WDIFF

draft-ietf-idr-bgp4-17.txt --> draft-ietf-idr-bgp4-19.txt

Network Working Group Y. Rekhter
INTERNET DRAFT Juniper Networks
T. Li
Procket Networks, Inc.
S. Hares
NextHop Technologies, Inc.
Editors

A Border Gateway Protocol 4 (BGP-4)
<draft-ietf-idr-bgp4-17.txt>
<draft-ietf-idr-bgp4-19.txt>

Status of this Memo

This document is an Internet-Draft and is in full conformance with
all provisions of Section 10 of RFC2026.

Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.

Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as ``work in progress.''

The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt

The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.

1. Acknowledgments

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

We would like to express our thanks to Guy Almes, Len Bosack, and
Jeffrey C. Honig for their contributions to the earlier version of
this document.

We like to explicitly thank Bob Braden for the review of the earlier
version

Specification of Requirements

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as well as his constructive and valuable
comments. described in RFC2119 [RFC2119].

Expiration Date July 2002 September 2003 [Page 1]

RFC DRAFT January 2002

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 earlier 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 IDR
Working Group with Yakov Rekhter and Tony Li as editors. Certain
sections March 2003

TTaabbllee ooff CCoonntteennttss

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1. Definition of the document borrowed heavily from IDRP [7], which is the
OSI counterpart commonly used terms . . . . . . . . . . . . . . 4
2. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 6
3. Summary of BGP. For this credit should be given to the ANSI
X3S3.3 group chaired by Lyman Chapin Operation . . . . . . . . . . . . . . . . . . . . . 7
3.1 Routes: Advertisement and to Charles Kunzinger who was
the IDRP editor within that group. We would also like to thank Enke
Chen, Edward Crabbe, Mike Craren, Vincent Gillet, Eric Gray, Jeffrey
Haas, Dimitry Haskin, John Krawczyk, David LeRoy, Dan Massey, Dan
Pei, Mathew Richardson, John Scudder, John Stewart III, Dave Thaler,
Paul Storage . . . . . . . . . . . . . . 9
3.2 Routing Information Bases . . . . . . . . . . . . . . . . . . 10
4. Message Formats . . . . . . . . . . . . . . . . . . . . . . . 11
4.1 Message Header Format . . . . . . . . . . . . . . . . . . . . 11
4.2 OPEN Message Format . . . . . . . . . . . . . . . . . . . . . 12
4.3 UPDATE Message Format . . . . . . . . . . . . . . . . . . . . 14
4.4 KEEPALIVE Message Format . . . . . . . . . . . . . . . . . . 21
4.5 NOTIFICATION Message Format . . . . . . . . . . . . . . . . . 21
5. Path Attributes . . . . . . . . . . . . . . . . . . . . . . . 23
5.1 Path Attribute Usage . . . . . . . . . . . . . . . . . . . . 25
5.1.1 ORIGIN . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.1.2 AS_PATH . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.1.3 NEXT_HOP . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.1.4 MULTI_EXIT_DISC . . . . . . . . . . . . . . . . . . . . . . 28
5.1.5 LOCAL_PREF . . . . . . . . . . . . . . . . . . . . . . . . 28
5.1.6 ATOMIC_AGGREGATE . . . . . . . . . . . . . . . . . . . . . 29
5.1.7 AGGREGATOR . . . . . . . . . . . . . . . . . . . . . . . . 30
6. BGP Error Handling . . . . . . . . . . . . . . . . . . . . . . 30
6.1 Message Header error handling . . . . . . . . . . . . . . . . 30
6.2 OPEN message error handling . . . . . . . . . . . . . . . . . 31
6.3 UPDATE message error handling . . . . . . . . . . . . . . . . 32
6.4 NOTIFICATION message error handling . . . . . . . . . . . . . 34
6.5 Hold Timer Expired error handling . . . . . . . . . . . . . . 34
6.6 Finite State Machine error handling . . . . . . . . . . . . . 34
6.7 Cease . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.8 BGP connection collision detection . . . . . . . . . . . . . 35
7. BGP Version Negotiation . . . . . . . . . . . . . . . . . . . 36
8. BGP Finite State machine . . . . . . . . . . . . . . . . . . . 36
8.1 Events for the BGP FSM . . . . . . . . . . . . . . . . . . . 37
8.1.1 Administrative Events . . . . . . . . . . . . . . . . . . 37
8.1.2 Timer Events . . . . . . . . . . . . . . . . . . . . . . . 38
8.1.3 TCP connection based Events . . . . . . . . . . . . . . . . 39
8.1.4 BGP Messages based Events . . . . . . . . . . . . . . . . . 41
8.2 Description of FSM . . . . . . . . . . . . . . . . . . . . . 43
8.2.1 FSM Definition . . . . . . . . . . . . . . . . . . . . . . 43
8.2.1.1 Terms "active" and "passive" . . . . . . . . . . . . . . 43
8.2.1.2 FSM and collision detection . . . . . . . . . . . . . . . 44
8.2.2 Finite State Machine . . . . . . . . . . . . . . . . . . . 44
9. UPDATE Message Handling . . . . . . . . . . . . . . . . . . . 57
9.1 Decision Process . . . . . . . . . . . . . . . . . . . . . . 58
9.1.1 Phase 1: Calculation of Degree of Preference . . . . . . . 59

Expiration Date September 2003 [Page 2]

RFC DRAFT March 2003

9.1.2 Phase 2: Route Selection . . . . . . . . . . . . . . . . . 60
9.1.2.1 Route Resolvability Condition . . . . . . . . . . . . . . 61
9.1.2.2 Breaking Ties (Phase 2) . . . . . . . . . . . . . . . . . 62
9.1.3 Phase 3: Route Dissemination . . . . . . . . . . . . . . . 64
9.1.4 Overlapping Routes . . . . . . . . . . . . . . . . . . . . 65
9.2 Update-Send Process . . . . . . . . . . . . . . . . . . . . . 66
9.2.1 Controlling Routing Traffic Overhead . . . . . . . . . . . 67
9.2.1.1 Frequency of Route Advertisement . . . . . . . . . . . . 67
9.2.1.2 Frequency of Route Origination . . . . . . . . . . . . . 68
9.2.2 Efficient Organization of Routing Information . . . . . . . 68
9.2.2.1 Information Reduction . . . . . . . . . . . . . . . . . . 68
9.2.2.2 Aggregating Routing Information . . . . . . . . . . . . . 69
9.3 Route Selection Criteria . . . . . . . . . . . . . . . . . . 72
9.4 Originating BGP routes . . . . . . . . . . . . . . . . . . . 72
10. BGP Timers . . . . . . . . . . . . . . . . . . . . . . . . . 72
Appendix A. Comparison with RFC1771 . . . . . . . . . . . . . . . 73
Appendix B. Comparison with RFC1267 . . . . . . . . . . . . . . . 74
Appendix C. Comparison with RFC 1163 . . . . . . . . . . . . . . 75
Appendix D. Comparison with RFC 1105 . . . . . . . . . . . . . . 75
Appendix E. TCP options that may be used with BGP . . . . . . . . 76
Appendix F. Implementation Recommendations . . . . . . . . . . . 76
Appendix F.1 Multiple Networks Per Message . . . . . . . . . . . 76
Appendix F.2 Reducing route flapping . . . . . . . . . . . . . . 77
Appendix F.3 Path attribute ordering . . . . . . . . . . . . . . 77
Appendix F.4 AS_SET sorting . . . . . . . . . . . . . . . . . . . 77
Appendix F.5 Control over version negotiation . . . . . . . . . . 78
Appendix F.6 Complex AS_PATH aggregation . . . . . . . . . . . . 78
Security Considerations . . . . . . . . . . . . . . . . . . . . . 79
IANA Considerations . . . . . . . . . . . . . . . . . . . . . . . 79
References . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Authors Information . . . . . . . . . . . . . . . . . . . . . . . 80

Expiration Date September 2003 [Page 3]

RFC DRAFT March 2003

Abstract

The Border Gateway Protocol (BGP) is an inter-Autonomous System rout-
ing protocol.

