Network Working Group F. Baker Request for Comments: 2747 Cisco Category: Standards Track B. Lindell USC/ISI M. Talwar Microsoft January 2000 RSVP Cryptographic Authentication 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 (2000). All Rights Reserved.
AbstractThis document describes the format and use of RSVP's INTEGRITY object to provide hop-by-hop integrity and authentication of RSVP messages. 1] is a protocol for setting up distributed state in routers and hosts, and in particular for reserving resources to implement integrated service. RSVP allows particular users to obtain preferential access to network resources, under the control of an admission control mechanism. Permission to make a reservation will depend both upon the availability of the requested resources along the path of the data, and upon satisfaction of policy rules. To ensure the integrity of this admission control mechanism, RSVP requires the ability to protect its messages against corruption and spoofing. This document defines a mechanism to protect RSVP message integrity hop-by-hop. The proposed scheme transmits an authenticating digest of the message, computed using a secret Authentication Key and a keyed-hash algorithm. This scheme provides protection against forgery or message modification. The INTEGRITY object of each RSVP message is tagged with a one-time-use sequence
number. This allows the message receiver to identify playbacks and hence to thwart replay attacks. The proposed mechanism does not afford confidentiality, since messages stay in the clear; however, the mechanism is also exportable from most countries, which would be impossible were a privacy algorithm to be used. Note: this document uses the terms "sender" and "receiver" differently from . They are used here to refer to systems that face each other across an RSVP hop, the "sender" being the system generating RSVP messages. The message replay prevention algorithm is quite simple. The sender generates packets with monotonically increasing sequence numbers. In turn, the receiver only accepts packets that have a larger sequence number than the previous packet. To start this process, a receiver handshakes with the sender to get an initial sequence number. This memo discusses ways to relax the strictness of the in-order delivery of messages as well as techniques to generate monotonically increasing sequence numbers that are robust across sender failures and restarts. The proposed mechanism is independent of a specific cryptographic algorithm, but the document describes the use of Keyed-Hashing for Message Authentication using HMAC-MD5 . As noted in , there exist stronger hashes, such as HMAC-SHA1; where warranted, implementations will do well to make them available. However, in the general case,  suggests that HMAC-MD5 is adequate to the purpose at hand and has preferable performance characteristics.  also offers source code and test vectors for this algorithm, a boon to those who would test for interoperability. HMAC-MD5 is required as a baseline to be universally included in RSVP implementations providing cryptographic authentication, with other proposals optional (see Section 6 on Conformance Requirements). The RSVP checksum MAY be disabled (set to zero) when the INTEGRITY object is included in the message, as the message digest is a much stronger integrity check. 8]. 3,5], we would choose not to use it. This was discussed at length in the working group, and the use of IPSEC was rejected for the following reasons.
The security associations in IPSEC are based on destination address. It is not clear that RSVP messages are well defined for either source or destination based security associations, as a router must forward PATH and PATH TEAR messages using the same source address as the sender listed in the SENDER TEMPLATE. RSVP traffic may otherwise not follow exactly the same path as data traffic. Using either source or destination based associations would require opening a new security association among the routers for which a reservation traverses. In addition, it was noted that neighbor relationships between RSVP systems are not limited to those that face one another across a communication channel. RSVP relationships across non-RSVP clouds, such as those described in Section 2.9 of , are not necessarily visible to the sending system. These arguments suggest the use of a key management strategy based on RSVP router to RSVP router associations instead of IPSEC.
o Flags: An 8-bit field with the following format: Flags 0 1 2 3 4 5 6 7 +---+---+---+---+---+---+---+---+ | H | | | F | 0 | +---+---+---+---+---+---+---+---+ Currently only one flag (HF) is defined. The remaining flags are reserved for future use and MUST be set to 0. o Bit 0: Handshake Flag (HF) concerns the integrity handshake mechanism (Section 4.3). Message senders willing to respond to integrity handshake messages SHOULD set this flag to 1 whereas those that will reject integrity handshake messages SHOULD set this to 0. o Key Identifier: An unsigned 48-bit number that MUST be unique for a given sender. Locally unique Key Identifiers can be generated using some combination of the address (IP or MAC or LIH) of the sending interface and the key number. The combination of the Key Identifier and the sending system's IP address uniquely identifies the security association (Section 2.2). o Sequence Number: An unsigned 64-bit monotonically increasing, unique sequence number. Sequence Number values may be any monotonically increasing sequence that provides the INTEGRITY object [of each RSVP message] with a tag that is unique for the associated key's lifetime. Details on sequence number generation are presented in Section 3. o Keyed Message Digest: The digest MUST be a multiple of 4 octets long. For HMAC-MD5, it will be 16 bytes long.
