RFC2478
Network Working Group E. Baize Request for Comments: 2478 D. Pinkas Category: Standards Track Bull
December 1998
The Simple and Protected GSS-API Negotiation Mechanism
Status of this Memo
This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (1998). All Rights Reserved.
Contents
ABSTRACT
This document specifies a Security Negotiation Mechanism for the Generic Security Service Application Program Interface (GSS-API) which is described in [1].
The GSS-API provides a generic interface which can be layered atop different security mechanisms such that if communicating peers acquire GSS-API credentials for the same security mechanism, then a security context may be established between them (subject to policy). However, GSS-API doesn't prescribe the method by which GSS-API peers can establish whether they have a common security mechanism.
The Simple and Protected GSS-API Negotiation Mechanism defined here is a pseudo-security mechanism, represented by the object identifier iso.org.dod.internet.security.mechanism.snego (1.3.6.1.5.5.2) which enables GSS-API peers to determine in-band whether their credentials share common GSS-API security mechanism(s), and if so, to invoke normal security context establishment for a selected common security mechanism. This is most useful for applications that are based on GSS-API implementations which support multiple security mechanisms.
This allows to negotiate different security mechanisms, different options within a given security mechanism or different options from several security mechanisms.
Once the common security mechanism is identified, the security mechanism may also negotiate mechanism-specific options during its context establishment. This will be inside the mechanism tokens, and invisible to the SPNEGO protocol.
The simple and protected GSS-API mechanism negotiation is based on the following negotiation model : the initiator proposes one security mechanism or an ordered list of security mechanisms, the target either accepts the proposed security mechanism, or chooses one from an offered set, or rejects the proposed value(s). The target then informs the initiator of its choice.
In its basic form this protocol requires an extra-round trip. Network connection setup is a critical performance characteristic of any network infrastructure and extra round trips over WAN links, packet radio networks, etc. really make a difference. In order to avoid such an extra round trip the initial security token of the preferred mechanism for the initiator may be embedded in the initial token. If the target preferred mechanism matches the initiator's preferred mechanism, no additional round trips are incurred by using the negotiation protocol.
The simple and protected GSS-API mechanism negotiation provides a technique to protect the negotiation that must be used when the underlying mechanism selected by the target is capable of integrity protection.
When all the mechanisms proposed by the initiator support integrity protection or when the selected mechanism supports integrity protection, then the negotiation mechanism becomes protected since this guarantees that the appropriate mechanism supported by both peers has been selected.
The Simple and Protected GSS-API Negotiation Mechanism uses the concepts developed in the GSS-API specification [1]. The negotiation data is encapsulated in context-level tokens. Therefore, callers of the GSS-API do not need to be aware of the existence of the negotiation tokens but only of the new pseudo-security mechanism. A failure in the negotiation phase causes a major status code to be returned: GSS_S_BAD_MECH.
NEGOTIATION MODEL
Negotiation description
The model for security mechanism negotiation reuses a subset of the concepts specified in [2].
Each OID represents one GSS-API mechanism or one variant of it.
- When one security mechanism is proposed by the initiator, it represents the only security mechanism supported or selected (when the additional APIs defined in the Annex A are used) by the initiator.
- When several security mechanisms are proposed by the initiator, they represent a set of security mechanisms supported or selected (when the additional APIs defined in the Annex A are used) by the initiator.
The first negotiation token sent by the initiator contains an ordered list of mechanisms, a set of options (e.g. deleg, replay, conf flags) that should be supported by the selected mechanism and optionally the initial security token for the desired mechanism of the initiator (i.e. the first of the list).
The first negotiation token sent by the target contains the result of the negotiation (accept_completed, accept_incomplete or reject) and, in case of accept, the agreed security mechanism. It may also include the response to the initial security token from the initiator, when the first proposed mechanism of the initiator has been selected. When the first mechanism is acceptable to the target,it should respond to the initial security token for the desired mechanism of the initiator when it is present. However, if this is not possible, the target can simply ignore it and omit the responseToken from the first reply.