The primary function of a BGP speaking system is to exchange network
reachability information with other BGP systems. This network reacha-
bility information includes information on the list of Autonomous
Systems (ASs) that reachability information traverses. This informa-
tion 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 set of mechanisms for supporting Classless Inter-
Domain Routing (CIDR) [RFC1518, RFC1519]. These mechanisms include
support for advertising a set of destinations as an IP prefix and
eliminating the concept of network "class" within BGP. BGP-4 also
introduces mechanisms which allow aggregation of routes, including
aggregation of AS paths.

1. Definition of commonly used terms

This section provides definition for terms that have a specific mean-
ing to the BGP protocol and that are used throughout the text.

Autonomous System (AS)
The classic definition of an Autonomous System is a set of routers
under a single technical administration, using an interior gateway
protocol (IGP) and common metrics to determine how to route pack-
ets within the AS, and using an inter-AS routing protocol to
determine how to route packets to other ASs. Since this classic
definition was developed, it has become common for a single AS to
use several IGPs 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 adminis-
tration of an AS appears to other ASs to have a single coherent
interior routing plan and presents a consistent picture of what
destinations are reachable through it.

Expiration Date September 2003 [Page 4]

RFC DRAFT March 2003

BGP speaker
A router that implements BGP.

BGP Identifier
A 4-octet unsigned integer indicating the BGP Identifier of the
sender of BGP messages. 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.

Internal peer
Peer that is in the same Autonomous System as the local system.

IBGP
Internal BGP (BGP connection between internal peers).

External peer
Peer that is in a different Autonomous System than the local sys-
tem.

EBGP
External BGP (BGP connection between external peers).

NLRI
Network Layer Reachability Information.

Route
A unit of information that pairs a set of destinations with the
attributes of a path to those destinations. The set of destina-
tions are systems whose IP addresses are contained in one IP
address prefix carried in the Network Layer Reachability Informa-
tion (NLRI) field of an UPDATE message. The path is the informa-
tion reported in the path attributes field of the same UPDATE mes-
sage.

RIB
Routing Information Base.

Adj-RIB-In
The Adj-RIBs-In contain unprocessed routing information that has
been advertised to the local BGP speaker by its peers.

Loc-RIB
The Loc-RIB contains the routes that have been selected by the
local BGP speaker's Decision Process.

Adj-RIB-Out
The Adj-RIBs-Out contains the routes for advertisement to specific

Expiration Date September 2003 [Page 5]

RFC DRAFT March 2003

peers by means of the local speaker's UPDATE messages.

IGP
Interior Gateway Protocol - a routing protocol used to exchange
routing information among routers within a single Autonomous Sys-
tem.

Feasible route
A route that is available for use.

Unfeasible route
A previously advertised feasible route that is no longer available
for use.

2. Acknowledgments

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

We would like to express our thanks to Guy Almes, Len Bosack, and
Jeffrey C. Honig for their contributions to the earlier version
(BGP-1) of this document.

We would like to specially acknowledge numerous contributions by Den-
nis Ferguson to the earlier version of this document.

We like to explicitly thank Bob Braden for the review of the earlier
version (BGP-2) of this document as well as his constructive and
valuable comments.

Certain sections of the document borrowed heavily from IDRP
[IS10747], which is the OSI counterpart of BGP. For this credit
should be given to the ANSI X3S3.3 group chaired by Lyman Chapin and
to Charles Kunzinger who was the IDRP editor within that group.

We would also like to thank Benjamin Abarbanel, Enke Chen, Edward
Crabbe, Mike Craren, Vincent Gillet, Eric Gray, Jeffrey Haas, Dimitry
Haskin, John Krawczyk, David LeRoy, Dan Massey, Jonathan Natale, Dan
Pei, Mathew Richardson, John Scudder, John Stewart III, Dave Thaler,

Expiration Date September 2003 [Page 6]

RFC DRAFT March 2003

Paul Traina, Russ White, Curtis Villamizar, and Alex Zinin for their
comments.

Many thanks Alex Zinin for their
comments.

We would like to specially acknowledge Andrew Lange for his help in
preparing the final version of this document.

Finally, we would like to thank all the members of the IDR Working
Group for their ideas and support they have given to this document.

3. Summary of Operation

The Border Gateway Protocol (BGP) is an inter-Autonomous System rout-
ing protocol. It is built on experience gained with EGP as defined in
[RFC904] and EGP usage in the NSFNET Backbone as described in
[RFC1092] and [RFC1093].

In the context of this document we assume that a BGP speaker adver-
tises to its peers only those routes that it itself uses (in this
context a BGP speaker is said to "use" a BGP route if it is the most
preferred BGP route and is used in forwarding). All other cases are
outside the scope of this document.

In the context of this document the term "IP address" refers to an IP
Version 4 address [RFC791].

Routing information exchanged via BGP supports only the destination-
based forwarding paradigm, which assumes that a router forwards a
packet based solely on the destination address carried in the IP
header of the packet. This, in turn, reflects the set of policy deci-
sions that can (and can not) be enforced using BGP. Note that some
policies can not be supported by the destination-based forwarding
paradigm, and thus require techniques such as source routing (aka
explicit routing) to be enforced. Such policies can not be enforced
using BGP either. For example, BGP does not enable one AS to send
traffic to a neighboring AS for forwarding to some destination
(reachable through but) beyond that neighboring AS intending that the
traffic take a different route to that taken by the traffic originat-
ing in the neighboring AS (for that same destination). On the other
hand, BGP can support any policy conforming to the destination-based

Expiration Date September 2003 [Page 7]

RFC DRAFT March 2003

forwarding paradigm.

BGP-4 provides a new set of mechanisms for supporting Classless
Inter-Domain Routing (CIDR) [RFC1518, RFC1519]. These mechanisms
include support for advertising a set of destinations as an IP prefix
and eliminating the concept of network "class" within BGP. BGP-4
also introduces mechanisms which allow aggregation of routes, includ-
ing aggregation of AS paths.

This document 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
(IGP) and common metrics to determine how to route packets within the
AS, and using an inter-AS routing protocol to determine how to route
packets to other ASs. Since this classic definition was developed, it
has become common for a single AS to use several IGPs 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 met-
rics are used, the administration of an AS appears to other ASs to
have a single coherent interior routing plan and presents a consis-
tent picture of what destinations are reachable through it.

BGP uses TCP [RFC793] as its transport protocol. This eliminates the
need to implement explicit update fragmentation, retransmission,
acknowledgment, and sequencing. BGP listens on TCP port 179. The
error notification mechanism used in BGP assumes that TCP supports a
"graceful" close, i.e., that all outstanding data will be delivered
before the connection is closed.

Two systems form a TCP connection between one another. They exchange
messages to open and confirm the connection parameters.

The initial data flow is the portion of the BGP routing table that is
allowed by the export policy, called the Adj-Ribs-Out (see 3.2).
Incremental updates are sent as the routing tables change. BGP does
not require periodic refresh of the routing table. To allow local
policy changes to have the correct effect without resetting any BGP
connections, a BGP speaker SHOULD either (a) retain the current ver-
sion of the routes advertised to it by all of its peers for the dura-
tion of the connection, or (b) make use of the Route Refresh exten-
sion [RFC2918].

KEEPALIVE messages may be 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 closed.

A peer in a different AS is referred to as an external peer, while a

Expiration Date September 2003 [Page 8]

RFC DRAFT March 2003

peer in the same AS is referred to as an internal peer. Internal BGP
and external BGP are commonly abbreviated IBGP and EBGP.

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 IGP used within the AS. For the
purpose of this document, it is assumed that a consistent view of the
routes exterior to the AS is provided by having all BGP speakers
within the AS maintain IBGP with each other. 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 tran-
sit service is being provided.

This document specifies the base behavior of the BGP protocol. This
behavior can and is modified by extention specifications. When the
protocol is extended the new behavior is fully documented in the
extention specifications.

3.1 Routes: Advertisement and Storage

For the purpose of this protocol, a route is defined as a unit of
information that pairs a set of destinations with the attributes of a
path to those destinations. The set of destinations are systems whose
IP addresses are contained in one IP address prefix carried in the
Network Layer Reachability Information (NLRI) field of an UPDATE mes-
sage, and the path is the information reported in the path attributes
field of the same UPDATE message.

Routes are advertised between BGP speakers in UPDATE messages. Mul-
tiple routes that have the same path attributes can be advertised in
a single UPDATE message by including multiple prefixes in the NLRI
field of the UPDATE message.

