Network Working Group                                          P. Traina
Request for Comments: 1656                                 cisco Systems
Category: Informational                                        July 1994


    BGP-4 Protocol Document Roadmap and Implementation Experience

Status of this Memo

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

Introduction

  Border Gateway Protocol v4 (BGP-4) [1] is an inter-Autonomous System
  routing protocol.  It is built on experience gained with BGP as
  defined in RFC-1267 [2] and BGP usage in the connected Internet as
  described in RFC-1268 [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
  inter-domain 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 [4].

  The management information base has been defined [5] and security
  considerations are discussed in the protocol definition document [1].

Applicability Statement for BGP-4

  BGP-4 is explicitly designed for carrying reachability information
  between Autonomous Systems.  BGP-4 is not intended to replace
  interior gateway protocols such as OSPF [7] or RIP [6].

Implementations

  Four vendors have developed independent implementations at the time
  of this memo:



Traina                                                          [Page 1]

RFC 1656                  BGP-4 Implementation                 July 1994


       ANS (gated)
       Europanet
       3COM
       cisco

  The complete interoperability matrix between all known
  implementations of various versions of BGP is available under
  separate cover [9].

Implementation Testing

  One implementation has been extensively tested in a network designed
  to mirror the complex connectivity present at many major Internet
  borders.  This network consists of multiple BGP-3 and BGP-4 speakers
  carrying full routing information injected from Alternet, EBone,
  Sprint, CERFnet, and cisco.  In many cases additional AS adjacencies
  are simulated via the use of IP over IP tunnels to increase the
  complexity of the routing topology.

  The primary feature of BGP-4 is the ability to carry network
  reachability information without regard to classfull routing.  In
  addition to canonical routing information,  CIDR prefixes (both
  supernets and subnets) are being injected from IGP information and
  aggregated using the methods described in BGP-4.  AS set aggregation
  and policy decisions based upon AS sets have been tested.

  Secondary extensions incorporated as part of version 4 of this
  protocol include enhancements to use of the INTER_AS_METRIC (now
  called MULTI_EXIT_DISC), the addition of a LOCAL_PREF parameter to
  influence route selection within an AS,  and a specified method of
  damping route fluctuations.  All of these features have been tested
  in at least one implementation.

Observations

  All implementations, are able to carry and exchange network
  reachability information.

  Not all implementations are capable of generating aggregate
  information based upon the existence of more specific routes.

  No implementation supports automatic deaggregation (enumeration of
  all networks in an aggregate block for backwards compatibility with
  routing protocols that do not carry mask information (e.g. BGP-3)).
  However, most implementations do allow for staticly configured
  controlled deaggregation for minimal backwards compatibility with
  non-CIDR capable routers.




Traina                                                          [Page 2]

RFC 1656                  BGP-4 Implementation                 July 1994


  At least one implementation capable of running earlier versions of
  BGP deliberately does not automaticly negotiate to earlier versions.
  Connections to BGP-4 peers must be explicitly configured as such.

Conclusions

  The ability to carry and inject natural networks and CIDR supernets
  is the immediate requirement for BGP-4.  The ability to carry subnet
  information (useful when reassigning parts of class A networks to
  organizations with different routing policies) is of secondary
  concern.

  The ability to conditionally aggregate routing information may be
  worked around by injecting static or IGP network information into
  BGP, or aggregation may be performed by an upstream router that is
  capable.

  Deaggregation is dangerous.  It leads to information loss and unless
  tightly controlled by a manual mechanism,  will create a routing
  information explosion.

  Automatic version negotiation is dangerous due to the state-less
  nature.  Given packet losses or spontaneous restarts,  it is possible
  for two BGP peers capable of BGP-4 to negotiate a BGP-3 or BGP-2
  connection,  which is incapable of carrying super/subnet reachability
  information and AS set information.

Acknowledgments

  The author would like to acknowledge Yakov Rekhter (IBM) and Tony Li
  (cisco) for their advice, encouragement and insightful comments.

References

  [1] Rekhter, Y., and T. Li, "A Border Gateway Protocol 4 (BGP-4), RFC
      1654, cisco Systems, T.J. Watson Research Center, IBM Corp., July
      1994.

  [2] Lougheed K., and Y. Rekhter, "A Border Gateway Protocol 3 (BGP-
      3)", RFC 1267, cisco Systems, T.J. Watson Research Center, IBM
      Corp., October 1991.

  [3] Gross P., and Y. Rekhter, "Application of the Border Gateway
      Protocol in the Internet", RFC 1268, T.J. Watson Research Center,
      IBM Corp., ANS, October 1991.






Traina                                                          [Page 3]

RFC 1656                  BGP-4 Implementation                 July 1994


  [4] Fuller V., Li. T, Yu J., and K. Varadhan, "Supernetting: an
      Address Assignment and Aggregation Strategy", Work in Progress.
      [Note: This is an expired draft, and is also referred to in
      BGP4.6.]

  [5] Willis S., Burruss J., and J. Chu, "Definitions of Managed
      Objects for the Border Gateway Protocol (Version 4) using SMIv2",
      RFC 1657, Wellfleet Communications Inc., IBM Corp., July 1994.

  [6] Hedrick, C., "Routing Information Protocol", RFC 1058, Rutgers
      University, June 1988.

  [7] Moy J., "Open Shortest Path First Routing Protocol (Version 2)",
      RFC 1583, Proteon, March 1994.

  [8] Varadhan, K., Hares S., and Y. Rekhter, "BGP4/IDRP for IP---OSPF
      Interaction", Work in Progress, September 1993.

  [9] Li, T., and P. Traina, "BGP Interoperabilty Matrix", Work in
      Progress, November 1993.

Security Considerations

  Security issues are not discussed in this memo.

Author's  Address

  Paul Traina
  cisco Systems, Inc.
  1525 O'Brien Drive
  Menlo Park, CA 94025

  EMail: [email protected]


















Traina                                                          [Page 4]


You are viewing proxied material from gopher.fnord.one. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.