Network Working Group                                         Y. Rekhter
Request for Comments: 1654        T.J. Watson Research Center, IBM Corp.
Category: Standards Track                                          T. Li
                                                           cisco Systems
                                                                 Editors
                                                               July 1994


                  A Border Gateway Protocol 4 (BGP-4)

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.

1. Acknowledgements

   This document was originally published as RFC 1267 in October 1991,
   jointly authored by Kirk Lougheed (cisco Systems) and Yakov Rekhter
   (IBM).

   We would like to express our thanks to Guy Almes (Rice University),
   Len Bosack (cisco Systems), and Jeffrey C. Honig (Cornell University)
   for their contributions to the earlier version of this document.

   We like to explicitly thank Bob Braden (ISI) for the review of the
   earlier version of this document as well as his constructive and
   valuable comments.

   We would also like to thank Bob Hinden, Director for Routing of the
   Internet Engineering Steering Group, and the team of reviewers he
   assembled to review the previous version (BGP-2) of this document.
   This team, consisting of Deborah Estrin, Milo Medin, John Moy, Radia
   Perlman, Martha Steenstrup, Mike St. Johns, and Paul Tsuchiya, acted
   with a strong combination of toughness, professionalism, and
   courtesy.

   This updated version of the document is the product of the IETF BGP
   Working Group with Yakov Rekhter and Tony Li as editors. Certain
   sections of the document borrowed heavily from IDRP [7], which is the
   OSI counterpart of BGP. For this credit should be given to the ANSI
   X3S3.3 group chaired by Lyman Chapin (BBN) and to Charles Kunzinger
   (IBM Corp.) who was the IDRP editor within that group.  We would also
   like to thank Mike Craren (Proteon, Inc.), Dimitry Haskin
   (Wellfleet), John Krawczyk (Wellfleet), and Paul Traina (cisco) for



Rekhter & Li                                                    [Page 1]


RFC 1654                         BGP-4                         July 1994


   their insightful comments.

   We would like to specially acknowledge numerous contributions by
   Dennis Ferguson (ANS).

2.  Introduction

   The Border Gateway Protocol (BGP) is an inter-Autonomous System
   routing protocol.  It is built on experience gained with EGP as
   defined in RFC 904 [1] and EGP usage in the NSFNET Backbone as
   described in RFC 1092 [2] and RFC 1093 [3].

   The primary function of a BGP speaking system is to exchange network
   reachability information with other BGP systems.  This network
   reachability information includes information on the list of
   Autonomous Systems (ASs) that reachability information traverses.
   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-4 provides a new set of mechanisms for supporting classless
   interdomain routing.  These mechanisms include support for
   advertising an IP prefix and eliminates the concept of network
   "class" within BGP.  BGP-4 also introduces mechanisms which allow
   aggregation of routes, including aggregation of AS paths.  These
   changes provide support for the proposed supernetting scheme [8, 9].

   To characterize the set of policy decisions that can be enforced
   using BGP, one must focus on the rule that a BGP speaker advertise to
   its peers (other BGP speakers which it communicates with) in
   neighboring 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 neighboring AS intending that the traffic take a
   different route from that taken by traffic originating in the
   neighboring AS.  On the other hand, BGP can support any policy
   conforming to the "hop-by-hop" routing paradigm.  Since the current
   Internet uses only the "hop-by-hop" routing paradigm and since BGP
   can support any policy that conforms to that paradigm, BGP is highly
   applicable as an inter-AS routing protocol for the current Internet.

   A more complete discussion of what policies can and cannot be
   enforced with BGP is outside the scope of this document (but refer to
   the companion document discussing BGP usage [5]).





Rekhter & Li                                                    [Page 2]


RFC 1654                         BGP-4                         July 1994


   BGP runs over a reliable transport protocol.  This eliminates the
   need to implement explicit update fragmentation, retransmission,
   acknowledgement, and sequencing.  Any authentication scheme used by
   the transport protocol may be used in addition to BGP's own
   authentication mechanisms.  The error notification mechanism used in
   BGP assumes that the transport protocol supports a "graceful" close,
   i.e., that all outstanding data will be delivered before the
   connection is closed.

   BGP uses TCP [4] as its transport protocol.  TCP meets BGP's
   transport requirements and is present in virtually all commercial
   routers and hosts.  In the following descriptions the phrase
   "transport protocol connection" can be understood to refer to a TCP
   connection.  BGP uses TCP port 179 for establishing its connections.

   This memo uses the term `Autonomous System' (AS) throughout.  The
   classic definition of an Autonomous System is a set of routers under
   a single technical administration, using an interior gateway protocol
   and common metrics to route packets within the AS, and using an
   exterior gateway protocol to route packets to other ASs.  Since this
   classic definition was developed, it has become common for a single
   AS to use several interior gateway protocols and sometimes several
   sets of metrics within an AS.  The use of the term Autonomous System
   here stresses the fact that, even when multiple IGPs and metrics are
   used, the administration of an AS appears to other ASs to have a
   single coherent interior routing plan and presents a consistent
   picture of what networks are reachable through it.

   The planned use of BGP in the Internet environment, including such
   issues as topology, the interaction between BGP and IGPs, and the
   enforcement of routing policy rules is presented in a companion
   document [5].  This document is the first of a series of documents
   planned to explore various aspects of BGP application.  Please send
   comments to the BGP mailing list (iwg@ans.net).

3.  Summary of Operation

   Two systems form a transport protocol connection between one another.
   They exchange messages to open and confirm the connection parameters.
   The initial data flow is the entire BGP routing table.  Incremental
   updates are sent as the routing tables change.  BGP does not require
   periodic refresh of the entire BGP routing table.  Therefore, a BGP
   speaker must retain the current version of the entire BGP routing
   tables of all of its peers for the duration of the connection.
   KeepAlive messages are sent periodically to ensure the liveness of
   the connection.  Notification messages are sent in response to errors
   or special conditions.  If a connection encounters an error
   condition, a notification message is sent and the connection is



Rekhter & Li                                                    [Page 3]


RFC 1654                         BGP-4                         July 1994


   closed.

   The hosts executing the Border Gateway Protocol need not be routers.
   A non-routing host could exchange routing information with routers
   via EGP or even an interior routing protocol.  That non-routing host
   could then use BGP to exchange routing information with a border
   router in another Autonomous System.  The implications and
   applications of this architecture are for further study.

   If a particular AS has multiple BGP speakers and is providing transit
   service for other ASs, then care must be taken to ensure a consistent
   view of routing within the AS.  A consistent view of the interior
   routes of the AS is provided by the interior routing protocol.  A
   consistent view of the routes exterior to the AS can be provided by
   having all BGP speakers within the AS maintain direct BGP connections
   with each other.  Using a common set of policies, the BGP speakers
   arrive at an agreement as to which border routers will serve as
   exit/entry points for particular networks outside the AS.  This
   information is communicated to the AS's internal routers, possibly
   via the interior routing protocol.  Care must be taken to ensure that
   the interior routers have all been updated with transit information
   before the BGP speakers announce to other ASs that transit service is
   being provided.

   Connections between BGP speakers of different ASs are referred to as
   "external" links.  BGP connections between BGP speakers within the
   same AS are referred to as "internal" links.  Similarly, a peer in a
   different AS is referred to as an external peer, while a peer in the
   same AS may be described as an internal peer.

3.1 Routes: Advertisement and Storage

   For purposes of this protocol a route is defined as a unit of
   information that pairs a destination with the attributes of a path to
   that destination:

      - Routes are advertised between a pair of BGP speakers in UPDATE
      messages: the destination is the systems whose IP addresses are
      reported in the Network Layer Reachability Information (NLRI)
      field, and the the path is the information reported in the path
      attributes fields of the same UPDATE message.

      - Routes are stored in the Routing Information Bases (RIBs):
      namely, the Adj-RIBs-In, the Loc-RIB, and the Adj-RIBs-Out. Routes
      that will be advertised to other BGP speakers must be present in
      the Adj-RIB-Out; routes that will be used by the local BGP speaker
      must be present in the Loc-RIB, and the next hop for each of these
      routes must be present in the local BGP speaker's forwarding



Rekhter & Li                                                    [Page 4]


RFC 1654                         BGP-4                         July 1994


      information base; and routes that are received from other BGP
      speakers are present in the Adj-RIBs-In.

   If a BGP speaker chooses to advertise the route, it may add to or
   modify the path attributes of the route before advertising it to a
   peer.

   BGP provides mechanisms by which a BGP speaker can inform its peer
   that a previously advertised route is no longer available for use.
   There are three methods by which a given BGP speaker can indicate
   that a route has been withdrawn from service:

      a) the IP prefix that expresses destinations for a previously
      advertised route can be advertised in the WITHDRAWN ROUTES field
      in the UPDATE message, thus marking the associated route as being
      no longer available for use

      b) a replacement route with the same Network Layer Reachability
      Information can be advertised, or

      c) the BGP speaker - BGP speaker connection can be closed, which
      implicitly removes from service all routes which the pair of
      speakers had advertised to each other.

3.2 Routing Information Bases

   The Routing Information Base (RIB) within a BGP speaker consists of
   three distinct parts:

      a) Adj-RIBs-In: The Adj-RIBs-In store routing information that has
      been learned from inbound UPDATE messages. Their contents
      represent routes that are available as an input to the Decision
      Process.

      b) Loc-RIB: The Loc-RIB contains the local routing information
      that the BGP speaker has selected by applying its local policies
      to the routing information contained in its Adj-RIBs-In.

      c) Adj-RIBs-Out: The Adj-RIBs-Out store the information that the
      local BGP speaker has selected for advertisement to its peers. The
      routing information stored in the Adj-RIBs-Out will be carried in
      the local BGP speaker's UPDATE messages and advertised to its
      peers.

   In summary, the Adj-RIBs-In contain unprocessed routing information
   that has been advertised to the local BGP speaker by its peers; the
   Loc-RIB contains the routes that have been selected by the local BGP
   speaker's Decision Process; and the Adj-RIBs-Out organize the routes



Rekhter & Li                                                    [Page 5]


RFC 1654                         BGP-4                         July 1994


   for advertisement to specific peers by means of the local speaker's
   UPDATE messages.

   Although the conceptual model distinguishes between Adj-RIBs-In,
   Loc-RIB, and Adj-RIBs-Out, this neither implies nor requires that an
   implementation must maintain three separate copies of the routing
   information. The choice of implementation (for example, 3 copies of
   the information vs 1 copy with pointers) is not constrained by the
   protocol.

4.  Message Formats

   This section describes message formats used by BGP.

   Messages are sent over a reliable transport protocol connection.  A
   message is processed only after it is entirely received.  The maximum
   message size is 4096 octets.  All implementations are required to
   support this maximum message size.  The smallest message that may be
   sent consists of a BGP header without a data portion, or 19 octets.

4.1 Message Header Format

   Each message has a fixed-size header.  There may or may not be a data
   portion following the header, depending on the message type.  The
   layout of these fields is shown below:

       0                   1                   2                   3
       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                                                               |
      +                                                               +
      |                                                               |
      +                                                               +
      |                           Marker                              |
      +                                                               +
      |                                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |          Length               |      Type     |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Marker:

         This 16-octet field contains a value that the receiver of the
         message can predict.  If the Type of the message is OPEN, or if
         the Authentication Code used in the OPEN message of the
         connection is zero, then the Marker must be all ones.
         Otherwise, the value of the marker can be predicted by some a
         computation specified as part of the authentication mechanism



Rekhter & Li                                                    [Page 6]


RFC 1654                         BGP-4                         July 1994


         used.  The Marker can be used to detect loss of synchronization
         between a pair of BGP peers, and to authenticate incoming BGP
         messages.

      Length:

         This 2-octet unsigned integer indicates the total length of the
         message, including the header, in octets.  Thus, e.g., it
         allows one to locate in the transport-level stream the (Marker
         field of the) next message.  The value of the Length field must
         always be at least 19 and no greater than 4096, and may be
         further constrained, depending on the message type.  No
         "padding" of extra data after the message is allowed, so the
         Length field must have the smallest value required given the
         rest of the message.

      Type:

         This 1-octet unsigned integer indicates the type code of the
         message.  The following type codes are defined:

                                    1 - OPEN
                                    2 - UPDATE
                                    3 - NOTIFICATION
                                    4 - KEEPALIVE

4.2 OPEN Message Format

   After a transport protocol connection is established, the first
   message sent by each side is an OPEN message.  If the OPEN message is
   acceptable, a KEEPALIVE message confirming the OPEN is sent back.
   Once the OPEN is confirmed, UPDATE, KEEPALIVE, and NOTIFICATION
   messages may be exchanged.

