RFC 4724






Network Working Group                                          S. Sangli
Request for Comments: 4724                                       E. Chen
Category: Standards Track                                  Cisco Systems
                                                             R. Fernando
                                                              J. Scudder
                                                              Y. Rekhter
                                                        Juniper Networks
                                                            January 2007


                   Graceful Restart Mechanism for BGP

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 IETF Trust (2007).

Abstract



   This document describes a mechanism for BGP that would help minimize
   the negative effects on routing caused by BGP restart.  An End-of-RIB
   marker is specified and can be used to convey routing convergence
   information.  A new BGP capability, termed "Graceful Restart
   Capability", is defined that would allow a BGP speaker to express its
   ability to preserve forwarding state during BGP restart.  Finally,
   procedures are outlined for temporarily retaining routing information
   across a TCP session termination/re-establishment.

   The mechanisms described in this document are applicable to all
   routers, both those with the ability to preserve forwarding state
   during BGP restart and those without (although the latter need to
   implement only a subset of the mechanisms described in this
   document).











Sangli, et al.              Standards Track                     [Page 1]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


Table of Contents



   1. Introduction ....................................................2
      1.1. Specification of Requirements ..............................2
   2. Marker for End-of-RIB ...........................................3
   3. Graceful Restart Capability .....................................3
   4. Operation .......................................................6
      4.1. Procedures for the Restarting Speaker ......................6
      4.2. Procedures for the Receiving Speaker .......................7
   5. Changes to BGP Finite State Machine .............................9
   6. Deployment Considerations ......................................11
   7. Security Considerations ........................................12
   8. Acknowledgments ................................................13
   9. IANA Considerations ............................................13
   10. References ....................................................13
      10.1. Normative References .....................................13
      10.2. Informative References ...................................13

1.  Introduction



   Usually, when BGP on a router restarts, all the BGP peers detect that
   the session went down and then came up.  This "down/up" transition
   results in a "routing flap" and causes BGP route re-computation,
   generation of BGP routing updates, and unnecessary churn to the
   forwarding tables.  It could spread across multiple routing domains.
   Such routing flaps may create transient forwarding blackholes and/or
   transient forwarding loops.  They also consume resources on the
   control plane of the routers affected by the flap.  As such, they are
   detrimental to the overall network performance.

   This document describes a mechanism for BGP that would help minimize
   the negative effects on routing caused by BGP restart.  An End-of-RIB
   marker is specified and can be used to convey routing convergence
   information.  A new BGP capability, termed "Graceful Restart
   Capability", is defined that would allow a BGP speaker to express its
   ability to preserve forwarding state during BGP restart.  Finally,
   procedures are outlined for temporarily retaining routing information
   across a TCP session termination/re-establishment.

1.1  Specification of Requirements



   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in RFC 2119 [RFC2119].







Sangli, et al.              Standards Track                     [Page 2]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


2.  Marker for End-of-RIB



   An UPDATE message with no reachable Network Layer Reachability
   Information (NLRI) and empty withdrawn NLRI is specified as the End-
   of-RIB marker that can be used by a BGP speaker to indicate to its
   peer the completion of the initial routing update after the session
   is established.  For the IPv4 unicast address family, the End-of-RIB
   marker is an UPDATE message with the minimum length [BGP-4].  For any
   other address family, it is an UPDATE message that contains only the
   MP_UNREACH_NLRI attribute [BGP-MP] with no withdrawn routes for that
   <AFI, SAFI>.

   Although the End-of-RIB marker is specified for the purpose of BGP
   graceful restart, it is noted that the generation of such a marker
   upon completion of the initial update would be useful for routing
   convergence in general, and thus the practice is recommended.

   In addition, it would be beneficial for routing convergence if a BGP
   speaker can indicate to its peer up-front that it will generate the
   End-of-RIB marker, regardless of its ability to preserve its
   forwarding state during BGP restart.  This can be accomplished using
   the Graceful Restart Capability described in the next section.