Routes are stored in the Routing Information Bases (RIBs): namely,
the Adj-RIBs-In, the Loc-RIB, and the Adj-RIBs-Out, as described in
Section 3.2.

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:

Expiration Date September 2003 [Page 9]

RFC DRAFT March 2003

a) the IP prefix that expresses the destination 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 NLRI 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.

Changing attribute of a route is accomplished by advertising a
replacement route. The replacement route carries new (changed)
attributes and has the same NLRI as the original route.

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 received from other BGP
speakers. 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. These are
the routes that will be used by the local BGP speaker. The next
hop for each of these routes MUST be resolvable via the local BGP
speaker's Routing Table.

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
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

Expiration Date September 2003 [Page 10]

RFC DRAFT March 2003

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.

Routing information that the BGP speaker uses to forward packets (or
to construct the forwarding table that is used for packet forwarding)
is maintained in the Routing Table. The Routing Table accumulates
routes to directly connected networks, static routes, routes learned
from the IGP protocols, and routes learned from BGP. Whether or not
a specific BGP route should be installed in the Routing Table, and
whether a BGP route should override a route to the same destination
installed by another source is a local policy decision, not specified
in this document. Besides actual packet forwarding, the Routing Table
is used for resolution of the next-hop addresses specified in BGP
updates (see Section 5.1.3).

4. Message Formats

This section describes message formats used by BGP.

BGP messages are sent over a TCP 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 mes-
sage size. The smallest message that may be sent consists of a BGP
header without a data portion, or 19 octets.

All multi-octet fields are in network byte order.

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 lay-
out of these fields is shown below:

Expiration Date September 2003 [Page 11]

RFC DRAFT March 2003

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 is included for compatibility; it MUST be
set to all ones.

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 TCP 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. This document defines the following type codes:

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

[RFC2918] defines one more type code.

4.2 OPEN Message Format

After a TCP is established, the first message sent by each side is an
OPEN message. If the OPEN message is acceptable, a KEEPALIVE message

Expiration Date September 2003 [Page 12]

RFC DRAFT March 2003

confirming the OPEN is sent back.

In addition to the fixed-size BGP header, the OPEN 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
+-+-+-+-+-+-+-+-+
| Version |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| My Autonomous System |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Hold Time |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| BGP Identifier |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Opt Parm Len |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
| Optional Parameters (variable) |
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

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:

Expiration Date September 2003 [Page 13]

RFC DRAFT March 2003

This 4-octet unsigned integer indicates the BGP Identifier of
the sender. A given BGP speaker sets the value of its BGP Iden-
tifier 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.

Optional Parameters Length:

This 1-octet unsigned integer indicates the total length of the
Optional Parameters field in octets. If the value of this field
is zero, no Optional Parameters are present.

Optional Parameters:

This field contains a list of optional parameters, where each
parameter is encoded as a <Parameter Type, Parameter Length,
Parameter Value> triplet.

0 1
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
| Parm. Type | Parm. Length | Parameter Value (variable)
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...

Parameter Type is a one octet field that unambiguously identi-
fies individual parameters. Parameter Length is a one octet
field that contains the length of the Parameter Value field in
octets. Parameter Value is a variable length field that is
interpreted according to the value of the Parameter Type field.

[RFC2842] defines the Capabilities Optional Parameter.

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

4.3 UPDATE Message Format

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

An UPDATE message is used to Sue Hares for her contributions advertise feasible routes sharing common
path attributes to the document, a peer, or to withdraw multiple unfeasible routes

Expiration Date September 2003 [Page 14]

RFC DRAFT March 2003

from service (see 3.1). An UPDATE message MAY simultaneously adver-
tise a feasible route and
especially for her work on withdraw multiple unfeasible routes from
service. The UPDATE message always includes the fixed-size BGP Finite State Machine.

We would like
header, and also includes the other fields as shown below (note, some
of the shown fields may not be present in every UPDATE message):

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

Withdrawn Routes Length:

This 2-octets unsigned integer indicates the total length of
the Withdrawn Routes field in octets. Its value allows the
length of the Network Layer Reachability Information field to specially acknowledge numerous contributions by
Dennis Ferguson.

2. Introduction

The Border Gateway Protocol (BGP)
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 an inter-Autonomous System
routing protocol. It 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 built on experience gained with EGP encoded as
defined in RFC 904 [1] and EGP usage in a 2-tuple of the NSFNET Backbone as
form <length, prefix>, whose fields are described in RFC 1092 [2] below:

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

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

a) Length:

Expiration Date September 2003 [Page 15]

RFC 1093 [3]. DRAFT March 2003

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

b) Prefix:

The Prefix field contains an IP address prefix followed by
the minimum number of trailing bits needed to exchange network
reachability information with other BGP systems. This network
reachability information includes information on make the list end
of
Autonomous Systems (ASs) the field fall on an octet boundary. Note that reachability information traverses.
This information the value
of trailing bits is sufficient to construct a graph irrelevant.

Total Path Attribute Length:

This 2-octet unsigned integer indicates the total length of AS
connectivity from which routing loops may be pruned and some policy
decisions at the AS level may
Path Attributes field in octets. Its value allows the length of
the Network Layer Reachability field to be enforced.

BGP-4 provides a new set determined as speci-
fied below.

A value of mechanisms for supporting Classless
Inter-Domain Routing (CIDR) [8, 9]. These mechanisms include support 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 message, except for advertising an IP prefix and eliminates UPDATE message that carries
only the concept of network
"class" within BGP. BGP-4 also introduces mechanisms which allow
aggregation withdrawn routes. Each path attribute is a triple
<attribute type, attribute length, attribute value> of routes, including aggregation variable
length.

Attribute Type is a two-octet field that consists of AS paths.

Expiration Date July 2002 [Page 2]

RFC DRAFT January 2002

To characterize the set of policy decisions that can be enforced
using BGP, one must focus on
Attribute Flags octet followed by the rule that a BGP speaker advertises
to its peers (other BGP speakers which it communicates with) in
neighboring ASs only those routes that it itself uses. This rule
reflects 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 "hop-by-hop" routing paradigm generally used throughout Attribute Flags octet is the current Internet. Note that some policies cannot be supported by
Optional bit. It defines whether the "hop-by-hop" routing paradigm and thus require techniques such as
source routing (aka explicit routing) to enforce. For example, BGP
does not enable one AS attribute is optional (if
set to send traffic 1) or well-known (if set to a neighboring AS intending
that the traffic take a different route from that taken by traffic
originating in 0).

The second high-order bit (bit 1) of the neighboring AS. On Attribute Flags octet
is the other hand, BGP can support
any policy conforming Transitive bit. It defines whether an optional attribute
is transitive (if set to 1) or non-transitive (if set to 0).
For well-known attributes, the "hop-by-hop" routing paradigm. Since the
current Internet uses only the "hop-by-hop" inter-AS routing paradigm
and since BGP can support any policy that conforms Transitive bit MUST be set to that paradigm,
BGP is highly applicable as an inter-AS routing protocol 1.

Expiration Date September 2003 [Page 16]

RFC DRAFT March 2003

(See Section 5 for the
current Internet.

A more complete a discussion of what policies can and cannot be
enforced with BGP transitive attributes.)

The third high-order bit (bit 2) of the Attribute Flags octet
is outside the scope of this document (but refer to Partial bit. It defines whether the companion document discussing BGP usage [5]).

BGP runs over a reliable transport protocol. This eliminates information con-
tained in the need optional transitive attribute is partial (if set
to implement explicit update fragmentation, retransmission,
acknowledgment, 1) or complete (if set to 0). For well-known attributes and sequencing. Any authentication scheme used by
for optional non-transitive attributes the
transport protocol (e.g., RFC2385 [10]) may Partial bit MUST be used in addition
set to
BGP's own authentication mechanisms. 0.

The error notification mechanism
used in BGP assumes that fourth high-order bit (bit 3) of the transport protocol supports a "graceful"
close, i.e., that all outstanding data will be delivered before Attribute Flags octet
is the
connection Extended Length bit. It defines whether the Attribute
Length is closed.

BGP uses TCP [4] as its transport protocol. TCP meets BGP's transport
requirements one octet (if set to 0) or two octets (if set to 1).

The lower-order four bits of the Attribute Flags octet are
unused. They MUST be zero when sent and is present MUST be ignored when
received.