   In addition to the fixed-size BGP header, the OPEN message contains
   the following fields:















Rekhter & Li                                                    [Page 7]


RFC 1654                         BGP-4                         July 1994


        0                   1                   2                   3
       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       +-+-+-+-+-+-+-+-+
       |    Version    |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |     My Autonomous System      |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |           Hold Time           |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |                         BGP Identifier                        |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |  Auth. Code   |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |                                                               |
       |                       Authentication Data                     |
       |                                                               |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Version:

         This 1-octet unsigned integer indicates the protocol version
         number of the message.  The current BGP version number is 4.

      My Autonomous System:

         This 2-octet unsigned integer indicates the Autonomous System
         number of the sender.

      Hold Time:

         This 2-octet unsigned integer indicates the number of seconds
         that the sender proposes for the value of the Hold Timer.  Upon
         receipt of an OPEN message, a BGP speaker MUST calculate the
         value of the Hold Timer by using the smaller of its configured
         Hold Time and the Hold Time received in the OPEN message.  The
         Hold Time MUST be either zero or at least three seconds.  An
         implementation may reject connections on the basis of the Hold
         Time.  The calculated value indicates the maximum number of
         seconds that may elapse between the receipt of successive
         KEEPALIVE, and/or UPDATE messages by the sender.

      BGP Identifier:

         This 4-octet unsigned integer indicates the BGP Identifier of
         the sender. A given BGP speaker sets the value of its BGP
         Identifier to an IP address assigned to that BGP speaker.  The
         value of the BGP Identifier is determined on startup and is the
         same for every local interface and every BGP peer.



Rekhter & Li                                                    [Page 8]


RFC 1654                         BGP-4                         July 1994


      Authentication Code:

         This 1-octet unsigned integer indicates the authentication
         mechanism being used.  Whenever an authentication mechanism is
         specified for use within BGP, three things must be included in
         the specification:

         - the value of the Authentication Code which indicates use of
           the mechanism,
         - the form and meaning of the Authentication Data, and
         - the algorithm for computing values of Marker fields.  Only
           one authentication mechanism is specified as part of this
           memo:
         - its Authentication Code is zero,
         - its Authentication Data must be empty (of zero length), and
         - the Marker fields of all messages must be all ones.  The
           semantics of non-zero Authentication Codes lies outside the
           scope of this memo.

         Note that a separate authentication mechanism may be used in
         establishing the transport level connection.

      Authentication Data:

         The form and meaning of this field is a variable-length field
         depend on the Authentication Code.  If the value of
         Authentication Code field is zero, the Authentication Data
         field must have zero length.  The semantics of the non-zero
         length Authentication Data field is outside the scope of this
         memo.

         Note that the length of the Authentication Data field can be
         determined from the message Length field by the formula:

            Message Length = 29 + Authentication Data Length

         The minimum length of the OPEN message is 29 octets (including
         message header).

4.3 UPDATE Message Format

   UPDATE messages are used to transfer routing information between BGP
   peers.  The information in the UPDATE packet can be used to construct
   a graph describing the relationships of the various Autonomous
   Systems.  By applying rules to be discussed, routing information
   loops and some other anomalies may be detected and removed from
   inter-AS routing.




Rekhter & Li                                                    [Page 9]


RFC 1654                         BGP-4                         July 1994


   An UPDATE message is used to advertise a single feasible route to a
   peer, or to withdraw multiple unfeasible routes from service (see
   3.1). An UPDATE message may simultaneously advertise a feasible route
   and withdraw multiple unfeasible routes from service.  The UPDATE
   message always includes the fixed-size BGP header, and can optionally
   include the other fields as shown below:

      +-----------------------------------------------------+
      |   Unfeasible Routes Length (2 octets)               |
      +-----------------------------------------------------+
      |  Withdrawn Routes (variable)                        |
      +-----------------------------------------------------+
      |   Total Path Attribute Length (2 octets)            |
      +-----------------------------------------------------+
      |    Path Attributes (variable)                       |
      +-----------------------------------------------------+
      |   Network Layer Reachability Information (variable) |
      +-----------------------------------------------------+

      Unfeasible Routes Length:

         This 2-octets unsigned integer indicates the total length of
         the Withdrawn Routes field in octets.  Its value must allow the
         length of the Network Layer Reachability Information field to
         be determined as specified below.

         A value of 0 indicates that no routes are being withdrawn from
         service, and that the WITHDRAWN ROUTES field is not present in
         this UPDATE message.

      Withdrawn Routes:


         This is a variable length field that contains a list of IP
         address prefixes for the routes that are being withdrawn from
         service.  Each IP address prefix is encoded as a 2-tuple of the
         form <length, prefix>, whose fields are described below:

                  +---------------------------+
                  |   Length (1 octet)        |
                  +---------------------------+
                  |   Prefix (variable)       |
                  +---------------------------+








Rekhter & Li                                                   [Page 10]


RFC 1654                         BGP-4                         July 1994


         The use and the meaning of these fields are as follows:

         a) Length:

            The Length field indicates the length in bits of the IP
            address prefix. A length of zero indicates a prefix that
            matches all IP addresses (with prefix, itself, of zero
            octets).

         b) Prefix:

            The Prefix field contains IP address prefixes followed by
            enough trailing bits to make the end of the field fall on an
            octet boundary. Note that the value of trailing bits is
            irrelevant.

      Total Path Attribute Length:

         This 2-octet unsigned integer indicates the total length of the
         Path Attributes field in octets.  Its value must allow the
         length of the Network Layer Reachability field to be determined
         as specified below.

         A value of 0 indicates that no Network Layer Reachability
         Information field is present in this UPDATE message.

      Path Attributes:

         A variable length sequence of path attributes is present in
         every UPDATE.  Each path attribute is a triple <attribute type,
         attribute length, attribute value> of variable length.

         Attribute Type is a two-octet field that consists of the
         Attribute Flags octet followed by the Attribute Type Code
         octet.

                0                   1
                0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
               +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
               |  Attr. Flags  |Attr. Type Code|
               +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

         The high-order bit (bit 0) of the Attribute Flags octet is the
         Optional bit.  It defines whether the attribute is optional (if
         set to 1) or well-known (if set to 0).

         The second high-order bit (bit 1) of the Attribute Flags octet
         is the Transitive bit.  It defines whether an optional



Rekhter & Li                                                   [Page 11]


RFC 1654                         BGP-4                         July 1994


         attribute is transitive (if set to 1) or non-transitive (if set
         to 0).  For well-known attributes, the Transitive bit must be
         set to 1.  (See Section 5 for a discussion of transitive
         attributes.)

         The third high-order bit (bit 2) of the Attribute Flags octet
         is the Partial bit.  It defines whether the information
         contained in the optional transitive attribute is partial (if
         set to 1) or complete (if set to 0).  For well-known attributes
         and for optional non-transitive attributes the Partial bit must
         be set to 0.

         The fourth high-order bit (bit 3) of the Attribute Flags octet
         is the Extended Length bit.  It defines whether the Attribute
         Length is one octet (if set to 0) or two octets (if set to 1).
         Extended Length may be used only if the length of the attribute
         value is greater than 255 octets.

         The lower-order four bits of the Attribute Flags octet are .
         unused. They must be zero (and must be ignored when received).

         The Attribute Type Code octet contains the Attribute Type Code.
         Currently defined Attribute Type Codes are discussed in Section
         5.

         If the Extended Length bit of the Attribute Flags octet is set
         to 0, the third octet of the Path Attribute contains the length
         of the attribute data in octets.

         If the Extended Length bit of the Attribute Flags octet is set
         to 1, then the third and the fourth octets of the path
         attribute contain the length of the attribute data in octets.

         The remaining octets of the Path Attribute represent the
         attribute value and are interpreted according to the Attribute
         Flags and the Attribute Type Code. The supported Attribute Type
         Codes, their attribute values and uses are the following:

         a)   ORIGIN (Type Code 1):

            ORIGIN is a well-known mandatory attribute that defines the
            origin of the path information.   The data octet can assume
            the following values:








Rekhter & Li                                                   [Page 12]


RFC 1654                         BGP-4                         July 1994


                  Value      Meaning

                  0         IGP - Network Layer Reachability Information
                               is interior to the originating AS

                  1         EGP - Network Layer Reachability Information
                               learned via EGP

                  2         INCOMPLETE - Network Layer Reachability
                               Information learned by some other means

            Its usage is defined in 5.1.1

         b) AS_PATH (Type Code 2):

            AS_PATH is a well-known mandatory attribute that is composed
            of a sequence of AS path segments. Each AS path segment is
            represented by a triple <path segment type, path segment
            length, path segment value>.

            The path segment type is a 1-octet long field with the
            following values defined:

                  Value      Segment Type

                  1         AS_SET: unordered set of ASs a route in the
                               UPDATE message has traversed

                  2         AS_SEQUENCE: ordered set of ASs a route in
                               the UPDATE message has traversed

            The path segment length is a 1-octet long field containing
            the number of ASs in the path segment value field.

            The path segment value field contains one or more AS
            numbers, each encoded as a 2-octets long field.

            Usage of this attribute is defined in 5.1.2.

         c)   NEXT_HOP (Type Code 3):

            This is a well-known mandatory attribute that defines the IP
            address of the border router that should be used as the next
            hop to the destinations listed in the Network Layer
            Reachability field of the UPDATE message.

            Usage of this attribute is defined in 5.1.3.




Rekhter & Li                                                   [Page 13]


RFC 1654                         BGP-4                         July 1994


         d) MULTI_EXIT_DISC (Type Code 4):

            This is an optional non-transitive attribute that is a four
            octet non-negative integer. The value of this attribute may
            be used by a BGP speaker's decision process to discriminate
            among multiple exit points to a neighboring autonomous
            system.

            Its usage is defined in 5.1.4.

         e) LOCAL_PREF (Type Code 5):

            LOCAL_PREF is a well-known discretionary attribute that is a
            four octet non-negative integer. It is used by a BGP speaker
            to inform other BGP speakers in its own autonomous system of
            the originating speaker's degree of preference for an
            advertised route. Usage of this attribute is described in
            5.1.5.

         f) ATOMIC_AGGREGATE (Type Code 6)

            ATOMIC_AGGREGATE is a well-known discretionary attribute of
            length 0. It is used by a BGP speaker to inform other BGP
            speakers that the local system selected a less specific
            route without selecting a more specific route which is
            included in it. Usage of this attribute is described in
            5.1.6.

         g) AGGREGATOR (Type Code 7)

            AGGREGATOR is an optional transitive attribute of length 6.
            The attribute contains the last AS number that formed the
            aggregate route (encoded as 2 octets), followed by the IP
            address of the BGP speaker that formed the aggregate route
            (encoded as 4 octets).  Usage of this attribute is described
            in 5.1.7

      Network Layer Reachability Information:

         This variable length field contains a list of IP address
         prefixes.  The length in octets of the Network Layer
         Reachability Information is not encoded explicitly, but can be
         calculated as:

            UPDATE message Length - 23 - Total Path Attributes Length -
            Unfeasible Routes Length

         where UPDATE message Length is the value encoded in the fixed-



Rekhter & Li                                                   [Page 14]


RFC 1654                         BGP-4                         July 1994


         size BGP header, Total Path Attribute Length and Unfeasible
         Routes Length  are the values encoded in the variable part of
         the UPDATE message, and 23 is a combined length of the fixed-
         size BGP header, the Total Path Attribute Length field and the
         Unfeasible Routes Length field.

         Reachability information is encoded as one or more 2-tuples of
         the form <length, prefix>, whose fields are described below:

                  +---------------------------+
                  |   Length (1 octet)        |
                  +---------------------------+
                  |   Prefix (variable)       |
                  +---------------------------+

         The use and the meaning of these fields are as follows:

         a) Length:

            The Length field indicates the length in bits of the IP
            address prefix. A length of zero indicates a prefix that
            matches all IP addresses (with prefix, itself, of zero
            octets).

         b) Prefix:

            The Prefix field contains IP address prefixes followed by
            enough trailing bits to make the end of the field fall on an
            octet boundary. Note that the value of the trailing bits is
            irrelevant.

   The minimum length of the UPDATE message is 23 octets -- 19 octets
   for the fixed header + 2 octets for the Unfeasible Routes Length + 2
   octets for the Total Path Attribute Length (the value of Unfeasible
   Routes Length is 0  and the value of Total Path Attribute Length is
   0).