3.  Graceful Restart Capability



   The Graceful Restart Capability is a new BGP capability [BGP-CAP]
   that can be used by a BGP speaker to indicate its ability to preserve
   its forwarding state during BGP restart.  It can also be used to
   convey to its peer its intention of generating the End-of-RIB marker
   upon the completion of its initial routing updates.

   This capability is defined as follows:

      Capability code: 64

      Capability length: variable

      Capability value: Consists of the "Restart Flags" field, "Restart
      Time" field, and 0 to 63 of the tuples <AFI, SAFI, Flags for
      address family> as follows:











Sangli, et al.              Standards Track                     [Page 3]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


         +--------------------------------------------------+
         | Restart Flags (4 bits)                           |
         +--------------------------------------------------+
         | Restart Time in seconds (12 bits)                |
         +--------------------------------------------------+
         | Address Family Identifier (16 bits)              |
         +--------------------------------------------------+
         | Subsequent Address Family Identifier (8 bits)    |
         +--------------------------------------------------+
         | Flags for Address Family (8 bits)                |
         +--------------------------------------------------+
         | ...                                              |
         +--------------------------------------------------+
         | Address Family Identifier (16 bits)              |
         +--------------------------------------------------+
         | Subsequent Address Family Identifier (8 bits)    |
         +--------------------------------------------------+
         | Flags for Address Family (8 bits)                |
         +--------------------------------------------------+

   The use and meaning of the fields are as follows:

      Restart Flags:

         This field contains bit flags related to restart.

             0 1 2 3
            +-+-+-+-+
            |R|Resv.|
            +-+-+-+-+



         The most significant bit is defined as the Restart State (R)
         bit, which can be used to avoid possible deadlock caused by
         waiting for the End-of-RIB marker when multiple BGP speakers
         peering with each other restart.  When set (value 1), this bit
         indicates that the BGP speaker has restarted, and its peer MUST
         NOT
wait for the End-of-RIB marker from the speaker before
         advertising routing information to the speaker.

         The remaining bits are reserved and MUST be set to zero by the
         sender and ignored by the receiver.

      Restart Time:

         This is the estimated time (in seconds) it will take for the
         BGP session to be re-established after a restart.  This can be
         used to speed up routing convergence by its peer in case that
         the BGP speaker does not come back after a restart.



Sangli, et al.              Standards Track                     [Page 4]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


      Address Family Identifier (AFI), Subsequent Address Family
         Identifier (SAFI):

         The AFI and SAFI, taken in combination, indicate that Graceful
         Restart is supported for routes that are advertised with the
         same AFI and SAFI.  Routes may be explicitly associated with a
         particular AFI and SAFI using the encoding of [BGP-MP] or
         implicitly associated with <AFI=IPv4, SAFI=Unicast> if using
         the encoding of [BGP-4].

      Flags for Address Family:

         This field contains bit flags relating to routes that were
         advertised with the given AFI and SAFI.

             0 1 2 3 4 5 6 7
            +-+-+-+-+-+-+-+-+
            |F|   Reserved  |
            +-+-+-+-+-+-+-+-+

         The most significant bit is defined as the Forwarding State (F)
         bit, which can be used to indicate whether the forwarding state
         for routes that were advertised with the given AFI and SAFI has
         indeed been preserved during the previous BGP restart.  When
         set (value 1), the bit indicates that the forwarding state has
         been preserved.

         The remaining bits are reserved and MUST be set to zero by the
         sender and ignored by the receiver.

   When a sender of this capability does not include any <AFI, SAFI> in
   the capability, it means that the sender is not capable of preserving
   its forwarding state during BGP restart, but supports procedures for
   the Receiving Speaker (as defined in Section 4.2 of this document).
   In that case, the value of the "Restart Time" field advertised by the
   sender is irrelevant.

   A BGP speaker MUST NOT include more than one instance of the Graceful
   Restart Capability in the capability advertisement [BGP-CAP].  If
   more than one instance of the Graceful Restart Capability is carried
   in the capability advertisement, the receiver of the advertisement
   MUST ignore all but the last instance of the Graceful Restart
   Capability.