o Authentication algorithm and algorithm mode being used. o Key used with the authentication algorithm. o Lifetime of the key. o Associated sending interface and other security association selection criteria [REQUIRED at Sending System]. o Source Address of the sending system [REQUIRED at Receiving System]. o Latest sending sequence number used with this key identifier [REQUIRED at Sending System]. o List of last N sequence numbers received with this key identifier [REQUIRED at Receiving System]. 9]. The sequence number field is chosen to be a 64-bit unsigned quantity. This is large enough to avoid exhaustion over the key lifetime. For example, if a key lifetime was conservatively defined as one year, there would be enough sequence number values to send RSVP messages at an average rate of about 585 gigaMessages per second. A 32-bit sequence number would limit this average rate to about 136 messages per second. The ability to generate unique monotonically increasing sequence numbers across a failure and restart implies some form of stable storage, either local to the device or remotely over the network. Three sequence number generation procedures are described below.
Section 4.3), arriving on a secured receiving interface contain the INTEGRITY object. If the INTEGRITY object is absent, the receiver discards the message. Security associations are simplex - the keys that a sending system uses to sign its messages may be different from the keys that its receivers use to sign theirs. Hence, each association is associated with a unique sending system and (possibly) multiple receiving systems. Each sender SHOULD have distinct security associations (and keys) per secured sending interface (or LIH). While administrators may configure all the routers and hosts on a subnet (or for that matter, in their network) using a single security association, implementations MUST assume that each sender may send using a distinct security association on each secured interface. At the sender, security association selection is based on the interface through which the message is sent. This selection MAY include additional criteria, such as the destination address (when sending the message unicast, over a broadcast LAN with a large number of hosts) or user identities at the sender or receivers . Finally, all intended message recipients should participate in this security association. Route flaps in a non RSVP cloud might cause messages for the same receiver to be sent on different interfaces at different times. In such cases, the receivers should participate in all possible security associations that may be selected for the interfaces through which the message might be sent. Receivers select keys based on the Key Identifier and the sending system's IP address. The Key Identifier is included in the INTEGRITY object. The sending system's address can be obtained either from the RSVP_HOP object, or if that's not present (as is the case with PathErr and ResvConf messages) from the IP source address. Since the Key Identifier is unique for a sender, this method uniquely identifies the key. The integrity mechanism slightly modifies the processing rules for RSVP messages, both when including the INTEGRITY object in a message sent over a secured sending interface and when accepting a message received on a secured receiving interface. These modifications are detailed below.
1], with these exceptions: (1) The RSVP checksum field is set to zero. If required, an RSVP checksum can be calculated when the processing of the INTEGRITY object is complete. (2) The INTEGRITY object is inserted in the appropriate place, and its location in the message is remembered for later use. (3) The sending interface and other appropriate criteria (as mentioned above) are used to determine the Authentication Key and the hash algorithm to be used. (4) The unused flags and the reserved field in the INTEGRITY object MUST be set to 0. The Handshake Flag (HF) should be set according to rules specified in Section 2.1. (5) The sending sequence number MUST be updated to ensure a unique, monotonically increasing number. It is then placed in the Sequence Number field of the INTEGRITY object. (6) The Keyed Message Digest field is set to zero. (7) The Key Identifier is placed into the INTEGRITY object. (8) An authenticating digest of the message is computed using the Authentication Key in conjunction with the keyed-hash algorithm. When the HMAC-MD5 algorithm is used, the hash calculation is described in . (9) The digest is written into the Cryptographic Digest field of the INTEGRITY object.