   An UPDATE message can advertise at most one route, which may be
   described by several path attributes. All path attributes contained
   in a given UPDATE messages apply to the destinations carried in the
   Network Layer Reachability Information field of the UPDATE message.

   An UPDATE message can list multiple routes to be withdrawn from
   service.  Each such route is identified by its destination (expressed
   as an IP prefix), which unambiguously identifies the route in the
   context of the BGP speaker - BGP speaker connection to which it has
   been previously been advertised.




Rekhter & Li                                                   [Page 15]


RFC 1654                         BGP-4                         July 1994


   An UPDATE message may advertise only routes to be withdrawn from
   service, in which case it will not include path attributes or Network
   Layer Reachability Information. Conversely, it may advertise only a
   feasible route, in which case the WITHDRAWN ROUTES field need not be
   present.

4.4 KEEPALIVE Message Format

   BGP does not use any transport protocol-based keep-alive mechanism to
   determine if peers are reachable.  Instead, KEEPALIVE messages are
   exchanged between peers often enough as not to cause the Hold Timer
   to expire.  A reasonable maximum time between KEEPALIVE messages
   would be one third of the Hold Time interval.  KEEPALIVE messages
   MUST NOT be sent more frequently than one per second.  An
   implementation MAY adjust the rate at which it sends KEEPALIVE
   messages as a function of the Hold Time interval.

   If the negotiated Hold Time interval is zero, then periodic KEEPALIVE
   messages MUST NOT be sent.

   KEEPALIVE message consists of only message header and has a length of
   19 octets.

4.5 NOTIFICATION Message Format

   A NOTIFICATION message is sent when an error condition is detected.
   The BGP connection is closed immediately after sending it.

   In addition to the fixed-size BGP header, the NOTIFICATION message
   contains the following fields:

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       | Error code    | Error subcode |           Data                |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+                               +
       |                                                               |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Error Code:

         This 1-octet unsigned integer indicates the type of
         NOTIFICATION.  The following Error Codes have been defined:








Rekhter & Li                                                   [Page 16]


RFC 1654                         BGP-4                         July 1994


            Error Code       Symbolic Name               Reference

              1         Message Header Error             Section 6.1

              2         OPEN Message Error               Section 6.2

              3         UPDATE Message Error             Section 6.3

              4         Hold Timer Expired               Section 6.5

              5         Finite State Machine Error       Section 6.6

              6         Cease                            Section 6.7


      Error subcode:

         This 1-octet unsigned integer provides more specific
         information about the nature of the reported error.  Each Error
         Code may have one or more Error Subcodes associated with it.
         If no appropriate Error Subcode is defined, then a zero
         (Unspecific) value is used for the Error Subcode field.

         Message Header Error subcodes:

                               1  - Connection Not Synchronized.
                               2  - Bad Message Length.
                               3  - Bad Message Type.

         OPEN Message Error subcodes:

                               1  - Unsupported Version Number.
                               2  - Bad Peer AS.
                               3  - Bad BGP Identifier.
                               4  - Unsupported Authentication Code.
                               5  - Authentication Failure.
                               6  - Unacceptable Hold Time.

         UPDATE Message Error subcodes:

                               1 - Malformed Attribute List.
                               2 - Unrecognized Well-known Attribute.
                               3 - Missing Well-known Attribute.
                               4 - Attribute Flags Error.
                               5 - Attribute Length Error.
                               6 - Invalid ORIGIN Attribute
                               7 - AS Routing Loop.
                               8 - Invalid NEXT_HOP Attribute.



Rekhter & Li                                                   [Page 17]


RFC 1654                         BGP-4                         July 1994


                               9 - Optional Attribute Error.
                              10 - Invalid Network Field.
                              11 - Malformed AS_PATH.

      Data:

         This variable-length field is used to diagnose the reason for
         the NOTIFICATION.  The contents of the Data field depend upon
         the Error Code and Error Subcode.  See Section 6 below for more
         details.

         Note that the length of the Data field can be determined from
         the message Length field by the formula:

                  Message Length = 21 + Data Length

   The minimum length of the NOTIFICATION message is 21 octets
   (including message header).

5.  Path Attributes

   This section discusses the path attributes of the UPDATE message.

   Path attributes fall into four separate categories:

               1. Well-known mandatory.
               2. Well-known discretionary.
               3. Optional transitive.
               4. Optional non-transitive.

   Well-known attributes must be recognized by all BGP implementations.
   Some of these attributes are mandatory and must be included in every
   UPDATE message.  Others are discretionary and may or may not be sent
   in a particular UPDATE message.

   All well-known attributes must be passed along (after proper
   updating, if necessary) to other BGP peers.

   In addition to well-known attributes, each path may contain one or
   more optional attributes.  It is not required or expected that all
   BGP implementations support all optional attributes.  The handling of
   an unrecognized optional attribute is determined by the setting of
   the Transitive bit in the attribute flags octet.  Paths with
   unrecognized transitive optional attributes should be accepted. If a
   path with unrecognized transitive optional attribute is accepted and
   passed along to other BGP peers, then the unrecognized transitive
   optional attribute of that path must be passed along with the path to
   other BGP peers with the Partial bit in the Attribute Flags octet set



Rekhter & Li                                                   [Page 18]


RFC 1654                         BGP-4                         July 1994


   to 1. If a path with recognized transitive optional attribute is
   accepted and passed along to other BGP peers and the Partial bit in
   the Attribute Flags octet is set to 1 by some previous AS, it is not
   set back to 0 by the current AS. Unrecognized non-transitive optional
   attributes must be quietly ignored and not passed along to other BGP
   peers.

   New transitive optional attributes may be attached to the path by the
   originator or by any other AS in the path.  If they are not attached
   by the originator, the Partial bit in the Attribute Flags octet is
   set to 1.  The rules for attaching new non-transitive optional
   attributes will depend on the nature of the specific attribute.  The
   documentation of each new non-transitive optional attribute will be
   expected to include such rules.  (The description of the
   MULTI_EXIT_DISC attribute gives an example.)  All optional attributes
   (both transitive and non-transitive) may be updated (if appropriate)
   by ASs in the path.

   The sender of an UPDATE message should order path attributes within
   the UPDATE message in ascending order of attribute type.  The
   receiver of an UPDATE message must be prepared to handle path
   attributes within the UPDATE message that are out of order.

   The same attribute cannot appear more than once within the Path
   Attributes field of a particular UPDATE message.

5.1 Path Attribute Usage

   The usage of each BGP path attributes is described in the following
   clauses.

5.1.1 ORIGIN

   ORIGIN is a well-known mandatory attribute.  The ORIGIN attribute
   shall be generated by the autonomous system that originates the
   associated routing information. It shall be included in the UPDATE
   messages of all BGP speakers that choose to propagate this
   information to other BGP speakers.

5.1.2  AS_PATH

   AS_PATH is a well-known mandatory attribute. This attribute
   identifies the autonomous systems through which routing information
   carried in this UPDATE message has passed. The components of this
   list can be AS_SETs or AS_SEQUENCEs.

   When a BGP speaker propagates a route which it has learned from
   another BGP speaker's UPDATE message, it shall modify the route's



Rekhter & Li                                                   [Page 19]


RFC 1654                         BGP-4                         July 1994


   AS_PATH attribute based on the location of the BGP speaker to which
   the route will be sent:

      a) When a given BGP speaker advertises the route to another BGP
      speaker located in its own autonomous system, the advertising
      speaker shall not modify the AS_PATH attribute associated with the
      route.

      b) When a given BGP speaker advertises the route to a BGP speaker
      located in a neighboring autonomous system, then the advertising
      speaker shall update the AS_PATH attribute as follows:

         1) if the first path segment of the AS_PATH is of type
         AS_SEQUENCE, the local system shall prepend its own AS number
         as the last element of the sequence (put it in the leftmost
         position).

         2) if the first path segment of the AS_PATH is of type AS_SET,
         the local system shall prepend a new path segment of type
         AS_SEQUENCE to the AS_PATH, including its own AS number in that
         segment.

      When a BGP speaker originates a route then:

         a) the originating speaker shall include its own AS number in
         the AS_PATH attribute of all UPDATE messages sent to BGP
         speakers located in neighboring autonomous systems. (In this
         case, the AS number of the originating speaker's autonomous
         system will be the only entry in the AS_PATH attribute).

         b) the originating speaker shall include an empty AS_PATH
         attribute in all UPDATE messages sent to BGP speakers located
         in its own autonomous system. (An empty AS_PATH attribute is
         one whose length field contains the value zero).

5.1.3 NEXT_HOP

   The NEXT_HOP path attribute defines the IP address of the border
   router that should be used as the next hop to the networks listed in
   the UPDATE message.  If a border router belongs to the same AS as its
   peer, then the peer is an internal border router. Otherwise, it is an
   external border router.  A BGP speaker can advertise any internal
   border router as the next hop provided that the interface associated
   with the IP address of this border router (as specified in the
   NEXT_HOP path attribute) shares a common subnet with both the local
   and remote BGP speakers. A BGP speaker can advertise any external
   border router as the next hop, provided that the IP address of this
   border router was learned from one of the BGP speaker's peers, and



Rekhter & Li                                                   [Page 20]


RFC 1654                         BGP-4                         July 1994


   the interface associated with the IP address of this border router
   (as specified in the NEXT_HOP path attribute) shares a common subnet
   with the local and remote BGP speakers.  A BGP speaker needs to be
   able to support disabling advertisement of external border routers.

   A BGP speaker must never advertise an address of a peer to that peer
   as a NEXT_HOP, for a route that the speaker is originating.  A BGP
   speaker must never install a route with itself as the next hop.

   When a BGP speaker advertises the route to a BGP speaker located in
   its own autonomous system, the advertising speaker shall not modify
   the NEXT_HOP attribute associated with the route.  When a BGP speaker
   receives the route via an internal link, it may forward packets to
   the NEXT_HOP address if the address contained in the attribute is on
   a common subnet with the local and remote BGP speakers.

5.1.4   MULTI_EXIT_DISC

   The MULTI_EXIT_DISC attribute may be used on external (inter-AS)
   links to discriminate among multiple exit or entry points to the same
   neighboring AS.  The value of the MULTI_EXIT_DISC attribute is a four
   octet unsigned number which is called a metric.  All other factors
   being equal, the exit or entry point with lower metric should be
   preferred.  If received over external links, the MULTI_EXIT_DISC
   attribute may be propagated over internal links to other BGP speakers
   within the same AS.  The MULTI_EXIT_DISC attribute is never
   propagated to other BGP speakers in neighboring AS's.

5.1.5   LOCAL_PREF

   LOCAL_PREF is a well-known discretionary attribute that shall be
   included in all UPDATE messages that a given BGP speaker sends to the
   other BGP speakers located in its own autonomous system. A BGP
   speaker shall calculate the degree of preference for each external
   route and include the degree of preference when advertising a route
   to its internal peers. The higher degree of preference should be
   preferred. A BGP speaker shall use the degree of preference learned
   via LOCAL_PREF in its decision process (see section 9.1.1).

   A BGP speaker shall not include this attribute in UPDATE messages
   that it sends to BGP speakers located in a neighboring autonomous
   system. If it is contained in an UPDATE message that is received from
   a BGP speaker which is not located in the same autonomous system as
   the receiving speaker, then this attribute shall be ignored by the
   receiving speaker.






Rekhter & Li                                                   [Page 21]


RFC 1654                         BGP-4                         July 1994


5.1.6   ATOMIC_AGGREGATE

   ATOMIC_AGGREGATE is a well-known discretionary attribute.  If a BGP
   speaker, when presented with a set of overlapping routes from one of
   its peers (see 9.1.4), selects the less specific route without
   selecting the more specific one, then the local system shall attach
   the ATOMIC_AGGREGATE attribute to the route when propagating it to
   other BGP speakers (if that attribute is not already present in the
   received less specific route). A BGP speaker that receives a route
   with the ATOMIC_AGGREGATE attribute shall not remove the attribute
   from the route when propagating it to other speakers. A BGP speaker
   that receives a route with the ATOMIC_AGGREGATE attribute shall not
   make any NLRI of that route more specific (as defined in 9.1.4) when
   advertising this route to other BGP speakers.  A BGP speaker that
   receives a route with the ATOMIC_AGGREGATE attribute needs to be
   cognizant of the fact that the actual path to destinations, as
   specified in the NLRI of the route, while having the loop-free
   property, may traverse ASs that are not listed in the AS_PATH
   attribute.