The Attribute Type Code octet contains the Attribute Type Code.
Currently defined Attribute Type Codes are discussed in virtually all commercial routers and
hosts. In Section
5.

If the following descriptions Extended Length bit of the phrase "transport protocol
connection" can be understood to refer Attribute Flags octet is set
to a TCP connection. BGP uses
TCP port 179 for establishing its connections.

This document uses 0, the term `Autonomous System' (AS) throughout. The
classic definition third octet of an Autonomous System the Path Attribute contains the length
of the attribute data in octets.

If the Extended Length bit of the Attribute Flags octet is a set of routers under
a single technical administration, using an interior gateway protocol
and common metrics to determine how
to route packets within 1, then the AS,
and using an exterior gateway protocol to determine how 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 third and sometimes several sets the fourth octets of metrics within an AS. the path
attribute contain the length of the attribute data in octets.

The use remaining octets of the term Autonomous System here stresses Path Attribute represent the fact that, even when
multiple IGPs
attribute value and metrics are used, the administration of an AS
appears to other ASs interpreted according to have 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 single coherent well-known mandatory attribute that defines the
origin of the path information. The data octet can assume
the following values:

Value Meaning

0 IGP - Network Layer Reachability Information
is interior routing plan
and presents a consistent picture of what destinations are reachable to the originating AS

1 EGP - Network Layer Reachability Information
learned via the EGP protocol [RFC904]

Expiration Date July 2002 September 2003 [Page 3] 17]

RFC DRAFT January 2002

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 March 2003

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

Usage of routing policy rules this attribute is presented defined in 5.1.1.

b) AS_PATH (Type Code 2):

AS_PATH is a companion
document [5]. This document well-known mandatory attribute that is the first composed
of a series sequence of documents
planned to explore various aspects 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 fol-
lowing values defined:

Value Segment Type

1 AS_SET: unordered set of BGP application.

3. Summary ASs a route in the
UPDATE message has traversed

2 AS_SEQUENCE: ordered set of Operation

Two systems form ASs a transport protocol connection between one another.
They exchange messages to open and confirm route in
the connection parameters. UPDATE message has traversed

The initial data flow path segment length is a 1-octet long field containing
the portion number of ASs (not the BGP routing table that is
allowed by the export policy, called number of octets) in the Adj-Ribs-Out (see 3.2).
Incremental updates are sent path
segment value field.

The path segment value field contains one or more AS num-
bers, each encoded as the routing tables change. BGP does
not require periodic refresh a 2-octets long field.

Usage of the routing table. Therefore, this attribute is defined in 5.1.2.

c) NEXT_HOP (Type Code 3):

This is a BGP
speaker must retain well-known mandatory attribute that defines the current version
(unicast) IP address of the routes advertised by
all of its peers for router that SHOULD be used as
the duration of next hop to the connection. If destinations listed in the
implementation decides to not store Network Layer
Reachability Information field of the routes UPDATE message.

Usage of this attribute is defined in 5.1.3.

d) MULTI_EXIT_DISC (Type Code 4):

This is an optional non-transitive attribute that have been
received from is a peer, but have been filtered out according to
configured local policy, the BGP Route Refresh extension [12] may four
octet unsigned integer. The value of this attribute MAY be
used to request the full set of routes from by a peer without resetting
the BGP session when the local policy configuration changes.

KEEPALIVE messages may be sent periodically speaker's decision process to ensure the liveness of
the connection. NOTIFICATION messages are sent in response discriminate
among multiple entry points to errors
or special conditions. If a connection encounters an error condition, a NOTIFICATION message neighboring autonomous

Expiration Date September 2003 [Page 18]

RFC DRAFT March 2003

system.

Usage of this attribute is sent and the connection defined in 5.1.4.

e) LOCAL_PREF (Type Code 5):

LOCAL_PREF is closed.

The hosts executing the Border Gateway Protocol need not be routers. a well-known attribute that is a four octet
unsigned integer. 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 speaker uses it to exchange routing information with a border
router inform other
internal peers of the advertising speaker's degree of pref-
erence for an advertised route.

Usage of this attribute is defined in another Autonomous System. The implications and
applications 5.1.5.

f) ATOMIC_AGGREGATE (Type Code 6)

ATOMIC_AGGREGATE is a well-known discretionary attribute of
length 0.

Usage of this architecture are for further study.

Connections between BGP speakers attribute is defined in 5.1.6.

g) AGGREGATOR (Type Code 7)

AGGREGATOR is an optional transitive attribute of different ASs are referred to length 6.
The attribute contains the last AS number that formed the
aggregate route (encoded as
"external" links. BGP connections between 2 octets), followed by the IP
address of the BGP speakers within speaker that formed the aggregate route
(encoded as 4 octets). This SHOULD be the same AS are referred to address as "internal" links. Similarly, a peer in a
different AS
the one used for the BGP Identifier of the speaker.

Usage of this attribute is referred to as an external peer, while defined in 5.1.7.

Network Layer Reachability Information:

This variable length field contains a peer list of IP address pre-
fixes. The length in octets of the
same AS may Network Layer Reachability
Information is not encoded explicitly, but can be described as an internal peer. Internal calculated
as:

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

where UPDATE message Length is the value encoded in the fixed-
size BGP header, Total Path Attribute Length and
external BGP Withdrawn
Routes Length are commonly abbreviated IBGP and EBGP.

Expiration Date July 2002 [Page 4]

RFC DRAFT January 2002

routes values encoded in the variable part of
the AS UPDATE message, and 23 is provided by the interior routing protocol. A
consistent view a combined length of the routes exterior to the AS can be provided by
having all fixed-
size BGP speakers within header, the AS maintain direct IBGP
connections with each other. Alternately Total Path Attribute Length field and the interior routing
protocol can pass BGP
Withdrawn Routes Length field.

Expiration Date September 2003 [Page 19]

RFC DRAFT March 2003

Reachability information among routers within an AS, taking
care not to lose BGP attributes that will be needed by EBGP speakers
if transit connectivity is being provided. For the purpose encoded as one or more 2-tuples of
discussion, it is assumed that BGP information is passed within an AS
using IBGP. Care must be taken to ensure that the interior routers
have all been updated with transit information before
the EBGP
speakers announce to other ASs that transit service is being
provided.

3.1 Routes: Advertisement form <length, prefix>, whose fields are described below:

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

The use and Storage

For the purpose meaning of this protocol, a route is defined these fields are as a unit of
information that pairs a set follows:

a) Length:

The Length field indicates the length in bits of destinations with the attributes IP
address prefix. A length of zero indicates a
path to those destinations. The set of destinations are the systems
whose prefix that
matches all IP addresses are reported in the Network Layer Reachability
Information (NLRI) (with prefix, itself, of zero
octets).

b) Prefix:

The Prefix field and the path is contains an IP address prefix followed by
enough trailing bits to make the information reported in end of the path attributes field of fall on an
octet boundary. Note that the same UPDATE message.

Routes are advertised between BGP speakers in UPDATE messages.

Routes are stored in value of the Routing Information Bases (RIBs): namely, trailing bits is
irrelevant.

The minimum length of the Adj-RIBs-In, UPDATE message is 23 octets -- 19 octets
for the Loc-RIB, and fixed header + 2 octets for the Adj-RIBs-Out. Withdrawn Routes that will
be advertised to other BGP speakers must be present in Length + 2
octets for the Adj-RIB-
Out. Total Path Attribute Length (the value of Withdrawn
Routes that will be used by the local BGP speaker must be
present in the Loc-RIB, Length is 0 and the next hop for each value of these routes
must be resolvable via the local BGP speaker's Routing Table. Routes
that are received from other BGP speakers are present in the Adj-
RIBs-In.

If a BGP speaker chooses to Total Path Attribute Length is
0).

An UPDATE message can advertise at most one set of path attributes,
but multiple destinations, provided that the route, it may add to or
modify the destinations share these
attributes. All 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 contained in a given BGP speaker can indicate
that a route has been withdrawn from service:

a) the IP prefix that expresses the destination for a previously
advertised route can be advertised UPDATE message
apply to all destinations carried in the WITHDRAWN ROUTES NLRI field
in of the UPDATE message, thus marking the associated
message.

