully to make the right thing happen, and tried
 to clearly distinguish logical interfaces from physical interfaces.
 However, we did not study this issue in detail, and we are not at
 all confident that all of the rules in the document are correct in
 the presence of multiple logical (sub)nets on the same wire.

(15) Congestion control and resource management. On the advice of the

 IETF's experts (Mankin and Ramakrishnan) we deprecated (SHOULD NOT)
 Source Quench and said little else concrete (Section 5.3.6).

(16) Developing a Link-Layer requirements document that would be common

 for both routers and hosts.

(17) Developing a common PPP LQM algorithm.

(18) Investigate of other information (above and beyond section [3.2])

 that passes between the layers, such as physical network MTU,
 mappings of IP precedence to Link Layer priority values, etc.

(19) Should the Link Layer notify IP if address resolution failed (just

 like it notifies IP when there is a Link Layer priority value
 problem)?

(20) Should all routers be required to implement a DNS resolver?

(21) Should a human user be able to use a host name anywhere you can use

 an IP address when configuring the router? Even in ping and
 traceroute?

(22) Almquist's draft ruminations on the next hop and ruminations on

 route leaking need to be reviewed, brought up to date, and
 published.

(23) Investigation is needed to determine if a redirect message for

 precedence is needed or not. If not, are the type-of-service
 redirects acceptable?

(24) RIPv2 and RIP+CIDR and variable length subnet masks.

(25) BGP-4 CIDR is going to be important, and everyone is betting on

 BGP-4. We can't avoid mentioning it.  Probably need to describe the
 differences between BGP-3 and BGP-4, and explore upgrade issues...

(26) Loose Source Route Mobile IP and some multicasting may require

 this.  Perhaps it should be elevated to a SHOULD (per Fred Baker's
 Suggestion).

APPENDIX D. Multicast Routing Protocols

Multicasting is a relatively new technology within the Internet Protocol family. It is not widely deployed or commonly in use yet. Its importance, however, is expected to grow over the coming years.

This Appendix describes some of the technologies being investigated for routing multicasts through the Internet.

A diligent implementor will keep abreast of developments in this area in order to properly develop multicast facilities.

This Appendix does not specify any standards or requirements.

D.1 Introduction

Multicast routing protocols enable the forwarding of IP multicast datagrams throughout a TCP/IP internet. Generally these algorithms forward the datagram based on its source and destination addresses. Additionally, the datagram may need to be forwarded to several multicast group members, at times requiring the datagram to be replicated and sent out multiple interfaces.

The state of multicast routing protocols is less developed than the protocols available for the forwarding of IP unicasts. Two multicast routing protocols have been documented for TCP/IP; both are currently considered to be experimental. Both also use the IGMP protocol (discussed in Section [4.4]) to monitor multicast group membership.

D.2 Distance Vector Multicast Routing Protocol - DVMRP

DVMRP, documented in [ROUTE:9], is based on Distance Vector or Bellman-Ford technology. It routes multicast datagrams only, and does so within a single Autonomous System. DVMRP is an implementation of the Truncated Reverse Path Broadcasting algorithm described in [ROUTE:10]. In addition, it specifies the tunneling of IP multicasts through non-multicast-routing-capable IP domains.

D.3 Multicast Extensions to OSPF - MOSPF

MOSPF, currently under development, is a backward-compatible addition to OSPF that allows the forwarding of both IP multicasts and unicasts within an Autonomous System. MOSPF routers can be mixed with OSPF routers within a routing domain, and they will interoperate in the forwarding of unicasts. OSPF is a link-state or SPF-based protocol. By adding link state advertisements that pinpoint group membership, MOSPF routers can calculate the path of a multicast datagram as a tree rooted at the datagram source. Those branches that do not contain group members can then be discarded, eliminating unnecessary datagram forwarding hops.

APPENDIX E Additional Next-Hop Selection Algorithms

Section [5.2.4.3] specifies an algorithm that routers ought to use when selecting a next-hop for a packet.