5.1.7   AGGREGATOR

   AGGREGATOR is an optional transitive attribute which may be included
   in updates which are formed by aggregation (see Section 9.2.4.2).  A
   BGP speaker which performs route aggregation may add the AGGREGATOR
   attribute which shall contain its own AS number and IP address.

6.  BGP Error Handling.

   This section describes actions to be taken when errors are detected
   while processing BGP messages.

   When any of the conditions described here are detected, a
   NOTIFICATION message with the indicated Error Code, Error Subcode,
   and Data fields is sent, and the BGP connection is closed.  If no
   Error Subcode is specified, then a zero must be used.

   The phrase "the BGP connection is closed" means that the transport
   protocol connection has been closed and that all resources for that
   BGP connection have been deallocated.  Routing table entries
   associated with the remote peer are marked as invalid.  The fact that
   the routes have become invalid is passed to other BGP peers before
   the routes are deleted from the system.

   Unless specified explicitly, the Data field of the NOTIFICATION
   message that is sent to indicate an error is empty.





Rekhter & Li                                                   [Page 22]


RFC 1654                         BGP-4                         July 1994


6.1 Message Header error handling.

   All errors detected while processing the Message Header are indicated
   by sending the NOTIFICATION message with Error Code Message Header
   Error.  The Error Subcode elaborates on the specific nature of the
   error.

   The expected value of the Marker field of the message header is all
   ones if the message type is OPEN.  The expected value of the Marker
   field for all other types of BGP messages determined based on the
   Authentication Code in the BGP OPEN message and the actual
   authentication mechanism (if the Authentication Code in the BGP OPEN
   message is non-zero). If the Marker field of the message header is
   not the expected one, then a synchronization error has occurred and
   the Error Subcode is set to Connection Not Synchronized.

   If the Length field of the message header is less than 19 or greater
   than 4096, or if the Length field of an OPEN message is less  than
   the minimum length of the OPEN message, or if the Length field of an
   UPDATE message is less than the minimum length of the UPDATE message,
   or if the Length field of a KEEPALIVE message is not equal to 19, or
   if the Length field of a NOTIFICATION message is less than the
   minimum length of the NOTIFICATION message, then the Error Subcode is
   set to Bad Message Length.  The Data field contains the erroneous
   Length field.

   If the Type field of the message header is not recognized, then the
   Error Subcode is set to Bad Message Type.  The Data field contains
   the erroneous Type field.

6.2 OPEN message error handling.

   All errors detected while processing the OPEN message are indicated
   by sending the NOTIFICATION message with Error Code OPEN Message
   Error.  The Error Subcode elaborates on the specific nature of the
   error.

   If the version number contained in the Version field of the received
   OPEN message is not supported, then the Error Subcode is set to
   Unsupported Version Number.  The Data field is a 2-octet unsigned
   integer, which indicates the largest locally supported version number
   less than the version the remote BGP peer bid (as indicated in the
   received OPEN message).

   If the Autonomous System field of the OPEN message is unacceptable,
   then the Error Subcode is set to Bad Peer AS.  The determination of
   acceptable Autonomous System numbers is outside the scope of this
   protocol.



Rekhter & Li                                                   [Page 23]


RFC 1654                         BGP-4                         July 1994


   If the Hold Time field of the OPEN message is unacceptable, then the
   Error Subcode MUST be set to Unacceptable Hold Time.  An
   implementation MUST reject Hold Time values of one or two seconds.
   An implementation MAY reject any proposed Hold Time.  An
   implementation which accepts a Hold Time MUST use the negotiated
   value for the Hold Time.

   If the BGP Identifier field of the OPEN message is syntactically
   incorrect, then the Error Subcode is set to Bad BGP Identifier.
   Syntactic correctness means that the BGP Identifier field represents
   a valid IP host address.

   If the Authentication Code of the OPEN message is not recognized,
   then the Error Subcode is set to Unsupported Authentication Code.  If
   the Authentication Code is zero, then the Authentication Data must be
   of zero length.  Otherwise, the Error Subcode is set to
   Authentication Failure.

   If the Authentication Code is non-zero, then the corresponding
   authentication procedure is invoked.  If the authentication procedure
   (based on Authentication Code and Authentication Data) fails, then
   the Error Subcode is set to Authentication Failure.

6.3 UPDATE message error handling.

   All errors detected while processing the UPDATE message are indicated
   by sending the NOTIFICATION message with Error Code UPDATE Message
   Error.  The error subcode elaborates on the specific nature of the
   error.

   Error checking of an UPDATE message begins by examining the path
   attributes.  If the Unfeasible Routes Length or Total Attribute
   Length is too large (i.e., if Unfeasible Routes Length + Total
   Attribute Length + 23 exceeds the message Length), then the Error
   Subcode is set to Malformed Attribute List.

   If any recognized attribute has Attribute Flags that conflict with
   the Attribute Type Code, then the Error Subcode is set to Attribute
   Flags Error.  The Data field contains the erroneous attribute (type,
   length and value).

   If any recognized attribute has Attribute Length that conflicts with
   the expected length (based on the attribute type code), then the
   Error Subcode is set to Attribute Length Error.  The Data field
   contains the erroneous attribute (type, length and value).

   If any of the mandatory well-known attributes are not present, then
   the Error Subcode is set to Missing Well-known Attribute.  The Data



Rekhter & Li                                                   [Page 24]


RFC 1654                         BGP-4                         July 1994


   field contains the Attribute Type Code of the missing well-known
   attribute.

   If any of the mandatory well-known attributes are not recognized,
   then the Error Subcode is set to Unrecognized Well-known Attribute.
   The Data field contains the unrecognized attribute (type, length and
   value).

   If the ORIGIN attribute has an undefined value, then the Error
   Subcode is set to Invalid Origin Attribute.  The Data field contains
   the unrecognized attribute (type, length and value).

   If the NEXT_HOP attribute field is syntactically incorrect, then the
   Error Subcode is set to Invalid NEXT_HOP Attribute.  The Data field
   contains the incorrect attribute (type, length and value).  Syntactic
   correctness means that the NEXT_HOP attribute represents a valid IP
   host address.  Semantic correctness applies only to the external BGP
   links. It means that the interface associated with the IP address, as
   specified in the NEXT_HOP attribute, shares a common subnet with the
   receiving BGP speaker and is not the IP address of the receiving BGP
   speaker.  If the NEXT_HOP attribute is semantically incorrect, the
   error should be logged, and the the route should be ignored.  In this
   case, no NOTIFICATION message should be sent.

   The AS_PATH attribute is checked for syntactic correctness.  If the
   path is syntactically incorrect, then the Error Subcode is set to
   Malformed AS_PATH.

   If an optional attribute is recognized, then the value of this
   attribute is checked.  If an error is detected, the attribute is
   discarded, and the Error Subcode is set to Optional Attribute Error.

   The Data field contains the attribute (type, length and value).

   If any attribute appears more than once in the UPDATE message, then
   the Error Subcode is set to Malformed Attribute List.

   The NLRI field in the UPDATE message is checked for syntactic
   validity.  If the field is syntactically incorrect, then the Error
   Subcode is set to Invalid Network Field.

6.4 NOTIFICATION message error handling.

   If a peer sends a NOTIFICATION message, and there is an error in that
   message, there is unfortunately no means of reporting this error via
   a subsequent NOTIFICATION message.  Any such error, such as an
   unrecognized Error Code or Error Subcode, should be noticed, logged
   locally, and brought to the attention of the administration of the



Rekhter & Li                                                   [Page 25]


RFC 1654                         BGP-4                         July 1994


   peer.  The means to do this, however, lies outside the scope of this
   document.

6.5 Hold Timer Expired error handling.

   If a system does not receive successive KEEPALIVE and/or UPDATE
   and/or NOTIFICATION messages within the period specified in the Hold
   Time field of the OPEN message, then the NOTIFICATION message with
   Hold Timer Expired Error Code must be sent and the BGP connection
   closed.

6.6 Finite State Machine error handling.

   Any error detected by the BGP Finite State Machine (e.g., receipt of
   an unexpected event) is indicated by sending the NOTIFICATION message
   with Error Code Finite State Machine Error.

6.7 Cease.

   In absence of any fatal errors (that are indicated in this section),
   a BGP peer may choose at any given time to close its BGP connection
   by sending the NOTIFICATION message with Error Code Cease.  However,
   the Cease NOTIFICATION message must not be used when a fatal error
   indicated by this section does exist.

6.8 Connection collision detection.

   If a pair of BGP speakers try simultaneously to establish a TCP
   connection to each other, then two parallel connections between this
   pair of speakers might well be formed.  We refer to this situation as
   connection collision.  Clearly, one of these connections must be
   closed.

   Based on the value of the BGP Identifier a convention is established
   for detecting which BGP connection is to be preserved when a
   collision does occur. The convention is to compare the BGP
   Identifiers of the peers involved in the collision and to retain only
   the connection initiated by the BGP speaker with the higher-valued
   BGP Identifier.

   Upon receipt of an OPEN message, the local system must examine all of
   its connections that are in the OpenConfirm state.  A BGP speaker may
   also examine connections in an OpenSent state if it knows the BGP
   Identifier of the peer by means outside of the protocol.  If among
   these connections there is a connection to a remote BGP speaker whose
   BGP Identifier equals the one in the OPEN message, then the local
   system performs the following collision resolution procedure:




Rekhter & Li                                                   [Page 26]


RFC 1654                         BGP-4                         July 1994


      1. The BGP Identifier of the local system is compared to the BGP
      Identifier of the remote system (as specified in the OPEN
      message).

      2. If the value of the local BGP Identifier is less than the
      remote one, the local system closes BGP connection that already
      exists (the one that is already in the OpenConfirm state), and
      accepts BGP connection initiated by the remote system.

      3. Otherwise, the local system closes newly created BGP connection
      (the one associated with the newly received OPEN message), and
      continues to use the existing one (the one that is already in the
      OpenConfirm state).

      Comparing BGP Identifiers is done by treating them as (4-octet
      long) unsigned integers.

      A connection collision with an existing BGP connection that is in
      Established states causes unconditional closing of the newly
      created connection. Note that a connection collision cannot be
      detected with connections that are in Idle, or Connect, or Active
      states.

      Closing the BGP connection (that results from the collision
      resolution procedure) is accomplished by sending the NOTIFICATION
      message with the Error Code Cease.

7.  BGP Version Negotiation.

   BGP speakers may negotiate the version of the protocol by making
   multiple attempts to open a BGP connection, starting with the highest
   version number each supports.  If an open attempt fails with an Error
   Code OPEN Message Error, and an Error Subcode Unsupported Version
   Number, then the BGP speaker has available the version number it
   tried, the version number its peer tried, the version number passed
   by its peer in the NOTIFICATION message, and the version numbers that
   it supports.  If the two peers do support one or more common
   versions, then this will allow them to rapidly determine the highest
   common version. In order to support BGP version negotiation, future
   versions of BGP must retain the format of the OPEN and NOTIFICATION
   messages.

8.  BGP Finite State machine.

   This section specifies BGP operation in terms of a Finite State
   Machine (FSM).  Following is a brief summary and overview of BGP
   operations by state as determined by this FSM.  A condensed version
   of the BGP FSM is found in Appendix 1.



Rekhter & Li                                                   [Page 27]


RFC 1654                         BGP-4                         July 1994


      Initially BGP is in the Idle state.

      Idle state:

         In this state BGP refuses all incoming BGP connections.  No
         resources are allocated to the peer.  In response to the Start
         event (initiated by either system or operator) the local system
         initializes all BGP resources, starts the ConnectRetry timer,
         initiates a transport connection to other BGP peer, while
         listening for connection that may be initiated by the remote
         BGP peer, and changes its state to Connect.  The exact value of
         the ConnectRetry timer is a local matter, but should be
         sufficiently large to allow TCP initialization.

         If a BGP speaker detects an error, it shuts down the connection
         and changes its state to Idle. Getting out of the Idle state
         requires generation of the Start event.  If such an event is
         generated automatically, then persistent BGP errors may result
         in persistent flapping of the speaker.  To avoid such a
         condition it is recommended that Start events should not be
         generated immediately for a peer that was previously
         transitioned to Idle due to an error. For a peer that was
         previously transitioned to Idle due to an error, the time
         between consecutive generation of Start events, if such events
         are generated automatically, shall exponentially increase. The
         value of the initial timer shall be 60 seconds. The time shall
         be doubled for each consecutive retry.

         Any other event received in the Idle state is ignored.

      Connect state:

         In this state BGP is waiting for the transport protocol
         connection to be completed.