(3) The Key Identifier field and the sending system address are used to uniquely determine the Authentication Key and the hash algorithm to be used. Processing of this packet might be delayed when the Key Management System (Appendix 1) is queried for this information. (4) A new keyed-digest is calculated using the indicated algorithm and the Authentication Key. (5) If the calculated digest does not match the received digest, the message is discarded without further processing. (6) If the message is of type "Integrity Response", verify that the CHALLENGE object identically matches the originated challenge. If it matches, save the sequence number in the INTEGRITY object as the largest sequence number received to date. Otherwise, for all other RSVP Messages, the sequence number is validated to prevent replay attacks, and messages with invalid sequence numbers are ignored by the receiver. When a message is accepted, the sequence number of that message could update a stored value corresponding to the largest sequence number received to date. Each subsequent message must then have a larger (modulo 2^64) sequence number to be accepted. This simple processing rule prevents message replay attacks, but it must be modified to tolerate limited out-of-order message delivery. For example, if several messages were sent in a burst (in a periodic refresh generated by a router, or as a result of a tear down function), they might get reordered and then the sequence numbers would not be received in an increasing order. An implementation SHOULD allow administrative configuration that sets the receiver's tolerance to out-of-order message delivery. A simple approach would allow administrators to specify a message window corresponding to the worst case reordering behavior. For example, one might specify that packets reordered within a 32 message window would be accepted. If no reordering can occur, the window is set to one. The receiver must store a list of all sequence numbers seen within the reordering window. A received sequence number is valid if (a) it is greater than the maximum sequence number received or (b) it is a past sequence number lying within the reordering window and not recorded in the list. Acceptance of
a sequence number implies adding it to the list and removing a number from the lower end of the list. Messages received with sequence numbers lying below the lower end of the list or marked seen in the list are discarded. When an "Integrity Challenge" message is received on a secured sending interface it is processed in the following manner: (1) An "Integrity Response" message is formed using the Challenge object received in the challenge message. (2) The message is sent back to the receiver, based on the source IP address of the challenge message, using the "Message Generation" steps outlined above. The selection of the Authentication Key and the hash algorithm to be used is determined by the key identifier supplied in the challenge message. Section 2.5.3 of ). It is suggested that the cookie be an MD5 hash of a local secret and a timestamp to provide uniqueness (see Section 9). An RSVP Integrity Challenge message will carry a message type of 11. The message format is as follows: <Integrity Challenge message> ::= <Common Header> <CHALLENGE>
he CHALLENGE object has the following format: CHALLENGE Object: Class = 64, C-Type = 1 +-------------+-------------+-------------+-------------+ | 0 (Reserved) | | +-------------+-------------+ + | Key Identifier | +-------------+-------------+-------------+-------------+ | Challenge Cookie | | | +-------------+-------------+-------------+-------------+ The sender accepts the "Integrity Challenge" without doing an integrity check. It returns an RSVP "Integrity Response" message that contains the original CHALLENGE object. It also includes an INTEGRITY object, signed with the key specified by the Key Identifier included in the "Integrity Challenge". An RSVP Integrity Response message will carry a message type of 12. The message format is as follows: <Integrity Response message> ::= <Common Header> <INTEGRITY> <CHALLENGE> The "Integrity Response" message is accepted by the receiver (challenger) only if the returned CHALLENGE object matches the one sent in the "Integrity Challenge" message. This prevents replay of old "Integrity Response" messages. If the match is successful, the receiver saves the Sequence Number from the INTEGRITY object as the latest sequence number received with the key identifier included in the CHALLENGE. If a response is not received within a given period of time, the challenge is repeated. When the integrity handshake successfully completes, the receiver begins accepting normal RSVP signaling messages from that sender and ignores any other "Integrity Response" messages. The Handshake Flag (HF) is used to allow implementations the flexibility of not including the integrity handshake mechanism. By setting this flag to 1, message senders that implement the integrity handshake distinguish themselves from those that do not. Receivers SHOULD NOT attempt to handshake with senders whose INTEGRITY object has HF = 0.
An integrity handshake may not be necessary in all environments. A common use of RSVP integrity will be between peering domain routers, which are likely to be processing a steady stream of RSVP messages due to aggregation effects. When a router restarts after a crash, valid RSVP messages from peering senders will probably arrive within a short time. Assuming that replay messages are injected into the stream of valid RSVP messages, there may be only a small window of opportunity for a replay attack before a valid message is processed. This valid message will set the largest sequence number seen to a value greater than any number that had been stored prior to the crash, preventing any further replays. On the other hand, not using an integrity handshake could allow exposure to replay attacks if there is a long period of silence from a given sender following a restart of a receiver. Hence, it SHOULD be an administrative decision whether or not the receiver performs an integrity handshake with senders that are willing to respond to "Integrity Challenge" messages, and whether it accepts any messages from senders that refuse to do so. These decisions will be based on assumptions related to a particular network environment.
In general, no key is ever used outside its lifetime (but see Section 5.3). Possible mechanisms for managing key lifetime include the Network Time Protocol and hardware time-of-day clocks. To maintain security, it is advisable to change the RSVP Authentication Key on a regular basis. It should be possible to switch the RSVP Authentication Key without loss of RSVP state or denial of reservation service, and without requiring people to change all the keys at once. This requires an RSVP implementation to support the storage and use of more than one active RSVP Authentication Key at the same time. Hence both the sender and receivers might have multiple active keys for a given security association. Since keys are shared between a sender and (possibly) multiple receivers, there is a region of uncertainty around the time of key switch-over during which some systems may still be using the old key and others might have switched to the new key. The size of this uncertainty region is related to clock synchrony of the systems. Administrators should configure the overlap between the expiration time of the old key (KeyEndValid) and the validity of the new key (KeyStartValid) to be at least twice the size of this uncertainty interval. This will allow the sender to make the key switch-over at the midpoint of this interval and be confident that all receivers are now accepting the new key. For the duration of the overlap in key lifetimes, a receiver must be prepared to authenticate messages using either key. During a key switch-over, it will be necessary for each receiver to handshake with the sender using the new key. As stated before, a receiver has the choice of initiating a handshake during the switchover or postponing the handshake until the receipt of a message using that key.