This appendix provides historical perspective for the next-hop selection problem. It also presents several additional pruning rules and next-hop selection algorithms that might be found in the Internet.

This appendix presents material drawn from an earlier, unpublished, work by Philip Almquist; Ruminations on the Next Hop.

This Appendix does not specify any standards or requirements.

E.1. Some Historical Perspective

It is useful to briefly review the history of the topic, beginning with what is sometimes called the "classic model" of how a router makes routing decisions. This model predates IP. In this model, a router speaks some single routing protocol such as RIP. The protocol completely determines the contents of the router's FIB. The route lookup algorithm is trivial: the router looks in the FIB for a route whose destination attribute exactly matches the network number portion of the destination address in the packet. If one is found, it is used; if none is found, the destination is unreachable. Because the routing protocol keeps at most one route to each destination, the problem of what to do when there are multiple routes which match the same destination cannot arise.

Over the years, this classic model has been augmented in small ways. With the advent of default routes, subnets, and host routes, it became possible to have more than one routing table entry which in some sense matched the destination. This was easily resolved by a consensus that there was a hierarchy of routes: host routes should be preferred over subnet routes, subnet routes over net routes, and net routes over default routes.

With the advent of variable length subnet masks, the general approach remained the same although its description became a little more complicated. We now say that each route has a bit mask associated with it. If a particular bit in a route's bit mask is set, the corresponding bit in the route's destination attribute is significant. A route cannot be used to route a packet unless each significant bit in the route's destination attribute matches the corresponding bit in the packet's destination address, and routes with more bits set in their masks are preferred over routes which have fewer bits set in their masks. This is simply a generalization

of the hierarchy of routes described above, and will be referred to for the rest of this memo as choosing a route by preferring longest match.

Another way the classic model has been augmented is through a small amount of relaxation of the notion that a routing protocol has complete control over the contents of the routing table. First, static routes were introduced. For the first time, it was possible to simultaneously have two routes (one dynamic and one static) to the same destination. When this happened, a router had to have a policy (in some cases configurable, and in other cases chosen by the author of the router's software) which determined whether the static route or the dynamic route was preferred. However, this policy was only used as a tie-breaker when longest match didn't uniquely determine which route to use. Thus, for example, a static default route would never be preferred over a dynamic net route even if the policy preferred static routes over dynamic routes.

The classic model had to be further augmented when inter-domain routing protocols were invented. Traditional routing protocols came to be called "interior gateway protocols" (IGPs), and at each Internet site there was a strange new beast called an "exterior gateway", a router which spoke EGP to several "BBN Core Gateways" (the routers which made up the Internet backbone at the time) at the same time as it spoke its IGP to the other routers at its site. Both protocols wanted to determine the contents of the router's routing table. Theoretically, this could result in a router having three routes (EGP, IGP, and static) to the same destination. Because of the Internet topology at the time, it was resolved with little debate that routers would be best served by a policy of preferring IGP routes over EGP routes. However, the sanctity of longest match remained unquestioned: a default route learned from the IGP would never be preferred over a net route from learned EGP.

Although the Internet topology, and consequently routing in the Internet, have evolved considerably since then, this slightly augmented version of the classic model has survived pretty much intact to this day in the Internet (except that BGP has replaced EGP). Conceptually (and often in implementation) each router has a routing table and one or more routing protocol processes. Each of these processes can add any entry that it pleases, and can delete or modify any entry that it has created. When routing a packet, the router picks the best route using longest match, augmented with a policy mechanism to break ties. Although this augmented classic model has served us well, it has a number of shortcomings:

o It ignores (although it could be augmented to consider) path

  characteristics such as quality of service and MTU.

o It doesn't support routing protocols (such as OSPF and Integrated

  IS-IS) that require route lookup algorithms different than pure
  longest match.

o There has not been a firm consensus on what the tie-breaking

  mechanism ought to be. Tie-breaking mechanisms have often been
  found to be difficult if not impossible to configure in such a way
  that the router will always pick what the network manger considers
  to be the "correct" route.