         If the transport protocol connection succeeds, the local system
         clears the ConnectRetry timer, completes initialization, sends
         an OPEN message to its peer, and changes its state to OpenSent.

         If the transport protocol connect fails (e.g., retransmission
         timeout), the local system restarts the ConnectRetry timer,
         continues to listen for a connection that may be initiated by
         the remote BGP peer, and changes its state to Active state.

         In response to the ConnectRetry timer expired event, the local
         system restarts the ConnectRetry timer, initiates a transport
         connection to other BGP peer, continues to listen for a
         connection that may be initiated by the remote BGP peer, and



Rekhter & Li                                                   [Page 28]


RFC 1654                         BGP-4                         July 1994


         stays in the Connect state.

         Start event is ignored in the Active state.

         In response to any other event (initiated by either system or
         operator), the local system releases all BGP resources
         associated with this connection and changes its state to Idle.

      Active state:

         In this state BGP is trying to acquire a peer by initiating a
         transport protocol connection.

         If the transport protocol connection succeeds, the local system
         clears the ConnectRetry timer, completes initialization, sends
         an OPEN message to its peer, sets its Hold Timer to a large
         value, and changes its state to OpenSent.  A Hold Timer value
         of 4 minutes is suggested.

         In response to the ConnectRetry timer expired event, the local
         system restarts the ConnectRetry timer, initiates a transport
         connection to other BGP peer, continues to listen for a
         connection that may be initiated by the remote BGP peer, and
         changes its state to Connect.

         If the local system detects that a remote peer is trying to
         establish BGP connection to it, and the IP address of the
         remote peer is not an expected one, the local system restarts
         the ConnectRetry timer, rejects the attempted connection,
         continues to listen for a connection that may be initiated by
         the remote BGP peer, and stays in the Active state.

         Start event is ignored in the Active state.

         In response to any other event (initiated by either system or
         operator), the local system releases all BGP resources
         associated with this connection and changes its state to Idle.

      OpenSent state:

         In this state BGP waits for an OPEN message from its peer.
         When an OPEN message is received, all fields are checked for
         correctness.  If the BGP message header checking or OPEN
         message checking detects an error (see Section 6.2), or a
         connection collision (see Section 6.8) the local system sends a
         NOTIFICATION message and changes its state to Idle.





Rekhter & Li                                                   [Page 29]


RFC 1654                         BGP-4                         July 1994


         If there are no errors in the OPEN message, BGP sends a
         KEEPALIVE message and sets a KeepAlive timer.  The Hold Timer,
         which was originally set to a large value (see above), is
         replaced with the negotiated Hold Time value (see section 4.2).
         If the negotiated Hold Time value is zero, then the Hold Time
         timer and KeepAlive timers are not started.  If the value of
         the Autonomous System field is the same as the local Autonomous
         System number, then the connection is an "internal" connection;
         otherwise, it is "external".  (This will effect UPDATE
         processing as described below.) Finally, the state is changed
         to OpenConfirm.

         If a disconnect notification is received from the underlying
         transport protocol, the local system closes the BGP connection,
         restarts the ConnectRetry timer, while continue listening for
         connection that may be initiated by the remote BGP peer, and
         goes into the Active state.

         If the Hold Timer expires, the local system sends NOTIFICATION
         message with error code Hold Timer Expired and changes its
         state to Idle.

         In response to the Stop event (initiated by either system or
         operator) the local system sends NOTIFICATION message with
         Error Code Cease and changes its state to Idle.

         Start event is ignored in the OpenSent state.

         In response to any other event the local system sends
         NOTIFICATION message with Error Code Finite State Machine Error
         and changes its state to Idle.

         Whenever BGP changes its state from OpenSent to Idle, it closes
         the BGP (and transport-level) connection and releases all
         resources associated with that connection.

      OpenConfirm state:

         In this state BGP waits for a KEEPALIVE or NOTIFICATION
         message.

         If the local system receives a KEEPALIVE message, it changes
         its state to Established.

         If the Hold Timer expires before a KEEPALIVE message is
         received, the local system sends NOTIFICATION message with
         error code Hold Timer Expired and changes its state to Idle.




Rekhter & Li                                                   [Page 30]


RFC 1654                         BGP-4                         July 1994


         If the local system receives a NOTIFICATION message, it changes
         its state to Idle.

         If the KeepAlive timer expires, the local system sends a
         KEEPALIVE message and restarts its KeepAlive timer.

         If a disconnect notification is received from the underlying
         transport protocol, the local system changes its state to Idle.

         In response to the Stop event (initiated by either system or
         operator) the local system sends NOTIFICATION message with
         Error Code Cease and changes its state to Idle.

         Start event is ignored in the OpenConfirm state.

         In response to any other event the local system sends
         NOTIFICATION message with Error Code Finite State Machine Error
         and changes its state to Idle.

         Whenever BGP changes its state from OpenConfirm to Idle, it
         closes the BGP (and transport-level) connection and releases
         all resources associated with that connection.

      Established state:

         In the Established state BGP can exchange UPDATE, NOTIFICATION,
         and KEEPALIVE messages with its peer.

         If the local system receives an UPDATE or KEEPALIVE message, it
         restarts its Hold Timer, if the negotiated Hold Time value is
         non-zero.

         If the local system receives a NOTIFICATION message, it changes
         its state to Idle.

         If the local system receives an UPDATE message and the UPDATE
         message error handling procedure (see Section 6.3) detects an
         error, the local system sends a NOTIFICATION message and
         changes its state to Idle.

         If a disconnect notification is received from the underlying
         transport protocol, the local system changes its state to Idle.

         If the Hold Timer expires, the local system sends a
         NOTIFICATION message with Error Code Hold Timer Expired and
         changes its state to Idle.





Rekhter & Li                                                   [Page 31]


RFC 1654                         BGP-4                         July 1994


         If the KeepAlive timer expires, the local system sends a
         KEEPALIVE message and restarts its KeepAlive timer.

         Each time the local system sends a KEEPALIVE or UPDATE message,
         it restarts its KeepAlive timer, unless the negotiated Hold
         Time value is zero.

         In response to the Stop event (initiated by either system or
         operator), the local system sends a NOTIFICATION message with
         Error Code Cease and changes its state to Idle.

         Start event is ignored in the Established state.

         In response to any other event, the local system sends
         NOTIFICATION message with Error Code Finite State Machine Error
         and changes its state to Idle.

         Whenever BGP changes its state from Established to Idle, it
         closes the BGP (and transport-level) connection, releases all
         resources associated with that connection, and deletes all
         routes derived from that connection.

9.  UPDATE Message Handling

   An UPDATE message may be received only in the Established state.
   When an UPDATE message is received, each field is checked for
   validity as specified in Section 6.3.

   If an optional non-transitive attribute is unrecognized, it is
   quietly ignored.  If an optional transitive attribute is
   unrecognized, the Partial bit (the third high-order bit) in the
   attribute flags octet is set to 1, and the attribute is retained for
   propagation to other BGP speakers.

   If an optional attribute is recognized, and has a valid value, then,
   depending on the type of the optional attribute, it is processed
   locally, retained, and updated, if necessary, for possible
   propagation to other BGP speakers.

   If the UPDATE message contains a non-empty WITHDRAWN ROUTES field,
   the previously advertised routes whose destinations (expressed as IP
   prefixes) contained in this field shall be removed from the Adj-RIB-
   In.  This BGP speaker shall run its Decision Process since the
   previously advertised route is not longer available for use.

   If the UPDATE message contains a feasible route, it shall be placed
   in the appropriate Adj-RIB-In, and the following additional actions
   shall be taken:



Rekhter & Li                                                   [Page 32]


RFC 1654                         BGP-4                         July 1994


   i) If its Network Layer Reachability Information (NLRI) is identical
   to the one of a route currently stored in the Adj-RIB-In, then the
   new route shall replace the older route in the Adj-RIB-In, thus
   implicitly withdrawing the older route from service. The BGP speaker
   shall run its Decision Process since the older route is no longer
   available for use.

   ii) If the new route is an overlapping route that is included (see
   9.1.4) in an earlier route contained in the Adj-RIB-In, the BGP
   speaker shall run its Decision Process since the more specific route
   has implicitly made a portion of the less specific route unavailable
   for use.

   iii) If the new route has identical path attributes to an earlier
   route contained in the Adj-RIB-In, and is more specific (see 9.1.4)
   than the earlier route, no further actions are necessary.

   iv) If the new route has NLRI that is not present in any of the
   routes currently stored in the Adj-RIB-In, then the new route shall
   be placed in the Adj-RIB-In. The BGP speaker shall run its Decision
   Process.

   v) If the new route is an overlapping route that is less specific
   (see 9.1.4) than an earlier route contained in the Adj-RIB-In, the
   BGP speaker shall run its Decision Process on the set of destinations
   described only by the less specific route.

9.1 Decision Process

   The Decision Process selects routes for subsequent advertisement by
   applying the policies in the local Policy Information Base (PIB) to
   the routes stored in its Adj-RIB-In. The output of the Decision
   Process is the set of routes that will be advertised to all peers;
   the selected routes will be stored in the local speaker's Adj-RIB-
   Out.

   The selection process is formalized by defining a function that takes
   the attribute of a given route as an argument and returns a non-
   negative integer denoting the degree of preference for the route.
   The function that calculates the degree of preference for a given
   route shall not use as its inputs any of the following: the existence
   of other routes, the non-existence of other routes, or the path
   attributes of other routes. Route selection then consists of
   individual application of the degree of preference function to each
   feasible route, followed by the choice of the one with the highest
   degree of preference.





Rekhter & Li                                                   [Page 33]


RFC 1654                         BGP-4                         July 1994


   The Decision Process operates on routes contained in each Adj-RIB-In,
   and is responsible for:

      - selection of routes to be advertised to BGP speakers located in
        the local speaker's autonomous system

      - selection of routes to be advertised to BGP speakers located in
        neighboring autonomous systems

      - route aggregation and route information reduction

   The Decision Process takes place in three distinct phases, each
   triggered by a different event:

      a) Phase 1 is responsible for calculating the degree of preference
      for each route received from a BGP speaker located in a
      neighboring autonomous system, and for advertising to the other
      BGP speakers in the local autonomous system the routes that have
      the highest degree of preference for each distinct destination.

      b) Phase 2 is invoked on completion of phase 1. It is responsible
      for choosing the best route out of all those available for each
      distinct destination, and for installing each chosen route into
      the appropriate Loc-RIB.

      c) Phase 3 is invoked after the Loc-RIB has been modified. It is
      responsible for disseminating routes in the Loc-RIB to each peer
      located in a neighboring autonomous system, according to the
      policies contained in the PIB. Route aggregation and information
      reduction can optionally be performed within this phase.

9.1.1 Phase 1: Calculation of Degree of Preference

   The Phase 1 decision function shall be invoked whenever the local BGP
   speaker receives an UPDATE message from a peer located in a
   neighboring autonomous system that advertises a new route, a
   replacement route, or a withdrawn route.

   The Phase 1 decision function is a separate process which completes
   when it has no further work to do.

   The Phase 1 decision function shall lock an Adj-RIB-In prior to
   operating on any route contained within it, and shall unlock it after
   operating on all new or unfeasible routes contained within it.

   For each newly received or replacement feasible route, the local BGP
   speaker shall determine a degree of preference. If the route is
   learned from a BGP speaker in the local autonomous system, either the



Rekhter & Li                                                   [Page 34]


RFC 1654                         BGP-4                         July 1994


   value of the LOCAL_PREF attribute shall be taken as the degree of
   preference, or the local system shall compute the degree of
   preference of the route based on preconfigured policy information. If
   the route is learned from a BGP speaker in a neighboring autonomous
   system, then the degree of preference shall be computed based on
   preconfigured policy information.  The exact nature of this policy
   information and the computation involved is a local matter.  The
   local speaker shall then run the internal update process of 9.2.1 to
   select and advertise the most preferable route.

9.1.2 Phase 2: Route Selection

   The Phase 2 decision function shall be invoked on completion of Phase
   1.  The Phase 2 function is a separate process which completes when
   it has no further work to do. The Phase 2 process shall consider all
   routes that are present in the Adj-RIBs-In, including those received
   from BGP speakers located in its own autonomous system and those
   received from BGP speakers located in neighboring autonomous systems.

   The Phase 2 decision function shall be blocked from running while the
   Phase 3 decision function is in process. The Phase 2 function shall
   lock all Adj-RIBs-In prior to commencing its function, and shall
   unlock them on completion.