An UPDATE message can list multiple routes to be withdrawn from ser-
vice. Each such route is identified by its destination (expressed as being
no longer available for use

Expiration Date July 2002 [Page 5]

RFC DRAFT January 2002

b) a replacement
an IP prefix), which unambiguously identifies the route with in the same NLRI can be advertised, or

c) con-
text of 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 which it has been learned from inbound
previously advertised.

An UPDATE messages. Their contents
represent message might advertise only 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 be withdrawn from
service, in the Adj-RIBs-Out which case it will be carried not include path attributes or Network
Layer Reachability Information. Conversely, it may advertise only a
feasible route, in which case the local BGP speaker's UPDATE messages and advertised to its
peers.

Expiration Date September 2003 [Page 20]

RFC DRAFT March 2003

An UPDATE message SHOULD NOT include the Adj-RIBs-Out organize same address prefix in the routes
for advertisement
WITHDRAWN ROUTES and Network Layer Reachability Information fields,
however a BGP speaker MUST be able to specific peers by means of the local speaker's process UPDATE messages.

Although the conceptual model distinguishes between Adj-RIBs-In, Loc-
RIB, and Adj-RIBs-Out, messages in this neither implies nor requires that
form. A BGP speaker SHOULD treat an
implementation must maintain three separate copies UPDATE message of this form as if
the routing
information. The choice of implementation (for example, 3 copies of WITHDRAWN ROUTES doesn't contain the information vs 1 copy with pointers) is address prefix.

4.4 KEEPALIVE Message Format

BGP does not constrained by the
protocol.

Routing information that the router uses use any TCP-based keep-alive mechanism to forward packets (or determine if
peers are reachable. Instead, KEEPALIVE messages are exchanged
between peers often enough as not to
construct the forwarding table that is used for packet forwarding) is
maintained in cause the Routing Table. The Routing Table accumulates routes Hold Timer to directly connected networks, static routes, routes learned from expire.
A reasonable maximum time between KEEPALIVE messages would be one
third of the IGP protocols, and routes learned from BGP. Whether or not a
specific BGP route should Hold Time interval. KEEPALIVE messages MUST NOT be installed in sent
more frequently than one per second. An implementation MAY adjust the Routing Table, and
whether a BGP route should override
rate at which it sends KEEPALIVE messages as a route to function of the same destination

Expiration Date July 2002 [Page 6]

RFC DRAFT January 2002

installed by another source is a local policy decision, not specified
in this document. Besides actual packet forwarding, Hold
Time interval.

If the Routing Table negotiated Hold Time interval is used for resolution zero, then periodic KEEPALIVE
messages MUST NOT be sent.

A KEEPALIVE message consists of the next-hop addresses specified in BGP
updates (see Section 9.1.2).

4. Message Formats

This section describes only message formats used by BGP.

Messages are sent over header and has a reliable transport protocol connection. length
of 19 octets.

4.5 NOTIFICATION Message Format

A NOTIFICATION message is processed only after it sent when an error condition is entirely received. detected.
The maximum
message size BGP connection is 4096 octets. All implementations are required closed immediately after sending it.

In addition to
support this maximum the fixed-size BGP header, the NOTIFICATION message size.
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 (variable) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Error Code:

This 1-octet unsigned integer indicates the type of NOTIFICA-
TION. The smallest message that may be
sent consists following Error Codes have been defined:

Error Code Symbolic Name Reference

Expiration Date September 2003 [Page 21]

RFC DRAFT March 2003

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 informa-
tion about the nature of a BGP header without a data portion, or 19 octets.

4.1 Message Header Format the reported error. Each message has a fixed-size header. There Error Code
may have one or may not be more Error Subcodes associated with it. If no
appropriate Error Subcode is defined, then a data
portion following the header, depending on the message type. The
layout of these fields zero (Unspecific)
value is shown below:

0 1 2 3
0 used for the Error Subcode field.

Message Header Error subcodes:

1 - Connection Not Synchronized.
2 - Bad Message Length.
3 4 5 6 7 8 9 0 - Bad Message Type.

OPEN Message Error subcodes:

1 - Unsupported Version Number.
2 - Bad Peer AS.
3 - Bad BGP Identifier.
4 - Unsupported Optional Parameter.
5 - [Deprecated - see Appendix A].
6 7 8 9 0 - 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 - [Deprecated - see Appendix A].
8 - Invalid NEXT_HOP Attribute.
9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
+ +
| |
+ +
| Marker |
+ +
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Length | Type |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Marker: - Optional Attribute Error.
10 - Invalid Network Field.

Expiration Date September 2003 [Page 22]

RFC DRAFT March 2003

11 - Malformed AS_PATH.

Data:

This 16-octet variable-length field contains a value 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 receiver length of the
message Data field can predict. If be determined from
the Type message Length field by the formula:

Message Length = 21 + Data Length

The minimum length of the NOTIFICATION message is OPEN, or if
the OPEN 21 octets (includ-
ing message carries no Authentication Information (as an
Optional Parameter), then the Marker must be all ones.
Otherwise, header).

5. Path Attributes

This section discusses the value path attributes of the marker can 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 predicted recognized by some all BGP implementations.
Some of these attributes are mandatory and MUST be included in every
UPDATE message that contains NLRI. Others are discretionary and MAY
or MAY NOT be sent in a
computation specified as part particular UPDATE message.

All well-known attributes MUST be passed along (after proper updat-
ing, 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 the authentication mechanism
(which an
unrecognized optional attribute is specified as part determined by the setting of the Authentication Information)
used. The Marker can
Transitive bit in the attribute flags octet. Paths with unrecognized
transitive optional attributes SHOULD be used accepted. If a path with
unrecognized transitive optional attribute is accepted and passed
along to detect loss other BGP peers, then the unrecognized transitive optional
attribute of synchronization that path MUST be passed along with the path to other

Expiration Date July 2002 September 2003 [Page 7] 23]

RFC DRAFT January 2002

between March 2003

BGP peers with the Partial bit in the Attribute Flags octet set to 1.
If a pair of BGP peers, path with recognized transitive optional attribute is accepted
and passed along to authenticate incoming other BGP
messages.

Length:

This 2-octet unsigned integer indicates the total length of the
message, including peers and the header, Partial bit in octets. Thus, e.g., the
Attribute Flags octet is set to 1 by some previous AS, it allows
one is not set
back to locate in the transport-level stream the (Marker field
of the) next message. The value of 0 by the Length field must always current AS. Unrecognized non-transitive optional
attributes MUST be at least 19 and no greater than 4096, quietly ignored and may not passed along to other BGP
peers.

New transitive optional attributes MAY be further
constrained, depending on attached to the message type. No "padding" of
extra data after path by the message is allowed, so
originator or by any other BGP speaker in the Length field
must have path. If they are not
attached by the smallest value required given originator, the rest of Partial bit in the
message.

Type:

This 1-octet unsigned integer indicates Attribute Flags
octet is set to 1. The rules for attaching new non-transitive
optional attributes will depend on the type code nature of the
message. specific
attribute. 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, documentation of each new non-transitive optional
attribute will be expected to include such rules. (The description of
the first
message sent MULTI_EXIT_DISC attribute gives an example.) All optional
attributes (both transitive and non-transitive) MAY be updated (if
appropriate) by each side is BGP speakers in the path.

The sender of an OPEN message. If UPDATE message SHOULD order path attributes within
the OPEN UPDATE message is
acceptable, a KEEPALIVE in ascending order of attribute type. The receiver
of an UPDATE message confirming the OPEN is sent back.
Once the OPEN is confirmed, UPDATE, KEEPALIVE, and NOTIFICATION
messages may MUST be exchanged.

In addition prepared to handle path attributes
within the fixed-size BGP header, the OPEN UPDATE message contains that are out of order.

The same attribute (attribute with the following fields:

Expiration Date July 2002 [Page 8]

RFC DRAFT January 2002

+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
| Optional Parameters (variable) |
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Version:

This 1-octet unsigned integer indicates same type) can not appear more
than once within the protocol version
number Path Attributes field of a particular UPDATE
message.