Implementations that can piggyback the initial token will be rewarded by faster connection setup.
In case of a successful negotiation, the security mechanism represents the value suitable for the target, and picked up from the list offered by the initiator. The policy by which the target chooses a mechanism is an implementation-specific local matter. In the absence of other policy, the target should chose the first mechanism in the list for which valid credentials are available.
Once a mechanism has been selected, the tokens specific to the selected mechanism are carried within the negotiation tokens (in the mechToken for the initiator and in the responseToken for the target).
Negotiation procedure
The negotiation procedure is summarised as follows:
(a) the GSS-API initiator invokes GSS_Init_sec_context as normal, but
requests (either explicitly, with the negotiation mechanism, or through accepting a default, when the default is the negotiation mechanism) that the Simple and Protected GSS-API Negotiation Mechanism be used;
(b) the initiator GSS-API implementation emits a negotiation token
containing a list of supported security mechanisms for the credentials used for this context establishment, and optionally an initial security token for the first mechanism from that list (i.e. the preferred mechanism), and indicates GSS_S_CONTINUE_NEEDED status;
(c) The GSS-API initiator sends the token to the target application;
(d) The GSS-API target deposits the token through invoking
GSS_Accept_sec_context. The target GSS-API implementation emits a negotiation token containing which if any of the proposed mechanisms it supports (or has selected).
If the mechanism selected by the target matches the preferred mechanism identified by the initiator and the initiator provides a mechToken, the negotiation token response may contain also an initial security token from that mechanism.
If the preferred mechanism is accepted, GSS_Accept_sec_context() indicates GSS_S_COMPLETE when unilateral or mutual authentication has been performed and involves a single token in either direction.
If a proposed mechanism is accepted, and it was not the preferred mechanism, or if the first negotiation token sent by the initiator did not included a mechToken, then the negotiation token response sent by the target may contain also a response token from that mechanism which transmits mechanism-specific information (e.g. to transmit a certificate). The initiator may ignore such an initial token if it is not prepared to process it.
If a proposed mechanism other than the preferred mechanism is accepted, or the preferred mechanism is accepted but involves multiple exchanges (e.g. challenge-response authentication), then GSS_Accept_sec_context() indicates GSS_S_CONTINUE_NEEDED status.
If the proposed mechanism(s) are rejected, GSS_Accept_sec_context() indicates GSS_S_BAD_MECH status. The security context initialisation has failed.
(e) The GSS-API target returns the token to the initiator
application;
(f) The GSS-API initiator deposits the token through invoking
GSS_Init_sec_context.
GSS_Init_sec_context() may then indicate GSS_S_CONTINUE_NEEDED, GSS_S_COMPLETE or GSS_S_BAD_MECH status.
The GSS_S_BAD_MECH status is returned when the negotiation token carries a reject result or when the negotiation token carries an accept result and the mechanism selected by the target is not included in the initial list sent by the initiator.
The GSS_S_BAD_MIC status is returned when the selected mechanism supports a MIC token but the MIC computed over the list of mechanisms sent by the initiator is missing or incorrect.
If the negotiation token carries a reject result, the context establishment is impossible. For example, a rejection will occur if the target doesn't support the initiator's proposed mechanism type(s). Upon failure of the mechanism negotiation procedure, the mech_type output parameter value is the negotiation mechanism type.
The GSS_S_CONTINUE_NEEDED status is returned when the negotiation token carries an accept result and further tokens must be transferred in order to complete context establishment for the selected mechanism. In that case GSS_Init_sec_context() returns an initial context token as output_token (with the selected mechanism's context token encapsulated within that output_token).
The initiator then sends the output_token to the target. The security context initialisation is then continued according to the standard GSS-API conventions for the selected mechanism, where the tokens of the selected mechanism are encapsulated until the GSS_S_COMPLETE is returned for both the initiator and the target. When GSS_S_CONTINUE_NEEDED is returned, the mech_type output parameter is not yet valid.