   If the NEXT_HOP attribute of a BGP route depicts an address to which
   the local BGP speaker doesn't have a route in its Loc-RIB, the BGP
   route SHOULD be excluded from the Phase 2 decision function.

   For each set of destinations for which a feasible route exists in the
   Adj-RIBs-In, the local BGP speaker shall identify the route that has:

      a) the highest degree of preference of any route to the same set
         of destinations, or

      b) is the only route to that destination, or

      c) is selected as a result of the Phase 2 tie breaking rules
         specified in 9.1.2.1.

   The local speaker SHALL then install that route in the Loc-RIB,
   replacing any route to the same destination that is currently being
   held in the Loc-RIB. The local speaker MUST determine the immediate
   next hop to the address depicted by the NEXT_HOP attribute of the
   selected route by performing a lookup in the IGP and selecting one of
   the possible paths in the IGP.  This immediate next hop MUST be used
   when installing the selected route in the Loc-RIB.  If the route to
   the address depicted by the NEXT_HOP attribute changes such that the
   immediate next hop changes, route selection should be recalculated as



Rekhter & Li                                                   [Page 35]


RFC 1654                         BGP-4                         July 1994


   specified above.

   Unfeasible routes shall be removed from the Loc-RIB, and
   corresponding unfeasible routes shall then be removed from the Adj-
   RIBs-In.

9.1.2.1 Breaking Ties (Phase 2)

   In its Adj-RIBs-In a BGP speaker may have several routes to the same
   destination that have the same degree of preference. The local
   speaker can select only one of these routes for inclusion in the
   associated Loc-RIB. The local speaker considers all equally
   preferable routes, both those received from BGP speakers located in
   neighboring autonomous systems, and those received from other BGP
   speakers located in the local speaker's autonomous system.

   The following tie-breaking procedure assumes that for each candidate
   route all the BGP speakers within an autonomous system can ascertain
   the cost of a path (interior distance) to the address depicted by the
   NEXT_HOP attribute of the route.  Ties shall be broken according to
   the following algorithm:

      a) If the local system is configured to take into account
         MULTI_EXIT_DISC, and the candidate routes differ in their
         MULTI_EXIT_DISC attribute, select the route that has the
         lowest value of the MULTI_EXIT_DISC attribute.

      b) Otherwise, select the route that has the lowest cost
         (interior distance) to the entity depicted by the NEXT_HOP
         attribute of the route.  If there are several routes with the
         same cost, then the tie-breaking shall be broken as follows:

         - if at least one of the candidate routes was advertised by
           the BGP speaker in a neighboring autonomous system, select
           the route that was advertised by the BGP speaker in a
           neighboring autonomous system whose BGP Identifier has the
           lowest value among all other BGP speakers in neighboring
           autonomous systems;

         - otherwise, select the route that was advertised by the BGP
           speaker whose BGP Identifier has the lowest value.










Rekhter & Li                                                   [Page 36]


RFC 1654                         BGP-4                         July 1994


9.1.3 Phase 3: Route Dissemination

   The Phase 3 decision function shall be invoked on completion of Phase
   2, or when any of the following events occur:

      a) when routes in a Loc-RIB to local destinations have changed

      b) when locally generated routes learned by means outside of BGP
         have changed

      c) when a new BGP speaker - BGP speaker connection has been
         established

   The Phase 3 function is a separate process which completes when it
   has no further work to do. The Phase 3 Routing Decision function
   shall be blocked from running while the Phase 2 decision function is
   in process.

   All routes in the Loc-RIB shall be processed into a corresponding
   entry in the associated Adj-RIBs-Out. Route aggregation and
   information reduction techniques (see 9.2.4.1) may optionally be
   applied.

   For the benefit of future support of inter-AS multicast capabilities,
   a BGP speaker that participates in inter-AS multicast routing shall
   advertise a route it receives from one of its external peers and if
   it installs it in its Loc-RIB, it shall advertise it back to the peer
   from which the route was received. For a BGP speaker that does not
   participate in inter-AS multicast routing such an advertisement is
   optional. When doing such an advertisement, the NEXT_HOP attribute
   should be set to the address of the peer. An implementation may also
   optimize such an advertisement by truncating information in the
   AS_PATH attribute to include only its own AS number and that of the
   peer that advertised the route (such truncation requires the ORIGIN
   attribute to be set to INCOMPLETE).  In addition an implementation is
   not required to pass optional or discretionary path attributes with
   such an advertisement.

   When the updating of the Adj-RIBs-Out and the Forwarding Information
   Base (FIB) is complete, the local BGP speaker shall run the external
   update process of 9.2.2.

9.1.4 Overlapping Routes

   A BGP speaker may transmit routes with overlapping Network Layer
   Reachability Information (NLRI) to another BGP speaker. NLRI overlap
   occurs when a set of destinations are identified in non-matching
   multiple routes. Since BGP encodes NLRI using IP prefixes, overlap



Rekhter & Li                                                   [Page 37]


RFC 1654                         BGP-4                         July 1994


   will always exhibit subset relationships.  A route describing a
   smaller set of destinations (a longer prefix) is said to be more
   specific than a route describing a larger set of destinations (a
   shorted prefix); similarly, a route describing a larger set of
   destinations (a shorter prefix) is said to be less specific than a
   route describing a smaller set of destinations (a longer prefix).

   The precedence relationship effectively decomposes less specific
   routes into two parts:

      -  a set of destinations described only by the less specific
         route, and

      -  a set of destinations described by the overlap of the less
         specific and the more specific routes

   When overlapping routes are present in the same Adj-RIB-In, the more
   specific route shall take precedence, in order from more specific to
   least specific.

   The set of destinations described by the overlap represents a portion
   of the less specific route that is feasible, but is not currently in
   use.  If a more specific route is later withdrawn, the set of
   destinations described by the overlap will still be reachable using
   the less specific route.

   If a BGP speaker receives overlapping routes, the Decision Process
   shall take into account the semantics of the overlapping routes. In
   particular, if a BGP speaker accepts the less specific route while
   rejecting the more specific route from the same peer, then the
   destinations represented by the overlap may not forward along the ASs
   listed in the AS_PATH attribute of that route. Therefore, a BGP
   speaker has the following choices:

      a)   Install both the less and the more specific routes

      b)   Install the more specific route only

      c)   Install the non-overlapping part of the less specific
           route only (that implies de-aggregation)

      d)   Aggregate the two routes and install the aggregated route

      e)   Install the less specific route only

      f)   Install neither route





Rekhter & Li                                                   [Page 38]


RFC 1654                         BGP-4                         July 1994


   If a BGP speaker chooses e), then it should add ATOMIC_AGGREGATE
   attribute to the route. A route that carries ATOMIC_AGGREGATE
   attribute can not be de-aggregated. That is, the NLRI of this route
   can not be made more specific.  Forwarding along such a route does
   not guarantee that IP packets will actually traverse only ASs listed
   in the AS_PATH attribute of the route.  If a BGP speaker chooses a),
   it must not advertise the more general route without the more
   specific route.

9.2 Update-Send Process

   The Update-Send process is responsible for advertising UPDATE
   messages to all peers. For example, it distributes the routes chosen
   by the Decision Process to other BGP speakers which may be located in
   either the same autonomous system or a neighboring autonomous system.
   Rules for information exchange between BGP speakers located in
   different autonomous systems are given in 9.2.2; rules for
   information exchange between BGP speakers located in the same
   autonomous system are given in 9.2.1.

   Distribution of routing information between a set of BGP speakers,
   all of which are located in the same autonomous system, is referred
   to as internal distribution.

9.2.1 Internal Updates

   The Internal update process is concerned with the distribution of
   routing information to BGP speakers located in the local speaker's
   autonomous system.

   When a BGP speaker receives an UPDATE message from another BGP
   speaker located in its own autonomous system, the receiving BGP
   speaker shall not re-distribute the routing information contained in
   that UPDATE message to other BGP speakers located in its own
   autonomous system.

   When a BGP speaker receives a new route from a BGP speaker in a
   neighboring autonomous system, it shall advertise that route to all
   other BGP speakers in its autonomous system by means of an UPDATE
   message if any of the following conditions occur:

      1) the degree of preference assigned to the newly received route
      by the local BGP speaker is higher than the degree of preference
      that the local speaker has assigned to other routes that have been
      received from BGP speakers in neighboring autonomous systems, or

      2) there are no other routes that have been received from BGP
      speakers in neighboring autonomous systems, or



Rekhter & Li                                                   [Page 39]


RFC 1654                         BGP-4                         July 1994


      3) the newly received route is selected as a result of breaking a
      tie between several routes which have the highest degree of
      preference, and the same destination (the tie-breaking procedure
      is specified in 9.2.1.1).

   When a BGP speaker receives an UPDATE message with a non-empty
   WITHDRAWN ROUTES field, it shall remove from its Adj-RIB-In all
   routes whose destinations was carried in this field (as IP prefixes).
   The speaker shall take the following additional steps:

      1) if the corresponding feasible route had not been previously
      advertised, then no further action is necessary

      2) if the corresponding feasible route had been previously
      advertised, then:

         i) if a new route is selected for advertisement that has the
         same Network Layer Reachability Information as the unfeasible
         routes, then the local BGP speaker shall advertise the
         replacement route

         ii) if a replacement route is not available for advertisement,
         then the BGP speaker shall include the destinations  of the
         unfeasible route (in form of IP prefixes) in the WITHDRAWN
         ROUTES field of an UPDATE message, and shall send this message
         to each peer to whom it had previously advertised the
         corresponding feasible route.

   All feasible routes which are advertised shall be placed in the
   appropriate Adj-RIBs-Out, and all unfeasible routes which are
   advertised shall be removed from the Adj-RIBs-Out.

9.2.1.1 Breaking Ties (Internal Updates)

   If a local BGP speaker has connections to several BGP speakers in
   neighboring autonomous systems, there will be multiple Adj-RIBs-In
   associated with these peers. These Adj-RIBs-In might contain several
   equally preferable routes to the same destination, all of which were
   advertised by BGP speakers located in neighboring autonomous systems.
   The local BGP speaker shall select one of these routes according to
   the following rules:

      a) If the candidate route differ only in their NEXT_HOP and
      MULTI_EXIT_DISC attributes, and the local system is configured to
      take into account MULTI_EXIT_DISC attribute, select the routes
      that has the lowest value of the MULTI_EXIT_DISC attribute.





Rekhter & Li                                                   [Page 40]


RFC 1654                         BGP-4                         July 1994


      b) If the local system can ascertain the cost of a path to the
      entity depicted by the NEXT_HOP attribute of the candidate route,
      select the route with the lowest cost.

      c) In all other cases, select the route that was advertised by the
      BGP speaker whose BGP Identifier has the lowest value.

9.2.2 External Updates

   The external update process is concerned with the distribution of
   routing information to BGP speakers located in neighboring autonomous
   systems. As part of Phase 3 route selection process, the BGP speaker
   has updated its Adj-RIBs-Out and its Forwarding Table. All newly
   installed routes and all newly unfeasible routes for which there is
   no replacement route shall be advertised to BGP speakers located in
   neighboring autonomous systems by means of UPDATE message.

   Any routes in the Loc-RIB marked as unfeasible shall be removed.
   Changes to the reachable destinations within its own autonomous
   system shall also be advertised in an UPDATE message.

9.2.3 Controlling Routing Traffic Overhead

   The BGP protocol constrains the amount of routing traffic (that is,
   UPDATE messages) in order to limit both the link bandwidth needed to
   advertise UPDATE messages and the processing power needed by the
   Decision Process to digest the information contained in the UPDATE
   messages.

9.2.3.1 Frequency of Route Advertisement

   The parameter MinRouteAdvertisementInterval determines the minimum
   amount of time that must elapse between advertisement of routes to a
   particular destination from a single BGP speaker. This rate limiting
   procedure applies on a per-destination basis, although the value of
   MinRouteAdvertisementInterval is set on a per BGP peer basis.

   Two UPDATE messages sent from a single BGP speaker that advertise
   feasible routes to some common set of destinations received from BGP
   speakers in neighboring autonomous systems must be separated by at
   least MinRouteAdvertisementInterval. Clearly, this can only be
   achieved precisely by keeping a separate timer for each common set of
   destinations. This would be unwarranted overhead. Any technique which
   ensures that the interval between two UPDATE messages sent from a
   single BGP speaker that advertise feasible routes to some common set
   of destinations received from BGP speakers in neighboring autonomous
   systems will be at least MinRouteAdvertisementInterval, and will also
   ensure a constant upper bound on the interval is acceptable.