E.2. Additional Pruning Rules

Section [5.2.4.3] defined several pruning rules to use to select routes from the FIB. There are other rules that could also be used.

o OSPF Route Class

  Routing protocols which have areas or make a distinction between
  internal and external routes divide their routes into classes,
  where classes are rank-ordered in terms of preference. A route is
  always chosen from the most preferred class unless none is
  available, in which case one is chosen from the second most
  preferred class, and so on. In OSPF, the classes (in order from
  most preferred to least preferred) are intra-area, inter-area,
  type 1 external (external routes with internal metrics), and type
  2 external. As an additional wrinkle, a router is configured to
  know what addresses ought to be accessible via intra-area routes,
  and will not use inter- area or external routes to reach these
  destinations even when no intra-area route is available.
  More precisely, we assume that each route has a class attribute,
  called route.class, which is assigned by the routing protocol.
  The set of candidate routes is examined to determine if it
  contains any for which route.class = intra-area.  If so, all
  routes except those for which route.class = intra-area are
  discarded.  Otherwise, router checks whether the packet's
  destination falls within the address ranges configured for the
  local area.  If so, the entire set of candidate routes is deleted.
  Otherwise, the set of candidate routes is examined to determine if
  it contains any for which route.class = inter-area.  If so, all
  routes except those for which route.class = inter-area are
  discarded.  Otherwise, the set of candidate routes is examined to
  determine if it contains any for which route.class = type 1
  external.  If so, all routes except those for which route.class =
  type 1 external are discarded.

o IS-IS Route Class

  IS-IS route classes work identically to OSPF's. However, the set
  of classes defined by Integrated IS-IS is different, such that
  there isn't a one-to-one mapping between IS-IS route classes and
  OSPF route classes. The route classes used by Integrated IS-IS are
  (in order from most preferred to least preferred) intra-area,
  inter-area, and external.
  The Integrated IS-IS internal class is equivalent to the OSPF
  internal class. Likewise, the Integrated IS-IS external class is
  equivalent to OSPF's type 2 external class. However, Integrated
  IS-IS does not make a distinction between inter-area routes and
  external routes with internal metrics - both are considered to be
  inter-area routes. Thus, OSPF prefers true inter-area routes over
  external routes with internal metrics, whereas Integrated IS-IS
  gives the two types of routes equal preference.

o IDPR Policy

  A specific case of Policy. The IETF's Inter-domain Policy Routing
  Working Group is devising a routing protocol called Inter-Domain
  Policy Routing (IDPR) to support true policy-based routing in the
  Internet. Packets with certain combinations of header attributes
  (such as specific combinations of source and destination addresses
  or special IDPR source route options) are required to use routes
  provided by the IDPR protocol. Thus, unlike other Policy pruning
  rules, IDPR Policy would have to be applied before any other
  pruning rules except Basic Match.
  Specifically, IDPR Policy examines the packet being forwarded to
  ascertain if its attributes require that it be forwarded using
  policy-based routes. If so, IDPR Policy deletes all routes not
  provided by the IDPR protocol.

E.3 Some Route Lookup Algorithms

This section examines several route lookup algorithms that are in use or have been proposed. Each is described by giving the sequence of pruning rules it uses. The strengths and weaknesses of each algorithm are presented