   Including <AFI=IPv4, SAFI=unicast> in the Graceful Restart Capability
   does not imply that the IPv4 unicast routing information should be
   carried by using the BGP multiprotocol extensions [BGP-MP] -- it
   could be carried in the NLRI field of the BGP UPDATE message.



Sangli, et al.              Standards Track                     [Page 5]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


4.  Operation



   A BGP speaker MAY advertise the Graceful Restart Capability for an
   address family to its peer if it has the ability to preserve its
   forwarding state for the address family when BGP restarts.  In
   addition, even if the speaker does not have the ability to preserve
   its forwarding state for any address family during BGP restart, it is
   still recommended that the speaker advertise the Graceful Restart
   Capability to its peer (as mentioned before this is done by not
   including any <AFI, SAFI> in the advertised capability).  There are
   two reasons for doing this.  The first is to indicate its intention
   of generating the End-of-RIB marker upon the completion of its
   initial routing updates, as doing this would be useful for routing
   convergence in general.  The second is to indicate its support for a
   peer which wishes to perform a graceful restart.

   The End-of-RIB marker MUST be sent by a BGP speaker to its peer once
   it completes the initial routing update (including the case when
   there is no update to send) for an address family after the BGP
   session is established.

   It is noted that the normal BGP procedures MUST be followed when the
   TCP session terminates due to the sending or receiving of a BGP
   NOTIFICATION message.

   A suggested default for the Restart Time is a value less than or
   equal to the HOLDTIME carried in the OPEN.

   In the following sections, "Restarting Speaker" refers to a router
   whose BGP has restarted, and "Receiving Speaker" refers to a router
   that peers with the restarting speaker.

   Consider that the Graceful Restart Capability for an address family
   is advertised by the Restarting Speaker, and is understood by the
   Receiving Speaker, and a BGP session between them is established.
   The following sections detail the procedures that MUST be followed by
   the Restarting Speaker as well as the Receiving Speaker once the
   Restarting Speaker restarts.

4.1.  Procedures for the Restarting Speaker



   When the Restarting Speaker restarts, it MUST retain, if possible,
   the forwarding state for the BGP routes in the Loc-RIB and MUST mark
   them as stale.  It MUST NOT differentiate between stale and other
   information during forwarding.

   To re-establish the session with its peer, the Restarting Speaker
   MUST set the "Restart State" bit in the Graceful Restart Capability



Sangli, et al.              Standards Track                     [Page 6]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


   of the OPEN message.  Unless allowed via configuration, the
   "Forwarding State" bit for an address family in the capability can be
   set only if the forwarding state has indeed been preserved for that
   address family during the restart.

   Once the session between the Restarting Speaker and the Receiving
   Speaker is re-established, the Restarting Speaker will receive and
   process BGP messages from its peers.  However, it MUST defer route
   selection for an address family until it either (a) receives the
   End-of-RIB marker from all its peers (excluding the ones with the
   "Restart State" bit set in the received capability and excluding the
   ones that do not advertise the graceful restart capability) or (b)
   the Selection_Deferral_Timer referred to below has expired.  It is
   noted that prior to route selection, the speaker has no routes to
   advertise to its peers and no routes to update the forwarding state.

   In situations where both Interior Gateway Protocol (IGP) and BGP have
   restarted, it might be advantageous to wait for IGP to converge
   before the BGP speaker performs route selection.

   After the BGP speaker performs route selection, the forwarding state
   of the speaker MUST be updated and any previously marked stale
   information MUST be removed.  The Adj-RIB-Out can then be advertised
   to its peers.  Once the initial update is complete for an address
   family (including the case that there is no routing update to send),
   the End-of-RIB marker MUST be sent.

   To put an upper bound on the amount of time a router defers its route
   selection, an implementation MUST support a (configurable) timer that
   imposes this upper bound.  This timer is referred to as the
   "Selection_Deferral_Timer".  The value of this timer should be large
   enough, so as to provide all the peers of the Restarting Speaker with
   enough time to send all the routes to the Restarting Speaker.