o An implementation MUST associate a specific lifetime (i.e., KeyStartValid and KeyEndValid) with each key and the corresponding Key Identifier. o An implementation MUST support manual key distribution (e.g., the privileged user manually typing in the key, key lifetime, and key identifier on the console). The lifetime may be infinite. o If more than one algorithm is supported, then the implementation MUST require that the algorithm be specified for each key at the time the other key information is entered. o Keys that are out of date MAY be automatically deleted by the implementation. o Manual deletion of active keys MUST also be supported. o Key storage SHOULD persist across a system restart, warm or cold, to ease operational usage. 7] MUST be implemented by all conforming implementations. A conforming implementation MAY also support other authentication algorithms such as NIST's Secure Hash Algorithm (SHA). Manual key distribution as described above MUST be supported by all conforming implementations. All implementations MUST support the smooth key roll over described under "Key Management Procedures." Implementations SHOULD support a standard key management protocol for secure distribution of RSVP Authentication Keys once such a key management protocol is standardized by the IETF.
10] MAY be used to generate the RSVP Authentication key used in generating a signature in the Integrity Object sent from a RSVP sender to a receiver. Kerberos key generation avoids the use of shared keys between RSVP senders and receivers such as hosts and routers. Kerberos allows for the use of trusted third party keying relationships between security principals (RSVP sender and receivers) where the Kerberos key distribution center(KDC) establishes an ephemeral session key that is subsequently shared between RSVP sender and receivers. In the multicast case all receivers of a multicast RSVP message MUST share a single key with the KDC (e.g. the receivers are in effect the same security principal with respect to Kerberos). The Key information determined by the sender MAY specify the use of Kerberos in place of configured shared keys as the mechanism for establishing a key between the sender and receiver. The Kerberos identity of the receiver is established as part of the sender's interface configuration or it can be established through other mechanisms. When generating the first RSVP message for a specific key identifier the sender requests a Kerberos service ticket and gets back an ephemeral session key and a Kerberos ticket from the KDC. The sender encapsulates the ticket and the identity of the sender in an Identity Policy Object. The sender includes the Policy Object in the RSVP message. The session key is then used by the sender as the RSVP Authentication key in section 4.1 step (3) and is stored as Key information associated with the key identifier. Upon RSVP Message reception, the receiver retrieves the Kerberos Ticket from the Identity Policy Object, decrypts the ticket and retrieves the session key from the ticket. The session key is the same key as used by the sender and is used as the key in section 4.2 step (3). The receiver stores the key for use in processing subsequent RSVP messages. Kerberos tickets have lifetimes and the sender MUST NOT use tickets that have expired. A new ticket MUST be requested and used by the sender for the receiver prior to the ticket expiring.
retrieve the cached key (and optionally other identity information) instead of passing tickets from sender to receiver in each RSVP message. A receiver may not have cached key state with an associated Key Identifier due to reboot or route changes. If the receiver's policy indicates the use of Kerberos keys for integrity checking, the receiver can send an integrity Challenge message back to the sender. Upon receiving an integrity Challenge message a sender MUST send an Identity object that includes the Kerberos ticket in the integrity Response message, thereby allowing the receiver to retrieve and store the session key from the Kerberos ticket for subsequent Integrity checking.  Braden, R., Zhang, L., Berson, S., Herzog, S. and S. Jamin, "Resource ReSerVation Protocol (RSVP) -- Version 1 Functional Specification", RFC 2205, September 1997.  Yadav, S., et al., "Identity Representation for RSVP", RFC 2752, January 2000.  Atkinson, R. and S. Kent, "Security Architecture for the Internet Protocol", RFC 2401, November 1998.  Maughan, D., Schertler, M., Schneider, M. and J. Turner, "Internet Security Association and Key Management Protocol (ISAKMP)", RFC 2408, November 1998.  Kent, S. and R. Atkinson, "IP Authentication Header", RFC 2402, November 1998.  Kent, S. and R. Atkinson, "IP Encapsulating Security Payload (ESP)", RFC 2406, November 1998.  Krawczyk, H., Bellare, M. and R. Canetti, "HMAC: Keyed-Hashing for Message Authentication", RFC 2104, March 1996.
Confidentiality is not provided by this mechanism. If confidentiality is required, IPSEC ESP  may be the best approach, although it is subject to the same criticisms as IPSEC Authentication, and therefore would be applicable only in specific environments. Protection against traffic analysis is also not provided. Mechanisms such as bulk link encryption might be used when protection against traffic analysis is required.
Acknowledgement Funding for the RFC Editor function is currently provided by the Internet Society.