The mandatory category refers to an attribute which MUST be present
in both IBGP and EBGP exchanges if NLRI are contained in the UPDATE
message. Attributes classified as optional for the purpose of the
protocol extension mechanism may be purely discretionary, or discre-
tionary, required, or disallowed in certain contexts.

attribute EBGP IBGP
ORIGIN mandatory mandatory
AS_PATH mandatory mandatory
NEXT_HOP mandatory mandatory
MULTI_EXIT_DISC discretionary discretionary
LOCAL_PREF see Section 5.1.5 required
ATOMIC_AGGREGATE see Section 5.1.6 and 9.1.4
AGGREGATOR discretionary discretionary

Expiration Date September 2003 [Page 24]

RFC DRAFT March 2003

5.1 Path Attribute Usage

The current usage of each BGP version number path attribute 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 described in the Hold Timer. Upon
receipt of an OPEN message, following
clauses.

5.1.1 ORIGIN

ORIGIN is a BGP well-known mandatory attribute. The ORIGIN attribute is
generated by the speaker MUST calculate that originates the associated routing
information. Its value of the Hold Timer SHOULD NOT be changed by using the smaller of its configured
Hold Time and any other speaker.

5.1.2 AS_PATH

AS_PATH is a well-known mandatory attribute. This attribute identi-
fies the Hold Time received autonomous systems through which routing information carried
in the OPEN message. this UPDATE message has passed. The
Hold Time MUST components of this list can be either zero
AS_SETs or at least three seconds. An
implementation may reject connections AS_SEQUENCEs.

When a BGP speaker propagates a route which it has learned from
another BGP speaker's UPDATE message, it modifies the route's AS_PATH
attribute based on the basis location of the Hold
Time. The calculated value indicates BGP speaker to which the maximum number of
seconds that may elapse between route
will be sent:

a) When a given BGP speaker advertises the receipt of successive
KEEPALIVE, and/or UPDATE messages by route to an internal
peer, the sender.

BGP Identifier:

This 4-octet unsigned integer indicates advertising speaker SHALL NOT modify the BGP Identifier of AS_PATH
attribute associated with the sender. A route.

b) When a given BGP speaker sets advertises the value of its BGP
Identifier route to an IP address assigned to that BGP speaker. The
value external
peer, then the advertising speaker updates the AS_PATH attribute
as follows:

1) if the first path segment of the BGP Identifier is determined on startup and AS_PATH is of type
AS_SEQUENCE, the
same for every local interface and every BGP peer.

Optional Parameters Length:

This 1-octet unsigned integer indicates system prepends its own AS number as the total length
last element of the
Optional Parameters field sequence (put it in octets. the leftmost position).
If the value act of this field prepending will cause an overflow in the AS_PATH
segment, i.e. more than 255 ASs, it is zero, no Optional Parameters are present.

Optional Parameters:

This field may contain legal to prepend a list new
segment of optional parameters, where
each parameter type AS_SEQUENCE and prepend its own AS number to
this new segment.

2) if the first path segment of the AS_PATH is encoded as of type AS_SET,
the local system prepends a <Parameter Type, Parameter new path segment of type
AS_SEQUENCE to the AS_PATH, including its own AS number in that

Expiration Date July 2002 September 2003 [Page 9] 25]

RFC DRAFT January 2002

Length, Parameter Value> triplet.

0 1
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...
| Parm. Type | Parm. Length | Parameter Value (variable)
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-...

Parameter Type is March 2003

segment.

When a one octet field that unambiguously
identifies individual parameters. Parameter Length is BGP speaker originates a one
octet field that contains route then:

a) the originating speaker includes its own AS number in a path
segment of type AS_SEQUENCE in the length AS_PATH attribute of all UPDATE
messages sent to an external peer. (In this case, the Parameter Value
field AS number of
the originating speaker's autonomous system will be the only entry
the path segment, and this path segment will be the only segment
in octets. Parameter Value the AS_PATH attribute).

b) the originating speaker includes an empty AS_PATH attribute in
all UPDATE messages sent to internal peers. (An empty AS_PATH
attribute is a variable one whose length field
that is interpreted according to contains the value zero).

Whenever the modification of the Parameter
Type field. AS_PATH attribute calls for includ-
ing or prepending the AS number of the local system, the local system
MAY include/prepend more than one instance of its own AS number in
the AS_PATH attribute. This document is controlled via local configuration.

5.1.3 NEXT_HOP

The NEXT_HOP is a well-known mandatory attribute that defines the following Optional Parameters:

a) Authentication Information (Parameter Type 1):

This optional parameter may IP
address of the router that SHOULD be used as the next hop to authenticate a BGP
peer. the des-
tinations listed in the UPDATE message. The Parameter Value field contains a 1-octet
Authentication Code followed by NEXT_HOP attribute is
calculated as follows.

1) When sending a variable length
Authentication Data.

0 1 2 3 4 5 6 7 8
+-+-+-+-+-+-+-+-+
| Auth. Code |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
| Authentication Data |
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Authentication Code:

This 1-octet unsigned integer indicates the
authentication mechanism being used. Whenever message to an
authentication mechanism internal peer, if the route is specified for use within
BGP, three things must be included in not
locally originated the
specification:

- BGP speaker SHOULD NOT modify the NEXT_HOP
attribute, unless it has been explicitly configured to announce
its own IP address as the value of NEXT_HOP. When announcing a locally
originated route to an internal peer, the Authentication Code which indicates BGP speaker SHOULD use of
as the mechanism,
- NEXT_HOP the form and meaning interface address of the Authentication Data, and
- router through which
the algorithm announced network is reachable for computing values of Marker fields.

Expiration Date July 2002 [Page 10]

RFC DRAFT January 2002

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

Authentication Data:

Authentication Data is a variable length field that speaker; if the route
is
interpreted according directly connected to the value of speaker, or the
Authentication Code field.

The minimum length interface address of
the OPEN message router through which the announced network 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 reachable for
the UPDATE packet can be used to construct
a graph describing speaker is the relationships internal peer's address, then the BGP speaker
SHOULD use for the NEXT_HOP attribute its own IP address (the
address 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.

An UPDATE message interface that is used to advertise feasible routes sharing common
path attribute to reach the peer).

2) When sending a peer, or message to withdraw multiple unfeasible routes an external peer X, and the peer is
one IP hop away from service (see 3.1). An UPDATE message may simultaneously
advertise a feasible the speaker:

- If the route and withdraw multiple unfeasible routes being announced was learned from service. The UPDATE message always includes an internal
peer or is locally originated, the fixed-size BGP
header, and also includes speaker can use for the other fields as shown below (note, some
NEXT_HOP attribute an interface address of the shown fields may not be present in every UPDATE message):

Withdrawn Routes Length: internal peer

Expiration Date July 2002 September 2003 [Page 11] 26]

RFC DRAFT January 2002

This 2-octets unsigned integer indicates March 2003

router (or the total length of internal router) through which the Withdrawn Routes field in octets. Its value must allow announced
network is reachable for the
length speaker, provided that peer X
shares a common subnet with this address. This is a form of
"third party" NEXT_HOP attribute.

- Otherwise, if the Network Layer Reachability Information field to
be determined as specified below.

A value of 0 indicates that no routes are route being withdrawn announced was learned from
service, and that an
external peer, the WITHDRAWN ROUTES field is not present speaker can use in the NEXT_HOP attribute an
IP address of any adjacent router (known from the received
NEXT_HOP attribute) that the speaker itself uses for local
route calculation, provided that peer X shares a common subnet
with this UPDATE message.

Withdrawn Routes: address. This is a variable length field that contains a list second form of IP
address prefixes for "third party"
NEXT_HOP attribute.

- Otherwise, if the routes that are external peer to which the route is being withdrawn from
service. Each
advertised shares a common subnet with one of the interfaces of
the announcing BGP speaker, the speaker MAY use the IP address prefix
associated with such an interface in the NEXT_HOP attribute.
This is encoded known as a 2-tuple "first party" NEXT_HOP attribute.

- By default (if none 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 above conditions apply), the length BGP
speaker SHOULD use in bits of the NEXT_HOP attribute the IP address prefix. A length of zero indicates a prefix
the interface that
matches all IP addresses (with prefix, itself, of zero
octets).

b) Prefix:

The Prefix field contains an IP address prefix followed by
enough trailing bits to make the end of speaker uses to establish the field fall on BGP con-
nection to peer X.

3) When sending a message to an
octet boundary. Note that external peer X, and the value of trailing bits peer 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
multiple IP hops away from the Network Layer Reachability field to speaker (aka "multihop EBGP"):

- The speaker MAY be determined
as specified below.