Rekhter & Li                                                   [Page 41]


RFC 1654                         BGP-4                         July 1994


   Since fast convergence is needed within an autonomous system, this
   procedure does not apply for routes receives from other BGP speakers
   in the same autonomous system. To avoid long-lived black holes, the
   procedure does not apply to the explicit withdrawal of unfeasible
   routes (that is, routes whose destinations (expressed as IP prefixes)
   are listed in the WITHDRAWN ROUTES field of an UPDATE message).

   This procedure does not limit the rate of route selection, but only
   the rate of route advertisement. If new routes are selected multiple
   times while awaiting the expiration of MinRouteAdvertisementInterval,
   the last route selected shall be advertised at the end of
   MinRouteAdvertisementInterval.

9.2.3.2 Frequency of Route Origination

   The parameter MinASOriginationInterval determines the minimum amount
   of time that must elapse between successive advertisements of UPDATE
   messages that report changes within the advertising BGP speaker's own
   autonomous systems.

9.2.3.3 Jitter

   To minimize the likelihood that the distribution of BGP messages by a
   given BGP speaker will contain peaks, jitter should be applied to the
   timers associated with MinASOriginationInterval, Keepalive, and
   MinRouteAdvertisementInterval. A given BGP speaker shall apply the
   same jitter to each of these quantities regardless of the
   destinations to which the updates are being sent; that is, jitter
   will not be applied on a "per peer" basis.

   The amount of jitter to be introduced shall be determined by
   multiplying the base value of the appropriate timer by a random
   factor which is uniformly distributed in the range from 0.75 to 1.0.

9.2.4 Efficient Organization of Routing Information

   Having selected the routing information which it will advertise, a
   BGP speaker may avail itself of several methods to organize this
   information in an efficient manner.

9.2.4.1 Information Reduction

   Information reduction may imply a reduction in granularity of policy
   control - after information is collapsed, the same policies will
   apply to all destinations and paths in the equivalence class.

   The Decision Process may optionally reduce the amount of information
   that it will place in the Adj-RIBs-Out by any of the following



Rekhter & Li                                                   [Page 42]


RFC 1654                         BGP-4                         July 1994


   methods:

      a)   Network Layer Reachability Information (NLRI):

      Destination IP addresses can be represented as IP address
      prefixes.  In cases where there is a correspondence between the
      address structure and the systems under control of an autonomous
      system administrator, it will be possible to reduce the size of
      the NLRI carried in the UPDATE messages.

      b)   AS_PATHs:

      AS path information can be represented as ordered AS_SEQUENCEs or
      unordered AS_SETs. AS_SETs are used in the route aggregation
      algorithm described in 9.2.4.2. They reduce the size of the
      AS_PATH information by listing each AS number only once,
      regardless of how many times it may have appeared in multiple
      AS_PATHs that were aggregated.

      An AS_SET implies that the destinations listed in the NLRI can be
      reached through paths that traverse at least some of the
      constituent autonomous systems. AS_SETs provide sufficient
      information to avoid routing information looping; however their
      use may prune potentially feasible paths, since such paths are no
      longer listed individually as in the form of AS_SEQUENCEs.  In
      practice this is not likely to be a problem, since once an IP
      packet arrives at the edge of a group of autonomous systems, the
      BGP speaker at that point is likely to have more detailed path
      information and can distinguish individual paths to destinations.

9.2.4.2 Aggregating Routing Information

   Aggregation is the process of combining the characteristics of
   several different routes in such a way that a single route can be
   advertised.  Aggregation can occur as part of the decision  process
   to reduce the amount of routing information that will be placed in
   the Adj-RIBs-Out.

   Aggregation reduces the amount of information that a BGP speaker must
   store and exchange with other BGP speakers. Routes can be aggregated
   by applying the following procedure separately to path attributes of
   like type and to the Network Layer Reachability Information.

   Routes that have the following attributes shall not be aggregated
   unless the corresponding attributes of each route are identical:
   MULTI_EXIT_DISC, NEXT_HOP.





Rekhter & Li                                                   [Page 43]


RFC 1654                         BGP-4                         July 1994


   Path attributes that have different type codes can not be aggregated
   together. Path of the same type code may be aggregated, according to
   the following rules:

      ORIGIN attribute: If at least one route among routes that are
      aggregated has ORIGIN with the value INCOMPLETE, then the
      aggregated route must have the ORIGIN attribute with the value
      INCOMPLETE. Otherwise, if at least one route among routes that are
      aggregated has ORIGIN with the value EGP, then the aggregated
      route must have the origin attribute with the value EGP. In all
      other case the value of the ORIGIN attribute of the aggregated
      route is INTERNAL.

      AS_PATH attribute: If routes to be aggregated have identical
      AS_PATH attributes, then the aggregated route has the same AS_PATH
      attribute as each individual route.

      For the purpose of aggregating AS_PATH attributes we model each AS
      within the AS_PATH attribute as a tuple <type, value>, where
      "type" identifies a type of the path segment the AS belongs to
      (e.g., AS_SEQUENCE, AS_SET), and "value" is the AS number.  If the
      routes to be aggregated have different AS_PATH attributes, then
      the aggregated AS_PATH attribute shall satisfy all of the
      following conditions:

         - all tuples of the type AS_SEQUENCE in the aggregated AS_PATH
         shall appear in all of the AS_PATH in the initial set of routes
         to be aggregated.

         - all tuples of the type AS_SET in the aggregated AS_PATH shall
         appear in at least one of the AS_PATH in the initial set (they
         may appear as either AS_SET or AS_SEQUENCE types).

         - for any tuple X of the type AS_SEQUENCE in the aggregated
         AS_PATH which precedes tuple Y in the aggregated AS_PATH, X
         precedes Y in each AS_PATH in the initial set which contains Y,
         regardless of the type of Y.

         - No tuple with the same value shall appear more than once in
         the aggregated AS_PATH, regardless of the tuple's type.

      An implementation may choose any algorithm which conforms to these
      rules.  At a minimum a conformant implementation shall be able to
      perform the following algorithm that meets all of the above
      conditions:

         - determine the longest leading sequence of tuples (as defined
         above) common to all the AS_PATH attributes of the routes to be



Rekhter & Li                                                   [Page 44]


RFC 1654                         BGP-4                         July 1994


         aggregated. Make this sequence the leading sequence of the
         aggregated AS_PATH attribute.

         - set the type of the rest of the tuples from the AS_PATH
         attributes of the routes to be aggregated to AS_SET, and append
         them to the aggregated AS_PATH attribute.

         - if the aggregated AS_PATH has more than one tuple with the
         same value (regardless of tuple's type), eliminate all, but one
         such tuple by deleting tuples of the type AS_SET from the
         aggregated AS_PATH attribute.

      Appendix 6, section 6.8 presents another algorithm that satisfies
      the conditions and  allows for more complex policy configurations.

      ATOMIC_AGGREGATE: If at least one of the routes to be aggregated
      has ATOMIC_AGGREGATE path attribute, then the aggregated route
      shall have this attribute as well.

      AGGREGATOR: All AGGREGATOR attributes of all routes to be
      aggregated should be ignored.

9.3   Route Selection Criteria

   Generally speaking, additional rules for comparing routes among
   several alternatives are outside the scope of this document.  There
   are two exceptions:

      - If the local AS appears in the AS path of the new route being
      considered, then that new route cannot be viewed as better than
      any other route.  If such a route were ever used, a routing loop
      would result.

      - In order to achieve successful distributed operation, only
      routes with a likelihood of stability can be chosen.  Thus, an AS
      must avoid using unstable routes, and it must not make rapid
      spontaneous changes to its choice of route.  Quantifying the terms
      "unstable" and "rapid" in the previous sentence will require
      experience, but the principle is clear.

9.4   Originating BGP routes

   A BGP speaker may originate BGP routes by injecting routing
   information acquired by some other means (e.g., via an IGP) into BGP.
   A BGP speaker that originates BGP routes shall assign the degree of
   preference to these routes by passing them through the Decision
   Process (see Section 9.1).  These routes may also be distributed to
   other BGP speakers within the local AS as part of the Internal update



Rekhter & Li                                                   [Page 45]


RFC 1654                         BGP-4                         July 1994


   process (see Section 9.2.1). The decision whether to distribute non-
   BGP acquired routes within an AS via BGP or not depends on the
   environment within the AS (e.g., type of IGP) and should be
   controlled via configuration.















































Rekhter & Li                                                   [Page 46]


RFC 1654                         BGP-4                         July 1994


Appendix 1.  BGP FSM State Transitions and Actions.

   This Appendix discusses the transitions between states in the BGP FSM
   in response to BGP events.  The following is the list of these states
   and events when the negotiated Hold Time value is non-zero.

       BGP States:

                1 - Idle
                2 - Connect
                3 - Active
                4 - OpenSent
                5 - OpenConfirm
                6 - Established


       BGP Events:

                1 - BGP Start
                2 - BGP Stop
                3 - BGP Transport connection open
                4 - BGP Transport connection closed
                5 - BGP Transport connection open failed
                6 - BGP Transport fatal error
                7 - ConnectRetry timer expired
                8 - Hold Timer expired
                9 - KeepAlive timer expired
               10 - Receive OPEN message
               11 - Receive KEEPALIVE message
               12 - Receive UPDATE messages
               13 - Receive NOTIFICATION message




















Rekhter & Li                                                   [Page 47]


RFC 1654                         BGP-4                         July 1994


   The following table describes the state transitions of the BGP FSM
   and the actions triggered by these transitions.

    Event                Actions               Message Sent   Next State
    --------------------------------------------------------------------
    Idle (1)
     1            Initialize resources            none             2
                  Start ConnectRetry timer
                  Initiate a transport connection
     others               none                    none             1

    Connect(2)
     1                    none                    none             2
     3            Complete initialization         OPEN             4
                  Clear ConnectRetry timer
     5            Restart ConnectRetry timer      none             3
     7            Restart ConnectRetry timer      none             2
                  Initiate a transport connection
     others       Release resources               none             1

    Active (3)
     1                    none                    none             3
     3            Complete initialization         OPEN             4
                  Clear ConnectRetry timer
     5            Close connection                                 3
                  Restart ConnectRetry timer
     7            Restart ConnectRetry timer      none             2
                  Initiate a transport connection
     others       Release resources               none             1

    OpenSent(4)
     1                    none                    none             4
     4            Close transport connection      none             3
                  Restart ConnectRetry timer
     6            Release resources               none             1
    10            Process OPEN is OK            KEEPALIVE          5
                  Process OPEN failed           NOTIFICATION       1
    others        Close transport connection    NOTIFICATION       1
                  Release resources












Rekhter & Li                                                   [Page 48]


RFC 1654                         BGP-4                         July 1994


    OpenConfirm (5)
     1                   none                     none             5
     4            Release resources               none             1
     6            Release resources               none             1
     9            Restart KeepAlive timer       KEEPALIVE          5
    11            Complete initialization         none             6
                  Restart Hold Timer
    13            Close transport connection                       1
                  Release resources
    others        Close transport connection    NOTIFICATION       1
                  Release resources




    Established (6)
     1                   none                     none             6
     4            Release resources               none             1
     6            Release resources               none             1
     9            Restart KeepAlive timer       KEEPALIVE          6
    11            Restart Hold Timer            KEEPALIVE          6
    12            Process UPDATE is OK          UPDATE             6
                  Process UPDATE failed         NOTIFICATION       1
    13            Close transport connection                       1
                  Release resources
    others        Close transport connection    NOTIFICATION       1
                  Release resources
   ---------------------------------------------------------------------























Rekhter & Li                                                   [Page 49]


RFC 1654                         BGP-4                         July 1994


      The following is a condensed version of the above state transition
      table.