When GSS_S_COMPLETE is returned, the mech_type output parameter indicates the selected mechanism. When the final negotiation token does not contain a MIC, the initiator GSS-API implementation must check the returned/selected mechanism is on the originally
submitted list of mechanisms and also verify that the selected mechanism is not able to support a MIC. When the final negotiation token contains a MIC over the initial mechanisms list sent by the initiator, the MIC must be verified.
Note that the *_req_flag input parameters for context establishment are relative to the selected mechanism, as are the *_state output parameters. i.e., these parameters are not applicable to the negotiation process per se.
The initiator GSS-API calling application may need to know when the negotiation exchanges were protected or not. For this, when GSS_S_COMPLETE is returned, it can simply test the integ_avail flag. When this flag is set it indicates that the negotiation was protected.
On receipt of a negotiation token on the target side, a GSS-API implementation that does not support negotiation would indicate the GSS_S_BAD_MECH status as if a particular basic security mechanism had been requested but was not supported.
When GSS_Acquire_cred is invoked with the negotiation mechanism as desired_mechs, an implementation-specific default credential is used to carry on the negotiation. A set of mechanisms as specified locally by the system administrator is then available for negotiation. If there is a desire for the caller to make its own choice, then an additional API has to be used (see Appendix A).
DATA ELEMENTS
Mechanism Type
MechType::= OBJECT IDENTIFIER
mechType
Each security mechanism is as defined in [1].
Negotiation Tokens
The syntax of the negotiation tokens follows the InitialContextToken syntax defined in [1]. The security mechanism of the initial negotiation token is identified by the Object Identifier iso.org.dod.internet.security.mechanism.snego (1.3.6.1.5.5.2).
Syntax
This section specifies the syntax of the corresponding "innerContextToken" field for the first token and subsequent negotiation tokens. During the mechanism negociation, the "innerContextToken" field contains the ASN.1 structure "NegociationToken" given below, encoded using the DER encoding conventions.
NegotiationToken ::= CHOICE {
negTokenInit [0] NegTokenInit,
negTokenTarg [1] NegTokenTarg }
MechTypeList ::= SEQUENCE OF MechType
NegTokenInit ::= SEQUENCE {
mechTypes [0] MechTypeList OPTIONAL,
reqFlags [1] ContextFlags OPTIONAL,
mechToken [2] OCTET STRING OPTIONAL,
mechListMIC [3] OCTET STRING OPTIONAL
}
ContextFlags ::= BIT STRING {
delegFlag (0),
mutualFlag (1),
replayFlag (2),
sequenceFlag (3),
anonFlag (4),
confFlag (5),
integFlag (6)
}
negTokenInit
Negotiation token sent by the initiator to the target, which contains, for the first token sent, one or more security mechanisms supported by the initiator (as indicated in the field mechTypes) and the service options (reqFlags) that are requested to establish the context. The context flags should be filled in from the req_flags parameter of init_sec_context().
The mechToken field is optional for the first token sent that all target implementations would not have to support. However for those targets that do support piggybacking the initial mechToken, an optimistic negotiation response is possible. Otherwise the mechToken is used to carry the tokens specific to the mechanism selected.
The mechListMIC is an optional field. In the case that the chosen mechanism supports integrity, the initiator may optionally include a mechListMIC which is the result of a GetMIC of the MechTypes in the initial NegTokenInit and return GSS_S_COMPLETE.
When the chosen mechanism uses an odd number of messages, the final mechanism token will be sent from the initiator to the acceptor. In this case, there is a tradeoff between using the optimal number of messages, or using an additional message from the acceptor to the initiator in order to give the initiator assurance that no modification of the initiator's mechanism list occurred. The implementation can choose which tradeoff to make (see section 4.2.2 for further details for the processing of that field).
NegTokenTarg ::= SEQUENCE {
negResult [0] ENUMERATED {
accept_completed (0),
accept_incomplete (1),
reject (2) } OPTIONAL,
supportedMech [1] MechType OPTIONAL,
responseToken [2] OCTET STRING OPTIONAL,
mechListMIC [3] OCTET STRING OPTIONAL
}
negTokenTarg
Negotiation token returned by the target to the initiator which contains, for the first token returned, a global negotiation result and the security mechanism selected (if any).
negResult
The result accept_completed indicates that a context has been successfully established, while the result accept_incomplete indicates that additional token exchanges are needed.