   If one wants to apply graceful restart only when the restart is
   planned (as opposed to both planned and unplanned restart), then one
   way to accomplish this would be to set the Forwarding State bit to 1
   after a planned restart, and to 0 in all other cases.  Other
   approaches to accomplish this are outside the scope of this document.

4.2.  Procedures for the Receiving Speaker



   When the Restarting Speaker restarts, the Receiving Speaker may or
   may not detect the termination of the TCP session with the Restarting
   Speaker, depending on the underlying TCP implementation, whether or
   not [BGP-AUTH] is in use, and the specific circumstances of the
   restart.  In case it does not detect the termination of the old TCP
   session and still considers the BGP session as being established, it



Sangli, et al.              Standards Track                     [Page 7]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


   MUST treat the subsequent open connection from the peer as an
   indication of the termination of the old TCP session and act
   accordingly (when the Graceful Restart Capability has been received
   from the peer).  See Section 8 for a description of this behavior in
   terms of the BGP finite state machine.

   "Acting accordingly" in this context means that the previous TCP
   session MUST be closed, and the new one retained.  Note that this
   behavior differs from the default behavior, as specified in [BGP-4],
   Section 6.8.  Since the previous connection is considered to be
   terminated, no NOTIFICATION message should be sent -- the previous
   TCP session is simply closed.

   When the Receiving Speaker detects termination of the TCP session for
   a BGP session with a peer that has advertised the Graceful Restart
   Capability, it MUST retain the routes received from the peer for all
   the address families that were previously received in the Graceful
   Restart Capability and MUST mark them as stale routing information.
   To deal with possible consecutive restarts, a route (from the peer)
   previously marked as stale MUST be deleted.  The router MUST NOT
   differentiate between stale and other routing information during
   forwarding.

   In re-establishing the session, the "Restart State" bit in the
   Graceful Restart Capability of the OPEN message sent by the Receiving
   Speaker MUST NOT be set unless the Receiving Speaker has restarted.
   The presence and the setting of the "Forwarding State" bit for an
   address family depend upon the actual forwarding state and
   configuration.

   If the session does not get re-established within the "Restart Time"
   that the peer advertised previously, the Receiving Speaker MUST
   delete all the stale routes from the peer that it is retaining.

   A BGP speaker could have some way of determining whether its peer's
   forwarding state is still viable, for example through Bidirectional
   Forwarding Detection [BFD] or through monitoring layer two
   information.  Specifics of such mechanisms are beyond the scope of
   this document.  In the event that it determines that its peer's
   forwarding state is not viable prior to the re-establishment of the
   session, the speaker MAY delete all the stale routes from the peer
   that it is retaining.

   Once the session is re-established, if the "Forwarding State" bit for
   a specific address family is not set in the newly received Graceful
   Restart Capability, or if a specific address family is not included
   in the newly received Graceful Restart Capability, or if the Graceful
   Restart Capability is not received in the re-established session at



Sangli, et al.              Standards Track                     [Page 8]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


   all, then the Receiving Speaker MUST immediately remove all the stale
   routes from the peer that it is retaining for that address family.

   The Receiving Speaker MUST send the End-of-RIB marker once it
   completes the initial update for an address family (including the
   case that it has no routes to send) to the peer.

   The Receiving Speaker MUST replace the stale routes by the routing
   updates received from the peer.  Once the End-of-RIB marker for an
   address family is received from the peer, it MUST immediately remove
   any routes from the peer that are still marked as stale for that
   address family.

   To put an upper bound on the amount of time a router retains the
   stale routes, an implementation MAY support a (configurable) timer
   that imposes this upper bound.

5.  Changes to BGP Finite State Machine



   As mentioned under "Procedures for the Receiving Speaker" above, this
   specification modifies the BGP finite state machine.

   The specific state machine modifications to [BGP-4], Section 8.2.2,
   are as follows.

   In the Idle state, make the following changes.

   Replace this text:

      - initializes all BGP resources for the peer connection,

   with

      - initializes all BGP resources for the peer connection, other
        than those resources required in order to retain routes
        according to section "Procedures for the Receiving Speaker" of
        this (Graceful Restart) specification,

   In the Established state, make the following changes.