A value of 0 indicates that no Network Layer Reachability

Expiration Date July 2002 [Page 12]

RFC DRAFT January 2002

Information field is present in configured to propagate the NEXT_HOP
attribute. 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 case when advertising a two-octet field route that consists of the
Attribute Flags octet followed by
speaker learned from one of its peers, 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) NEXT_HOP attribute
of the Attribute Flags octet advertised route is exactly the
Optional bit. It defines whether same as the NEXT_HOP
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 learned route (the speaker just doesn't modify
the Transitive bit. It defines whether an optional NEXT_HOP attribute).

- By default, the BGP speaker SHOULD use in the NEXT_HOP
attribute
is transitive (if set the IP address of the interface that the speaker uses
to 1) or non-transitive (if set establish the BGP connection to 0).
For well-known attributes, peer X.

Normally the Transitive bit must NEXT_HOP attribute is chosen such that the shortest
available path will be set taken. A BGP speaker MUST be able to 1.
(See Section 5 for a discussion support
disabling advertisement of transitive attributes.)

The third high-order bit (bit 2) party NEXT_HOP attributes to handle
imperfectly bridged media.

A route originated by a BGP speaker SHALL NOT be advertised to a peer
using an address of that peer as NEXT_HOP. A BGP speaker SHALL NOT
install a route with itself as the next hop.

The NEXT_HOP attribute is used by the BGP speaker to determine the

Expiration Date September 2003 [Page 27]

RFC DRAFT March 2003

actual outbound interface and immediate next-hop address that SHOULD
be used to forward transit packets to the Attribute Flags octet associated destinations.

The immediate next-hop address is determined by performing a recur-
sive route lookup operation for the Partial bit. It defines whether the information
contained IP address in the optional transitive NEXT_HOP
attribute is partial (if
set to 1) or complete (if set to 0). For well-known attributes
and for optional non-transitive attributes using the Partial bit must
be set to 0.

The fourth high-order bit (bit 3) contents of the Attribute Flags octet
is the Extended Length bit. It defines whether the Attribute
Length is Routing Table, selecting one octet (if set to 0) or two octets (if set to 1).

The lower-order four bits
entry if multiple entries of the Attribute Flags octet are
unused. They must be zero when sent and must be ignored when
received. equal cost exist. The Attribute Type Code octet contains Routing Table
entry which resolves the Attribute Type Code.

Expiration Date July 2002 [Page 13]

RFC DRAFT January 2002

Currently defined Attribute Type Codes are discussed IP address in Section
5. the NEXT_HOP attribute will
always specify the outbound interface. If the Extended Length bit of entry specifies an
attached subnet, but does not specify a next-hop address, then the Attribute Flags octet is set
to 0,
address in the third octet of NEXT_HOP attribute SHOULD be used as the Path Attribute contains immediate
next-hop address. If the length
of entry also specifies the next-hop address,
this address SHOULD be used as the immediate next-hop address for
packet forwarding.

5.1.4 MULTI_EXIT_DISC

The MULTI_EXIT_DISC is an optional non-transitive attribute data in octets.

If which is
intended to be used on external (inter-AS) links to discriminate
among multiple exit or entry points to the Extended Length bit same neighboring AS. The
value of the Attribute Flags MULTI_EXIT_DISC attribute is a four octet unsigned num-
ber which is set
to 1, then the third and called a metric. All other factors being equal, the fourth octets of exit
point with lower metric SHOULD be preferred. If received over EBGP,
the path MULTI_EXIT_DISC attribute contain the length of MAY be propagated over IBGP to other
BGP speakers within the attribute data in octets. same AS. The remaining octets of the Path Attribute represent the MULTI_EXIT_DISC attribute value and are interpreted according
received from a neighboring AS MUST NOT be propagated to other neigh-
boring ASs.

A BGP speaker MUST IMPLEMENT a mechanism based on local configuration
which allows the Attribute
Flags and the Attribute Type Code. The supported Attribute Type
Codes, their MULTI_EXIT_DISC attribute values and uses are the following:

a) ORIGIN (Type Code 1):

ORIGIN is to be removed from a well-known mandatory attribute that defines
route. This MAY be done prior to determining the
origin degree of preference
of the path information. The data octet can assume
the following values:

Value Meaning

0 IGP - Network Layer Reachability Information
is interior to the originating AS route and performing route selection (decision process phases
1 EGP - Network Layer Reachability Information
learned via and 2).

An implementation MAY also (based on local configuration) alter the EGP protocol

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
value of the MULTI_EXIT_DISC attribute that is composed received over EBGP. This MAY
be done prior to determining the degree of a sequence preference 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 route
and performing route selection (decision process phases 1 AS_SET: unordered set of ASs and 2). See
Section 9.1.2.2 for necessary restrictions on this.

5.1.5 LOCAL_PREF

LOCAL_PREF is a route well-known attribute that SHALL be included in all
UPDATE messages that a given BGP speaker sends to the other internal

Expiration Date July 2002 September 2003 [Page 14] 28]

RFC DRAFT January 2002

UPDATE message has traversed

2 AS_SEQUENCE: ordered set March 2003

peers. A BGP speaker SHALL calculate the degree of ASs a preference for
each external route in based on the UPDATE message has traversed

The path segment length is a 1-octet long field containing locally configured policy, and
include the number degree of ASs in the path segment value field.

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

Usage route to its
internal peers. The higher degree of preference MUST be preferred. A
BGP speaker uses the degree of preference learned via LOCAL_PREF in
its decision process (see Section 9.1.1).

A BGP speaker MUST NOT include this attribute in UPDATE messages that
it sends to external peers, except for the case of BGP Confederations
[RFC3065]. If it is defined contained in 5.1.2.

c) NEXT_HOP (Type Code 3):

This an UPDATE message that is received
from an external peer, then this attribute MUST be ignored by the
receiving speaker, except for the case of BGP Confederations
[RF3065].

5.1.6 ATOMIC_AGGREGATE

ATOMIC_AGGREGATE is a well-known mandatory attribute that defines discretionary attribute.

When a BGP speaker aggregates several routes for the IP
address purpose of
advertisement to a particular peer, the AS_PATH of the border router aggregated
route normally includes an AS_SET formed from the set of ASs from
which the aggregate was formed. In many cases the network adminis-
trator can determine that should the aggregate can safely be used as advertised
without the next
hop to AS_SET and not form route loops.

If an aggregate excludes at least some of the destinations listed AS numbers present in
the Network Layer
Reachability Information field AS_PATH of the UPDATE message.

Usage of this attribute is defined in 5.1.3.

d) MULTI_EXIT_DISC (Type Code 4):

This is an optional non-transitive attribute routes that is are aggregated as a four
octet non-negative integer. The value result of this attribute may
be used by a BGP speaker's decision process to discriminate
among multiple entry points dropping
the AS_SET, the aggregated route, when advertised 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 attribute that is a four octet
non-negative integer. the peer, SHOULD
include the ATOMIC_AGGREGATE attribute.

A BGP speaker uses that receives a route with the ATOMIC_AGGREGATE
attribute SHOULD NOT remove the attribute from the route when propa-
gating it to inform other
internal peers of speakers.

A BGP speaker that receives a route with the advertising speaker's degree of
preference for an advertised route. Usage of this ATOMIC_AGGREGATE
attribute
is described MUST NOT make any NLRI of that route more specific (as
defined in 5.1.5.

f) ATOMIC_AGGREGATE (Type Code 6)

ATOMIC_AGGREGATE is 9.1.4) when advertising this route to other BGP speakers.

A BGP speaker that receives a well-known discretionary route with the ATOMIC_AGGREGATE
attribute needs to be cognizant of
length 0. Usage the fact that the actual path to
destinations, as specified in the NLRI of this attribute is described the route, while having the
loop-free property, may not be the path specified in 5.1.6.

g) AGGREGATOR (Type Code 7) the AS_PATH
attribute of the route.

Expiration Date July 2002 September 2003 [Page 15] 29]

RFC DRAFT January 2002 March 2003

5.1.7 AGGREGATOR

AGGREGATOR is an optional transitive attribute of length 6.
The attribute contains the last AS number that which MAY be included
in updates which are formed the
aggregate route (encoded as 2 octets), followed by the IP
address of the aggregation (see Section 9.2.2.2). A
BGP speaker that formed the aggregate which performs route
(encoded as 4 octets). This should aggregation MAY add the AGGREGATOR
attribute which SHALL contain its own AS number and IP address. The
IP address SHOULD be the same address as the one used for the BGP Identifier of the speaker. Usage

6. BGP Error Handling.

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

When any of this attribute is the conditions described in 5.1.7.