Note: For the case where (a) a single-token context setup is
used and (b) the preferred mechanism does not support the
integrity facility which would cause a mechListMIC to be
generated and enclosed, this feature allows to make a
difference between a mechToken sent by the initiator but not
processed by the target (accept_incomplete) and a mechToken
sent by the initiator and processed by the target
(accept_completed).
For those targets that support piggybacking the initial mechToken, an optimistic negotiation response is possible and includes in that case a responseToken which may continue the authentication exchange (e.g. when mutual authentication has been requested or when unilateral authentication requires several round trips). Otherwise
the responseToken is used to carry the tokens specific to the mechanism selected. For subsequent tokens (if any) returned by the target, negResult, and supportedMech are not present.
For the last token returned by the target, the mechListMIC, when present, is a MIC computed over the MechTypes using the selected mechanism.
negResult
Result of the negotiation exchange, specified by the target.
This can be either :
accept_completed
The target accepts the preferred security mechanism,
and the context is established for the target or,
accept_incomplete
The target accepts one of the proposed security
mechanisms and further exchanges are necessary, or,
reject
The target rejects all the proposed security
mechanisms.
supportedMech
This field has to be present when negResult is "accept_completed" or "accept_incomplete". It is a choice from the mechanisms offered by the initiator.
responseToken
This field may be used either to transmit the response to the mechToken when sent by the initiator and when the first mechanism from the list has been selected by the target or to carry the tokens specific to the selected security mechanism.
mechListMIC
If the selected mechanism is capable of integrity protection, this field must be present in the last message of the negotiation, (i.e., when the underlying mechanism returns a non-empty token and a major status of GSS_S_COMPLETE); it contains the result of a GetMIC of the MechTypes field in the initial NegTokenInit. It allows to verify that the list initially sent by the initiator has been received unmodified by the target.
Processing of mechListMIC.
If the mechanism selected by the negotiation does not support integrity, then no mechListMIC is included, otherwise a mechListMIC must be used and validated as indicated below.
If the mechanism supports integrity and uses an even number of messages, then the target must compute a MIC as described above, and send this in the final NegTokenTarg along with the final mechToken. The initiator when receiving the last token must require that the mechListMIC field be present and valid. In the absence of a valid mechListMIC, the negotiation must fail as if the last context establishment token was invalid.
In the case that the chosen mechanism supports integrity and uses an odd number of messages, the final mechanism token will be sent from the initiator to the target. In this case, there is a tradeoff between using the optimal number of messages, or using an additional message from the target to the initiator in order to give the initiator assurance that no modification of the initiator's mechanism list occurred. The implementation can choose which tradeoff to make.
When generating the final NegTokenInit message, the NegTokenInit may optionally include a mechListMIC which is the result of a GetMIC of the MechTypes in the initial NegTokenInit and return GSS_S_COMPLETE. The target must check the presence of the MIC computed over the mechList sent in the initial NegTokenInit. Three cases may then be considered:
1) If the mechListMIC is present and correct, then
GSS_S_COMPLETE is returned to the target with no token; the
context is established by the target.
2) If the mechListMIC is present but invalid, then the context
establishment must fail. An error major status code is
returned to the target.
3) If the mechListMIC is not included in the final NegTokenInit,
then GSS_S_COMPLETE must be returned to the target with a
token. This token must be a NegTokenTarg, with a MIC included
as described above, and no responseToken. The application will
then send this token back to the initiator, which must verify
that the mechListMIC field is present and valid.
Note: If the MIC was originally sent by the initiator, but
thenafter deleted by an attacker, the target will send
back a token according to the description above, but the
initiator will be unable to process that returned token
and the context establishment must then fail.
EXAMPLES : SECURITY MECHANISM NEGOTIATION
Here are some examples of security mechanism negotiation between an initiator (I) and a target (T).