E.3.1 The Revised Classic Algorithm

  The Revised Classic Algorithm is the form of the traditional
  algorithm which was discussed in Section [E.1].  The steps of this
  algorithm are:
  1.  Basic match
  2.  Longest match
  3.  Best metric
  4.  Policy
  Some implementations omit the Policy step, since it is needed only
  when routes may have metrics that are not comparable (because they
  were learned from different routing domains).
  The advantages of this algorithm are:
  (1)  It is widely implemented.
  (2)  Except for the Policy step (which an implementor can choose
       to make arbitrarily complex) the algorithm is simple both to
       understand and to implement.
  Its disadvantages are:
  (1)  It does not handle IS-IS or OSPF route classes, and therefore
       cannot be used for Integrated IS-IS or OSPF.
  (2)  It does not handle TOS or other path attributes.
  (3)  The policy mechanisms are not standardized in any way, and
       are therefore are often implementation-specific.  This causes
       extra work for implementors (who must invent appropriate
       policy mechanisms) and for users (who must learn how to use
       the mechanisms.  This lack of a standardized mechanism also
       makes it difficult to build consistent configurations for
       routers from different vendors.  This presents a significant
       practical deterrent to multi-vendor interoperability.
  (4)  The proprietary policy mechanisms currently provided by
       vendors are often inadequate in complex parts of the
       Internet.
  (5)  The algorithm has not been written down in any generally
       available document or standard.  It is, in effect, a part of
       the Internet Folklore.

E.3.2 The Variant Router Requirements Algorithm

  Some Router Requirements Working Group members have proposed a
  slight variant of the algorithm described in the Section
  [5.2.4.3].  In this variant, matching the type of service
  requested is considered to be more important, rather than less
  important, than matching as much of the destination address as
  possible.  For example, this algorithm would prefer a default
  route which had the correct type of service over a network route
  which had the default type of service, whereas the algorithm in
  [5.2.4.3] would make the opposite choice.
  The steps of the algorithm are:
  1.  Basic match
  2.  Weak TOS
  3.  Longest match
  4.  Best metric
  5.  Policy
  Debate between the proponents of this algorithm and the regular
  Router Requirements Algorithm suggests that each side can show
  cases where its algorithm leads to simpler, more intuitive routing
  than the other's algorithm does.  In general, this variant has the
  same set of advantages and disadvantages that the algorithm
  specified in [5.2.4.3] does, except that pruning on Weak TOS
  before pruning on Longest Match makes this algorithm less
  compatible with OSPF and Integrated IS-IS than the standard Router
  Requirements Algorithm.

E.3.3 The OSPF Algorithm

  OSPF uses an algorithm which is virtually identical to the Router
  Requirements Algorithm except for one crucial difference: OSPF
  considers OSPF route classes.
  The algorithm is:
  1.  Basic match
  2.  OSPF route class
  3.  Longest match
  4.  Weak TOS
  5.  Best metric
  6.  Policy
  Type of service support is not always present.  If it is not
  present then, of course, the fourth step would be omitted
  This algorithm has some advantages over the Revised Classic
  Algorithm:
  (1)  It supports type of service routing.
  (2)  Its rules are written down, rather than merely being a part
       of the Internet folklore.
  (3)  It (obviously) works with OSPF.
  However, this algorithm also retains some of the disadvantages of
  the Revised Classic Algorithm:
  (1)  Path properties other than type of service (e.g. MTU) are
       ignored.
  (2)  As in the Revised Classic Algorithm, the details (or even the
       existence) of the Policy step are left to the discretion of
       the implementor.
  The OSPF Algorithm also has a further disadvantage (which is not
  shared by the Revised Classic Algorithm).  OSPF internal (intra-
  area or inter-area) routes are always considered to be superior to
  routes learned from other routing protocols, even in cases where
  the OSPF route matches fewer bits of the destination address.
  This is a policy decision that is inappropriate in some networks.
  Finally, it is worth noting that the OSPF Algorithm's TOS support
  suffers from a deficiency in that routing protocols which support
  TOS are implicitly preferred when forwarding packets which have
  non-zero TOS values.  This may not be appropriate in some cases.