   Replace this text:

        In response to an indication that the TCP connection is
        successfully established (Event 16 or Event 17), the second
        connection SHALL be tracked until it sends an OPEN message.

   with




Sangli, et al.              Standards Track                     [Page 9]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


        If the Graceful Restart Capability with one or more AFIs/SAFIs
        has not been received for the session, then in response to an
        indication that a TCP connection is successfully established
        (Event 16 or Event 17), the second connection SHALL be tracked
        until it sends an OPEN message.

        However, if the Graceful Restart Capability with one or more
        AFIs/SAFIs has been received for the session, then in response
        to Event 16 or Event 17 the local system:

         - retains all routes associated with this connection according
           to section "Procedures for the Receiving Speaker" of this
           (Graceful Restart) specification,

         - releases all other BGP resources,

         - drops the TCP connection associated with the ESTABLISHED
           session,

         - initializes all BGP resources for the peer connection, other
           than those required in order to retain routes according to
           section "Procedures for the Receiving Speaker" of this
           specification,

         - sets ConnectRetryCounter to zero,

         - starts the ConnectRetryTimer with the initial value, and

         - changes its state to Connect.

   Replace this text:

      If the local system receives a NOTIFICATION message (Event 24 or
      Event 25), or a TcpConnectionFails (Event 18) from the underlying
      TCP, the local system:

         - sets the ConnectRetryTimer to zero,

         - deletes all routes associated with this connection,

         - releases all the BGP resources,

         - drops the TCP connection,

         - increments the ConnectRetryCounter by 1,

         - changes its state to Idle.




Sangli, et al.              Standards Track                    [Page 10]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


   with

      If the local system receives a NOTIFICATION message (Event 24 or
      Event 25), or if the local system receives a TcpConnectionFails
      (Event 18) from the underlying TCP and the Graceful Restart
      capability with one or more AFIs/SAFIs has not been received for
      the session, the local system:

         - sets the ConnectRetryTimer to zero,

         - deletes all routes associated with this connection,

         - releases all the BGP resources,

         - drops the TCP connection,

         - increments the ConnectRetryCounter by 1, and

         - changes its state to Idle.

      However, if the local system receives a TcpConnectionFails (Event
      18) from the underlying TCP, and the Graceful Restart Capability
      with one or more AFIs/SAFIs has been received for the session, the
      local system:

         - sets the ConnectRetryTimer to zero,

         - retains all routes associated with this connection according
           to section "Procedures for the Receiving Speaker" of this
           (Graceful Restart) specification,

         - releases all other BGP resources,

         - drops the TCP connection,

         - increments the ConnectRetryCounter by 1, and

         - changes its state to Idle.

6.  Deployment Considerations



   Although the procedures described in this document would help
   minimize the effect of routing flaps, it is noted that when a BGP
   Graceful Restart-capable router restarts, or if it restarts without
   preserving its forwarding state (e.g., due to a power failure), there
   is a potential for transient routing loops or blackholes in the
   network if routing information changes before the involved routers
   complete routing updates and convergence.  Also, depending on the



Sangli, et al.              Standards Track                    [Page 11]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


   network topology, if not all IBGP speakers are Graceful Restart
   capable, there could be an increased exposure to transient routing
   loops or blackholes when the Graceful Restart procedures are
   exercised.

   The Restart Time, the upper bound for retaining routes, and the upper
   bound for deferring route selection may need to be tuned as more
   deployment experience is gained.

   Finally, it is noted that the benefits of deploying BGP Graceful
   Restart in an Autonomous System (AS) whose IGPs and BGP are tightly
   coupled (i.e., BGP and IGPs would both restart) and IGPs have no
   similar Graceful Restart Capability are reduced relative to the
   scenario where IGPs do have similar Graceful Restart Capability.