   Events| Idle | Connect | Active | OpenSent | OpenConfirm | Estab
         | (1)  |   (2)   |  (3)   |    (4)   |     (5)     |   (6)
         |---------------------------------------------------------------
    1    |  2   |    2    |   3    |     4    |      5      |    6
         |      |         |        |          |             |
    2    |  1   |    1    |   1    |     1    |      1      |    1
         |      |         |        |          |             |
    3    |  1   |    4    |   4    |     1    |      1      |    1
         |      |         |        |          |             |
    4    |  1   |    1    |   1    |     3    |      1      |    1
         |      |         |        |          |             |
    5    |  1   |    3    |   3    |     1    |      1      |    1
         |      |         |        |          |             |
    6    |  1   |    1    |   1    |     1    |      1      |    1
         |      |         |        |          |             |
    7    |  1   |    2    |   2    |     1    |      1      |    1
         |      |         |        |          |             |
    8    |  1   |    1    |   1    |     1    |      1      |    1
         |      |         |        |          |             |
    9    |  1   |    1    |   1    |     1    |      5      |    6
         |      |         |        |          |             |
   10    |  1   |    1    |   1    |  1 or 5  |      1      |    1
         |      |         |        |          |             |
   11    |  1   |    1    |   1    |     1    |      6      |    6
         |      |         |        |          |             |
   12    |  1   |    1    |   1    |     1    |      1      | 1 or 6
         |      |         |        |          |             |
   13    |  1   |    1    |   1    |     1    |      1      |    1
         |      |         |        |          |             |
         ---------------------------------------------------------------

















Rekhter & Li                                                   [Page 50]


RFC 1654                         BGP-4                         July 1994


Appendix 2. Comparison with RFC 1267

   BGP-4 is capable of operating in an environment where a set of
   reachable destinations may be expressed via a single IP prefix.  The
   concept of network classes, or subnetting is foreign to BGP-4.  To
   accommodate these capabilities BGP-4 changes semantics and encoding
   associated with the AS_PATH attribute. New text has been added to
   define semantics associated with IP prefixes.  These abilities allow
   BGP-4 to support the proposed supernetting scheme [9].

   To simplify configuration this version introduces a new attribute,
   LOCAL_PREF, that facilitates route selection procedures.

   The INTER_AS_METRIC attribute has been renamed to be MULTI_EXIT_DISC.
   A new attribute, ATOMIC_AGGREGATE, has been introduced to insure that
   certain aggregates are not de-aggregated.  Another new attribute,
   AGGREGATOR, can be added to aggregate routes in order to advertise
   which AS and which BGP speaker within that AS caused the aggregation.

   To insure that Hold Timers are symmetric, the Hold Time is now
   negotiated on a per-connection basis.  Hold Times of zero are now
   supported.

Appendix 3.  Comparison with RFC 1163

   All of the changes listed in Appendix 2, plus the following.

   To detect and recover from BGP connection collision, a new field (BGP
   Identifier) has been added to the OPEN message. New text (Section
   6.8) has been added to specify the procedure for detecting and
   recovering from collision.

   The new document no longer restricts the border router that is passed
   in the NEXT_HOP path attribute to be part of the same Autonomous
   System as the BGP Speaker.

   New document optimizes and simplifies the exchange of the information
   about previously reachable routes.

Appendix 4.  Comparison with RFC 1105

   All of the changes listed in Appendices 2 and 3, plus the following.

   Minor changes to the RFC1105 Finite State Machine were necessary to
   accommodate the TCP user interface provided by 4.3 BSD.

   The notion of Up/Down/Horizontal relations present in RFC1105 has
   been removed from the protocol.



Rekhter & Li                                                   [Page 51]


RFC 1654                         BGP-4                         July 1994


   The changes in the message format from RFC1105 are as follows:

      1.  The Hold Time field has been removed from the BGP header and
      added to the OPEN message.

      2.  The version field has been removed from the BGP header and
      added to the OPEN message.

      3.  The Link Type field has been removed from the OPEN message.

      4.  The OPEN CONFIRM message has been eliminated and replaced with
      implicit confirmation provided by the KEEPALIVE message.

      5.  The format of the UPDATE message has been changed
      significantly.  New fields were added to the UPDATE message to
      support multiple path attributes.

      6.  The Marker field has been expanded and its role broadened to
      support authentication.

      Note that quite often BGP, as specified in RFC 1105, is referred
      to as BGP-1, BGP, as specified in RFC 1163, is referred to as
      BGP-2, BGP, as specified in RFC1267 is referred to as BGP-3, and
      BGP, as specified in this document is referred to as BGP-4.

Appendix 5.  TCP options that may be used with BGP

   If a local system TCP user interface supports TCP PUSH function, then
   each BGP message should be transmitted with PUSH flag set.  Setting
   PUSH flag forces BGP messages to be transmitted promptly to the
   receiver.

   If a local system TCP user interface supports setting precedence for
   TCP connection, then the BGP transport connection should be opened
   with precedence set to Internetwork Control (110) value (see also
   [6]).

Appendix 6.  Implementation Recommendations

   This section presents some implementation recommendations.

6.1 Multiple Networks Per Message

   The BGP protocol allows for multiple networks with the same AS path
   and next-hop gateway to be specified in one message. Making use of
   this capability is highly recommended. With one network per message
   there is a substantial increase in overhead in the receiver. Not only
   does the system overhead increase due to the reception of multiple



Rekhter & Li                                                   [Page 52]


RFC 1654                         BGP-4                         July 1994


   messages, but the overhead of scanning the routing table for updates
   to BGP peers and other routing protocols (and sending the associated
   messages) is incurred multiple times as well. One method of building
   messages containing many networks per AS path and gateway from a
   routing table that is not organized per AS path is to build many
   messages as the routing table is scanned. As each network is
   processed, a message for the associated AS path and gateway is
   allocated, if it does not exist, and the new network is added to it.
   If such a message exists, the new network is just appended to it. If
   the message lacks the space to hold the new network, it is
   transmitted, a new message is allocated, and the new network is
   inserted into the new message. When the entire routing table has been
   scanned, all allocated messages are sent and their resources
   released.  Maximum compression is achieved when all networks share a
   gateway and common path attributes, making it possible to send many
   networks in one 4096-byte message.

   When peering with a BGP implementation that does not compress
   multiple networks into one message, it may be necessary to take steps
   to reduce the overhead from the flood of data received when a peer is
   acquired or a significant network topology change occurs. One method
   of doing this is to limit the rate of updates. This will eliminate
   the redundant scanning of the routing table to provide flash updates
   for BGP peers and other routing protocols. A disadvantage of this
   approach is that it increases the propagation latency of routing
   information.  By choosing a minimum flash update interval that is not
   much greater than the time it takes to process the multiple messages
   this latency should be minimized. A better method would be to read
   all received messages before sending updates.

6.2  Processing Messages on a Stream Protocol

   BGP uses TCP as a transport mechanism.  Due to the stream nature of
   TCP, all the data for received messages does not necessarily arrive
   at the same time. This can make it difficult to process the data as
   messages, especially on systems such as BSD Unix where it is not
   possible to determine how much data has been received but not yet
   processed.

   One method that can be used in this situation is to first try to read
   just the message header. For the KEEPALIVE message type, this is a
   complete message; for other message types, the header should first be
   verified, in particular the total length. If all checks are
   successful, the specified length, minus the size of the message
   header is the amount of data left to read. An implementation that
   would "hang" the routing information process while trying to read
   from a peer could set up a message buffer (4096 bytes) per peer and
   fill it with data as available until a complete message has been



Rekhter & Li                                                   [Page 53]


RFC 1654                         BGP-4                         July 1994


   received.

6.3 Reducing route flapping

   To avoid excessive route flapping a BGP speaker which needs to
   withdraw a destination and send an update about a more specific or
   less specific route shall combine them into the same UPDATE message.

6.4 BGP Timers

   BGP employs five timers: ConnectRetry, Hold Time, KeepAlive,
   MinASOriginationInterval, and MinRouteAdvertisementInterval The
   suggested value for the ConnectRetry timer is 120 seconds.  The
   suggested value for the Hold Time is 90 seconds.  The suggested value
   for the KeepAlive timer is 30 seconds.  The suggested value for the
   MinASOriginationInterval is 15 seconds.  The suggested value for the
   MinRouteAdvertisementInterval is 30 seconds.

   An implementation of BGP MUST allow these timers to be configurable.

6.5 Path attribute ordering

   Implementations which combine update messages as described above in
   6.1 may prefer to see all path attributes presented in a known order.
   This permits them to quickly identify sets of attributes from
   different update messages which are semantically identical.  To
   facilitate this, it is a useful optimization to order the path
   attributes according to type code.  This optimization is entirely
   optional.

6.6 AS_SET sorting

   Another useful optimization that can be done to simplify this
   situation is to sort the AS numbers found in an AS_SET.  This
   optimization is entirely optional.

6.7 Control over version negotiation

   Since BGP-4 is capable of carrying aggregated routes which cannot be
   properly represented in BGP-3, an implementation which supports BGP-4
   and another BGP version should provide the capability to only speak
   BGP-4 on a per-peer basis.

6.8 Complex AS_PATH aggregation

   An implementation which chooses to provide a path aggregation
   algorithm which retains significant amounts of path information may
   wish to use the following procedure:



Rekhter & Li                                                   [Page 54]


RFC 1654                         BGP-4                         July 1994


      For the purpose of aggregating AS_PATH attributes of two routes,
      we model each AS as a tuple <type, value>, where "type" identifies
      a type of the path segment the AS belongs to (e.g., AS_SEQUENCE,
      AS_SET), and "value" is the AS number.  Two ASs are said to be the
      same if their corresponding <type, value> tuples are the same.

      The algorithm to aggregate two AS_PATH attributes works as
      follows:

         a) Identify the same ASs (as defined above) within each AS_PATH
         attribute that are in the same relative order within both
         AS_PATH attributes.  Two ASs, X and Y, are said to be in the
         same order if either:
            - X precedes Y in both AS_PATH attributes, or - Y precedes X
            in both AS_PATH attributes.

         b) The aggregated AS_PATH attribute consists of ASs identified
         in (a) in exactly the same order as they appear in the AS_PATH
         attributes to be aggregated. If two consecutive ASs identified
         in (a) do not immediately follow each other in both of the
         AS_PATH attributes to be aggregated, then the intervening ASs
         (ASs that are between the two consecutive ASs that are the
         same) in both attributes are combined into an AS_SET path
         segment that consists of the intervening ASs from both AS_PATH
         attributes; this segment is then placed in between the two
         consecutive ASs identified in (a) of the aggregated attribute.
         If two consecutive ASs identified in (a) immediately follow
         each other in one attribute, but do not follow in another, then
         the intervening ASs of the latter are combined into an AS_SET
         path segment; this segment is then placed in between the two
         consecutive ASs identified in (a) of the aggregated attribute.

      If as a result of the above procedure a given AS number appears
      more than once within the aggregated AS_PATH attribute, all, but
      the last instance (rightmost occurrence) of that AS number should
      be removed from the aggregated AS_PATH attribute.

References

   [1] Mills, D., "Exterior Gateway Protocol Formal Specification", STD
       18, RFC 904, BBN, April 1984.

   [2] Rekhter, Y., "EGP and Policy Based Routing in the New NSFNET
       Backbone", RFC 1092, T.J. Watson Research Center, February 1989.

   [3] Braun, H-W., "The NSFNET Routing Architecture", RFC 1093,
       MERIT/NSFNET Project, February 1989.




Rekhter & Li                                                   [Page 55]


RFC 1654                         BGP-4                         July 1994


   [4] Postel, J., "Transmission Control Protocol - DARPA Internet
       Program Protocol Specification", RFC 793, DARPA, September 1981.

   [5] Rekhter, Y., and P. Gross, "Application of the Border Gateway
       Protocol in the Internet", T.J. Watson Research Center, IBM
       Corp., ANS, RFC 1655, T.J. Watson Research Center, MCI, July
       1994.

   [6] Postel, J., "Internet Protocol - DARPA Internet Program Protocol
       Specification", STD 5, RFC 791, DARPA, September 1981.

   [7] "Information Processing Systems - Telecommunications and
       Information Exchange between Systems - Protocol for Exchange of
       Inter-domain Routeing Information among Intermediate Systems to
       Support Forwarding of ISO 8473 PDUs", ISO/IEC IS10747, 1993

   [8] Fuller, V., Li, T., Yu, J., and K. Varadhan, "Classless Inter-
       Domain Routing (CIDR): an Address Assignment and Aggregation
       Strategy", RFC 1519, BARRNet, cisco, MERIT, OARnet, September
       1993.

   [9] Rekhter, Y., and T. Li, "An Architecture for IP Address
       Allocation with CIDR", RFC 1518, T.J. Watson Research Center,
       cisco, September 1993.

Security Considerations

   Security issues are not discussed in this memo.

Editors' Addresses

   Yakov Rekhter
   T.J. Watson Research Center IBM Corporation
   P.O. Box 218
   Yorktown Heights, NY 10598

   Phone:  (914) 945-3896
   EMail:  yakov@watson.ibm.com


   Tony Li
   cisco Systems, Inc.
   1525 O'Brien Drive
   Menlo Park, CA 94025

   EMail: tli@cisco.com





Rekhter & Li                                                   [Page 56]