E.3.4 The Integrated IS-IS Algorithm

  Integrated IS-IS uses an algorithm which is similar to but not
  quite identical to the OSPF Algorithm.  Integrated IS-IS uses a
  different set of route classes, and also differs slightly in its
  handling of type of service.  The algorithm is:
  1. Basic Match
  2. IS-IS Route Classes
  3. Longest Match
  4. Weak TOS
  5. Best Metric
  6. Policy
  Although Integrated IS-IS uses Weak TOS, the protocol is only
  capable of carrying routes for a small specific subset of the
  possible values for the TOS field in the IP header.  Packets
  containing other values in the TOS field are routed using the
  default TOS.
  Type of service support is optional; if disabled, the fourth step
  would be omitted.  As in OSPF, the specification does not include
  the Policy step.
  This algorithm has some advantages over the Revised Classic
  Algorithm:
  (1)  It supports type of service routing.
  (2)  Its rules are written down, rather than merely being a part
       of the Internet folklore.
  (3)  It (obviously) works with Integrated IS-IS.
  However, this algorithm also retains some of the disadvantages of
  the Revised Classic Algorithm:
  (1)  Path properties other than type of service (e.g. MTU) are
       ignored.
  (2)  As in the Revised Classic Algorithm, the details (or even the
       existence) of the Policy step are left to the discretion of
       the implementor.
  (3)  It doesn't work with OSPF because of the differences between
       IS-IS route classes and OSPF route classes.  Also, because
       IS-IS supports only a subset of the possible TOS values, some
       obvious implementations of the Integrated IS-IS algorithm
       would not support OSPF's interpretation of TOS.
  The Integrated IS-IS Algorithm also has a further disadvantage
  (which is not shared by the Revised Classic Algorithm): IS-IS
  internal (intra-area or inter-area) routes are always considered
  to be superior to routes learned from other routing protocols,
  even in cases where the IS-IS route matches fewer bits of the
  destination address and doesn't provide the requested type of
  service.  This is a policy decision that may not be appropriate in
  all cases.
  Finally, it is worth noting that the Integrated IS-IS Algorithm's
  TOS support suffers from the same deficiency noted for the OSPF
  Algorithm.

Security Considerations

Although the focus of this document is interoperability rather than security, there are obviously many sections of this document which have some ramifications on network security.

Security means different things to different people. Security from a router's point of view is anything that helps to keep its own networks operational and in addition helps to keep the Internet as a whole healthy. For the purposes of this document, the security services we are concerned with are denial of service, integrity, and authentication as it applies to the first two. Privacy as a security service is important, but only peripherally a concern of a router - at least as of the date of this document.

In several places in this document there are sections entitled ... Security Considerations. These sections discuss specific considerations that apply to the general topic under discussion.

Rarely does this document say do this and your router/network will be secure. More likely, it says this is a good idea and if you do it, it

  • may* improve the security of the Internet and your local system in

general.

Unfortunately, this is the state-of-the-art AT THIS TIME. Few if any of the network protocols a router is concerned with have reasonable, built-in security features. Industry and the protocol designers have been and are continuing to struggle with these issues. There is progress, but only small baby steps such as the peer-to-peer authentication available in the BGP and OSPF routing protocols.

In particular, this document notes the current research into developing and enhancing network security. Specific areas of research, development, and engineering that are underway as of this writing (December 1993) are in IP Security, SNMP Security, and common authentication technologies.

Notwithstanding all of the above, there are things both vendors and users can do to improve the security of their router. Vendors should get a copy of Trusted Computer System Interpretation [INTRO:8]. Even if a vendor decides not to submit their device for formal verification under these guidelines, the publication provides excellent guidance on general security design and practices for computing devices.

Acknowledgments

O that we now had here But one ten thousand of those men in England That do no work to-day!