Initial steps
(I) supports two security mechanism types (GSS-MECH1 and GSS-MECH2).
(I) invokes GSS_Init_sec_context() with :
Input
mech_type = OID for negotiation mechanism or NULL, if the negotiation mechanism is the default mechanism.
Output
major_status = GSS_S_CONTINUE_NEEDED output_token = negTokenInit
The negotiation token (negTokenInit) contains two security mechanisms with :
mechType = GSS-MECH1 or mechType = GSS-MECH2
(I) sends to (T) the negotiation token.
Successful negotiation steps
(T) supports GSS-MECH2 (T) receives the negotiation token (negTokenInit) from (I) (T) invokes GSS_Accept_sec_context() with :
Input
input_token = negTokenInit
Output
major_status = GSS_S_CONTINUE_NEEDED
output_token = negTokenTarg
The negotiation token (negTokenTarg) contains :
negResult = accept (the negotiation result)
supportedMech : mechType = GSS-MECH2
(T) returns the negotiation token (negTokenTarg) to (I) (I) invokes GSS_Init_sec_context() with :
Input
input_token = negTokenTarg
Output
major_status = GSS_S_COMPLETE
output_token = initialContextToken (initial context token
for GSS-MECH2)
mech_type = GSS-MECH2
The subsequent steps are security mechanism specific, and work as specified in [1]. The output tokens from the security mechanism are encapsulated in a NegTokenTarg message (with the supportedMech field omitted, and the mechListMIC included with the last token).
Failed negotiation steps
(T) supports GSS-MECH3. (T) receives the negotiation token (negTokenInit) from (I) (T) invokes GSS_Accept_sec_context() with :
Input
input_token = negTokenInit
Output
major_status = GSS_S_BAD_MECH
output_token = negTokenTarg
The negotiation token (negTokenTarg) contains :
negResult = reject (the negotiation result)
(T) returns the negotiation token (negTokenTarg) to (I) (I) invokes GSS_Init_sec_context() with :
Input
input_token = negTokenTarg
Output
major_status = GSS_S_BAD_MECH
The security context establishment has failed.
Successful Negotiation with preferred mechanism info
(I) supports two security mechanism types (GSS-MECH1 and GSS-MECH2).
(I) invokes GSS_Init_sec_context() with :
Input
mech_type = OID for negotiation mechanism or NULL, if the
negotiation mechanism is the default mechanism.
Output
major_status = GSS_S_CONTINUE_NEEDED
output_token = negTokenInit
The negotiation token (negTokenInit) contains two security mechanisms with :
mechType = GSS-MECH1 or
mechType = GSS-MECH2
mechToken = output_token from GSS_Init_sec_context ( first mechType) as described in [1]
(I) sends to (T) the negotiation token.
(T) supports GSS-MECH1. (T) receives the negotiation token (negTokenInit) from (I) (T) invokes GSS_Accept_sec_context() with :
Input
input_token = negTokenInit
Output
major_status = GSS_S_CONTINUE_NEEDED
output_token = negTokenTarg
The negotiation token (negTokenTarg) contains :
negResult = accept (the negotiation result)
supportedMech : mechType = GSS-MECH1
mechToken = output_token from
GSS_Accept_sec_context(mechToken )
(T) returns the negotiation token (negTokenTarg) to (I) (I) invokes GSS_Init_sec_context() with :
Input
input_token = negTokenTarg
Output
major_status = GSS_S_COMPLETE or GSS_S_CONTINUE_NEEDED as needed
output_token = ContextToken (initial or subsequent context token
for GSS-MECH1)
mech_type = GSS-MECH1
Specific implementations of the protocol can support the optimistic negotiation by completing the security context establishment using the agreed upon mechanism as described in [1]. As described above in section 5.2, the output tokens from the security mechanisms are encapsulated in a NegTokenTarg message (with the negResult and supportedMech fields omitted, and the mechListMIC included with the last token).