7.  Security Considerations



   Since with this proposal a new connection can cause an old one to be
   terminated, it might seem to open the door to denial of service
   attacks.  However, it is noted that unauthenticated BGP is already
   known to be vulnerable to denials of service through attacks on the
   TCP transport.  The TCP transport is commonly protected through use
   of [BGP-AUTH].  Such authentication will equally protect against
   denials of service through spurious new connections.

   If an attacker is able to successfully open a TCP connection
   impersonating a legitimate peer, the attacker's connection will
   replace the legitimate one, potentially enabling the attacker to
   advertise bogus routes.  We note, however, that the window for such a
   route insertion attack is small since through normal operation of the
   protocol the legitimate peer would open a new connection, in turn
   causing the attacker's connection to be terminated.  Thus, this
   attack devolves to a form of denial of service.

   It is thus concluded that this proposal does not change the
   underlying security model (and issues) of BGP-4.

   We also note that implementations may allow use of graceful restart
   to be controlled by configuration.  If graceful restart is not
   enabled, naturally the underlying security model of BGP-4 is
   unchanged.










Sangli, et al.              Standards Track                    [Page 12]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


8.  Acknowledgments



   The authors would like to thank Bruce Cole, Lars Eggert, Bill Fenner,
   Eric Gray, Jeffrey Haas, Sam Hartman, Alvaro Retana, Pekka Savola
   Naiming Shen, Satinder Singh, Mark Townsley, David Ward, Shane
   Wright, and Alex Zinin for their review and comments.

9.  IANA Considerations



   This document defines a new BGP capability - Graceful Restart
   Capability.  The Capability Code for Graceful Restart Capability is
   64.

10.  References



10.1.  Normative References



   [BGP-4]     Rekhter, Y., Li, T., and S. Hares, "A Border Gateway
               Protocol 4 (BGP-4)", RFC 4271, January 2006.

   [BGP-MP]    Bates, T., Rekhter, Y., Chandra, R., and D. Katz,
               "Multiprotocol Extensions for BGP-4", RFC 2858, June
               2000.

   [BGP-CAP]   Chandra, R. and J. Scudder, "Capabilities Advertisement
               with BGP-4", RFC 3392, November 2002.

   [BGP-AUTH]  Heffernan, A., "Protection of BGP Sessions via the TCP
               MD5 Signature Option", RFC 2385, August 1998.

   [RFC2119]   Bradner, S., "Key words for use in RFCs to Indicate
               Requirement Levels", BCP 14, RFC 2119, March 1997.

   [IANA-AFI]  http://www.iana.org/assignments/address-family-numbers

   [IANA-SAFI] http://www.iana.org/assignments/safi-namespace

10.2.  Informative References



   [BFD]       Katz, D. and D. Ward, "Bidirectional Forwarding
               Detection", Work in Progress.










Sangli, et al.              Standards Track                    [Page 13]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


Authors' Addresses



   Srihari R.  Sangli
   Cisco Systems, Inc.

   EMail: rsrihari@cisco.com


   Yakov Rekhter
   Juniper Networks, Inc.

   EMail: yakov@juniper.net


   Rex Fernando
   Juniper Networks, Inc.

   EMail: rex@juniper.net


   John G.  Scudder
   Juniper Networks, Inc.

   EMail: jgs@juniper.net


   Enke Chen
   Cisco Systems, Inc.

   EMail:  enkechen@cisco.com





















Sangli, et al.              Standards Track                    [Page 14]

RFC 4724           Graceful Restart Mechanism for BGP       January 2007


Full Copyright Statement



   Copyright (C) The IETF Trust (2007).

   This document is subject to the rights, licenses and restrictions
   contained in BCP 78, and except as set forth therein, the authors
   retain all their rights.

   This document and the information contained herein are provided on an
   "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
   OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND
   THE INTERNET ENGINEERING TASK FORCE DISCLAIM 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.

Intellectual Property



   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at
   http://www.ietf.org/ipr.

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights that may cover technology that may be required to implement
   this standard.  Please address the information to the IETF at
   ietf-ipr@ietf.org.

Acknowledgement



   Funding for the RFC Editor function is currently provided by the
   Internet Society.







Sangli, et al.              Standards Track                    [Page 15]