What's he that wishes so? My cousin Westmoreland? No, my fair cousin: If we are mark'd to die, we are enow To do our country loss; and if to live, The fewer men, the greater share of honour. God's will! I pray thee, wish not one man more. By Jove, I am not covetous for gold, Nor care I who doth feed upon my cost; It yearns me not if men my garments wear; Such outward things dwell not in my desires: But if it be a sin to covet honour, I am the most offending soul alive. No, faith, my coz, wish not a man from England: God's peace! I would not lose so great an honour As one man more, methinks, would share from me For the best hope I have. O, do not wish one more! Rather proclaim it, Westmoreland, through my host, That he which hath no stomach to this fight, Let him depart; his passport shall be made And crowns for convoy put into his purse: We would not die in that man's company That fears his fellowship to die with us. This day is called the feast of Crispian: He that outlives this day, and comes safe home, Will stand a tip-toe when the day is named, And rouse him at the name of Crispian. He that shall live this day, and see old age, Will yearly on the vigil feast his neighbours, And say 'To-morrow is Saint Crispian:' Then will he strip his sleeve and show his scars. And say 'These wounds I had on Crispin's day.' Old men forget: yet all shall be forgot, But he'll remember with advantages What feats he did that day: then shall our names. Familiar in his mouth as household words Harry the king, Bedford and Exeter, Warwick and Talbot, Salisbury and Gloucester, Be in their flowing cups freshly remember'd. This story shall the good man teach his son; And Crispin Crispian shall ne'er go by,

From this day to the ending of the world, But we in it shall be remember'd; We few, we happy few, we band of brothers; For he to-day that sheds his blood with me Shall be my brother; be he ne'er so vile, This day shall gentle his condition: And gentlemen in England now a-bed Shall think themselves accursed they were not here, And hold their manhoods cheap whiles any speaks That fought with us upon Saint Crispin's day.

This memo is a product of the IETF's Router Requirements Working Group. A memo such as this one is of necessity the work of many more people than could be listed here. A wide variety of vendors, network managers, and other experts from the Internet community graciously contributed their time and wisdom to improve the quality of this memo. The editor wishes to extend sincere thanks to all of them.

The current editor also wishes to single out and extend his heartfelt gratitude and appreciation to the original editor of this document; Philip Almquist. Without Philip's work, both as the original editor and as the Chair of the working group, this document would not have been produced.

Philip Almquist, Jeffrey Burgan, Frank Kastenholz, and Cathy Wittbrodt each wrote major chapters of this memo. Others who made major contributions to the document included Bill Barns, Steve Deering, Kent England, Jim Forster, Martin Gross, Jeff Honig, Steve Knowles, Yoni Malachi, Michael Reilly, and Walt Wimer.

Additional text came from Art Berggreen, John Cavanaugh, Ross Callon, John Lekashman, Brian Lloyd, Gary Malkin, Milo Medin, John Moy, Craig Partridge, Stephanie Price, Yakov Rekhter, Steve Senum, Richard Smith, Frank Solensky, Rich Woundy, and others who have been inadvertently overlooked.

Some of the text in this memo has been (shamelessly) plagiarized from earlier documents, most notably RFC-1122 by Bob Braden and the Host Requirements Working Group, and RFC-1009 by Bob Braden and Jon Postel. The work of these earlier authors is gratefully acknowledged.

Jim Forster was a co-chair of the Router Requirements Working Group during its early meetings, and was instrumental in getting the group off to a good start. Jon Postel, Bob Braden, and Walt Prue also contributed to the success by providing a wealth of good advice prior to the group's first meeting. Later on, Phill Gross, Vint Cerf, and Noel Chiappa all provided valuable advice and support.

Mike St. Johns coordinated the Working Group's interactions with the security community, and Frank Kastenholz coordinated the Working Group's interactions with the network management area. Allison Mankin and K.K. Ramakrishnan provided expertise on the issues of congestion control and resource allocation.

Many more people than could possibly be listed or credited here participated in the deliberations of the Router Requirements Working Group, either through electronic mail or by attending meetings. However, the efforts of Ross Callon and Vince Fuller in sorting out the difficult issues of route choice and route leaking are especially acknowledged.

The previous editor, Philip Almquist, wishes to extend his thanks and appreciation to his former employers, Stanford University and BARRNet, for allowing him to spend a large fraction (probably far more than they ever imagined when he started on this) of his time working on this project.

The current editor wishes to thank his employer, FTP Software, for allowing him to spend the time necessary to finish this document.

Editor's Address

The address of the current editor of this document is Frank J. Kastenholz FTP Software 2 High Street North Andover, MA, 01845-2620 USA

Phone: +1 508-685-4000

EMail: [email protected]