SECURITY CONSIDERATIONS
When the mechanism selected by the target from the list supplied by the initiator supports integrity protection, then the negotiation is protected.
When one of the mechanisms proposed by the initiator does not support integrity protection, then the negotiation is exposed to all threats a non secured service is exposed. In particular, an active attacker can force to use a security mechanism which is not the common preferred one (when multiple security mechanisms are shared between peers) but which is acceptable anyway to the target.
In any case, the communicating peers may be exposed to the denial of service threat.
ACKNOWLEDGMENTS
Acknowledgments are due to Stephen Farrell of SSE, Marc Horowitz of Stonecast, John Linn of RSA Laboratories, Piers McMahon of Platinum Technology, Tom Parker of ICL and Doug Rosenthal of EINet, for reviewing earlier versions of this document and for providing useful inputs. Acknowledgments are also due to Peter Brundrett of Microsoft for his proposal for an optimistic negotiation, and for Bill Sommerfeld of Epilogue Technology for his proposal for protecting the negotiation.
APPENDIX A
GSS-API NEGOTIATION SUPPORT API
In order to provide to a GSS-API caller (either the initiator or the target or both) the ability to choose among the set of supported mechanisms a reduced set of mechanisms for negotiation, two additional APIs are defined:
GSS_Get_neg_mechs() indicates the set of security mechanisms available on the local system to the caller for negotiation.
GSS_Set_neg_mechs() specifies the set of security mechanisms to be used on the local system by the caller for negotiation.
A.1. GSS_Set_neg_mechs call
Input:
cred_handle CREDENTIAL HANDLE
- NULL specifies default credentials
mech_set SET OF OBJECT IDENTIFIER
Outputs:
major_status INTEGER,
minor_status INTEGER,
Return major_status codes :
GSS_S_COMPLETE indicates that the set of security mechanisms available for negotiation has been set to mech_set. GSS_S_FAILURE indicates that the requested operation could not be performed for reasons unspecified at the GSS-API level.
Allows callers to specify the set of security mechanisms that may be negotiated with the credential identified by cred_handle. This call is intended for support of specialised callers who need to restrict the set of negotiable security mechanisms from the set of all security mechanisms available to the caller (based on available credentials). Note that if more than one mechanism is specified in mech_set, the order in which those mechanisms are specified implies a relative mechanism preference for the target.
A.2. GSS_Get_neg_mechs call
Input:
cred_handle CREDENTIAL HANDLE
- NULL specifies default credentials
Outputs:
major_status INTEGER,
minor_status INTEGER,
mech_set SET OF OBJECT IDENTIFIER
Return major_status codes :
GSS_S_COMPLETE indicates that the set of security mechanisms
available for negotiation has been returned in
mech_option_set.
GSS_S_FAILURE indicates that the requested operation could not
be performed for reasons unspecified at the GSS-API level.
Allows callers to determine the set of security mechanisms available for negotiation with the credential identified by cred_handle. This call is intended for support of specialised callers who need to reduce the set of negotiable security mechanisms from the set of supported security mechanisms available to the caller (based on available credentials).
Note: The GSS_Indicate_mechs() function indicates the full set of mechanism types available on the local system. Since this call has no input parameter, the returned set is not necessarily available for all credentials.
REFERENCES
[1] Linn, J., "Generic Security Service Application Program
Interface", RFC 2078, January 1997.
[2] Standard ECMA-206, "Association Context Management including
Security Context Management", December 1993. Available on http://www.ecma.ch
AUTHORS' ADDRESSES
Eric Baize Bull - 300 Concord Road Billerica, MA 01821 - USA
Phone: +1 978 294 61 37 Fax: +1 978 294 61 09 EMail: [email protected]
Denis Pinkas Bull Rue Jean-Jaures BP 68 78340 Les Clayes-sous-Bois - FRANCE
Phone: +33 1 30 80 34 87 Fax: +33 1 30 80 33 21 EMail: [email protected]
Full Copyright Statement
Copyright (C) The Internet Society (1998). All Rights Reserved.
This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English.
The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns.
This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