Network Layer Reachability Information:

This variable length field contains here are detected, a list of IP address
prefixes. The length in octets of NOTIFICA-
TION message with the Network Layer
Reachability Information indicated Error Code, Error Subcode, and Data
fields is sent, and the BGP connection is closed, unless it is
explicitly stated that no NOTIFICATION message is to be sent and the
BGP connection is not encoded explicitly, but can to be
calculated as:

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

where UPDATE message Length closed. If no Error Subcode is specified,
then a zero MUST be used.

The phrase "the BGP connection is closed" means that the value encoded in TCP connec-
tion has been closed, the fixed-
size BGP header, Total Path Attribute Length associated Adj-RIB-In has been cleared, and Withdrawn
Routes Length are the values encoded
that all resources for that BGP connection have been deallocated.
Entries in the variable part of Loc-RIB associated with the UPDATE message, and 23 is a combined length of remote peer are marked as
invalid. The fact that the fixed-
size routes have become invalid is passed to
other BGP header, peers before the Total Path Attribute Length routes are deleted from the system.

Unless specified explicitly, the Data field and of the
Withdrawn Routes Length field.

Reachability information NOTIFICATION mes-
sage that is encoded as one or more 2-tuples of sent to indicate an error is empty.

6.1 Message Header error handling.

All errors detected while processing the form <length, prefix>, whose fields Message Header are described below:

+---------------------------+
| Length (1 octet) |
+---------------------------+
| Prefix (variable) |
+---------------------------+ indicated
by sending the NOTIFICATION message with Error Code Message Header
Error. The use and Error Subcode elaborates on the meaning specific nature of these fields are as follows:

a) Length:

The Length field indicates the length in bits
error.

The expected value of the IP
address prefix. A length Marker field of zero indicates a prefix that
matches the message header is all IP addresses (with prefix, itself,
ones. If the Marker field of zero
octets).

b) Prefix:

The Prefix the message header is not as expected,
then a synchronization error has occurred and the Error Subcode is
set to Connection Not Synchronized.

If the Length field contains IP address prefixes followed by of the message header is less than 19 or greater

Expiration Date July 2002 September 2003 [Page 16] 30]

RFC DRAFT January 2002

enough trailing bits to make the end of March 2003

than 4096, or if the Length field fall on an
octet boundary. Note that the value of the trailing bits an OPEN message is
irrelevant.

The less than the
minimum length of the OPEN message, or if the Length field of an
UPDATE message is 23 octets -- 19 octets
for less than the fixed header + 2 octets for minimum length of the Withdrawn Routes Length + 2
octets for UPDATE message,
or if the Total Path Attribute Length (the value field of Withdrawn
Routes Length a KEEPALIVE message is 0 and not equal to 19, or
if the value of Total Path Attribute Length is
0).

An UPDATE message can advertise at most one set field of path attributes,
but multiple destinations, provided that the destinations share these
attributes. All path attributes contained in a given UPDATE NOTIFICATION message
apply is less than the mini-
mum length of the NOTIFICATION message, then the Error Subcode is set
to all destinations carried in Bad Message Length. The Data field contains the NLRI erroneous Length
field.

If the Type field of the UPDATE
message.

An UPDATE message can list multiple routes to be withdrawn from
service. Each such route header is identified 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 its destination (expressed
as an IP prefix), which unambiguously identifies sending the NOTIFICATION message with Error Code OPEN Message
Error. The Error Subcode elaborates on the route specific nature of the
error.

If the version number contained in the
context Version field of the BGP speaker - BGP speaker connection to which it has
been previously advertised.

An UPDATE received
OPEN message might advertise only routes to be withdrawn from
service, in which case it will is not include path attributes or Network
Layer Reachability Information. Conversely, it may advertise only supported, then the Error Subcode is set to
Unsupported Version Number. The Data field is a
feasible route, in 2-octets unsigned
integer, which case indicates the WITHDRAWN ROUTES field need not be
present.

An UPDATE message should not include largest locally supported version number
less than the same address prefix in version the
WITHDRAWN ROUTES and Network Layer Reachability Information fields,
however a remote BGP speaker MUST be able to process UPDATE messages peer bid (as indicated in this
form. A BGP speaker should treat an UPDATE message of this form as the
received OPEN message), or if the WITHDRAWN ROUTES doesn't contain smallest locally supported version
number is greater than the address prefix.

4.4 KEEPALIVE Message Format version the remote 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 peer bid, then the Hold Timer
smallest locally supported version number.

If the Autonomous System field of the OPEN message is unacceptable,
then the Error Subcode is set to expire. A reasonable maximum time between KEEPALIVE messages would
be one third Bad Peer AS. The determination of
acceptable Autonomous System numbers is outside the scope of this
protocol.

If the Hold Time interval. KEEPALIVE messages field of the OPEN message is unacceptable, then the
Error Subcode MUST NOT be sent more frequently than set to Unacceptable Hold Time. An implementa-
tion MUST reject Hold Time values of one per second. or two seconds. An implementation imple-
mentation MAY
adjust the rate at reject any proposed Hold Time. An implementation which it sends KEEPALIVE messages as
accepts a function of
the Hold Time interval.

If MUST use the negotiated value for the Hold Time interval is zero, then periodic KEEPALIVE

Expiration Date July 2002 [Page 17]

RFC DRAFT January 2002

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 Time.

If the fixed-size BGP header, the NOTIFICATION message
contains the following fields:

Error Code:

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

Error Code Symbolic Name Reference

1 Message Header Error Section 6.1

2 the 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 message is syntactically
incorrect, then the nature Error Subcode is set to Bad BGP Identifier. Syn-
tactic correctness means that the BGP Identifier field represents a
valid IP host address.

If one of the reported error. Each Error Optional Parameters in the OPEN message is not

Expiration Date July 2002 September 2003 [Page 18] 31]

RFC DRAFT January 2002

Code may have one or more Error Subcodes associated with it. If
no appropriate March 2003

recognized, then the Error Subcode is defined, then a zero
(Unspecific) value set to Unsupported Optional
Parameters.

If one of the Optional Parameters in the OPEN message is used for recognized,
but is malformed, then the Error Subcode field.

Message Header is set to 0 (Unspecific).

6.3 UPDATE message error handling.

All errors detected while processing the UPDATE message are indicated
by sending the NOTIFICATION message with Error subcodes:

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

OPEN Code UPDATE Message
Error. The error subcode elaborates on the specific nature of the
error.

Error subcodes:

1 - Unsupported Version Number.
2 - Bad Peer AS.
3 - Bad BGP Identifier.
4 - Unsupported Optional Parameter.
5 - Authentication Failure.
6 - Unacceptable Hold Time. checking of an UPDATE Message message begins by examining the path
attributes. If the Withdrawn Routes Length or Total Attribute Length
is too large (i.e., if Withdrawn Routes Length + Total Attribute
Length + 23 exceeds the message Length), then the Error subcodes:

1 - Subcode is
set to Malformed Attribute List.
2 - Unrecognized Well-known Attribute.
3 - Missing Well-known Attribute.
4 -

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

If any recognized attribute has Attribute Length Error.
6 - Invalid ORIGIN Attribute
8 - Invalid NEXT_HOP Attribute.
9 - Optional that conflicts with
the expected length (based on the attribute type code), then the
Error Subcode is set to Attribute Length Error.
10 - Invalid Network Field.
11 - Malformed AS_PATH.

Data:

This variable-length The Data field con-
tains the erroneous attribute (type, length and value).

If any of the mandatory well-known attributes are not present, then
the Error Subcode is used set to diagnose Missing Well-known Attribute. The Data
field contains the reason for Attribute Type Code of the NOTIFICATION. The contents 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 depend upon contains the Error Code unrecognized attribute (type, length and Error Subcode. See Section 6 below for more
details.

Note that
value).

If the length of ORIGIN attribute has an undefined value, then the Error Sub-
code is set to Invalid Origin Attribute. The Data field can be determined from contains the message Length
unrecognized attribute (type, length and value).

If the NEXT_HOP attribute field by is syntactically incorrect, then the formula:

Message Length = 21 + Data Length
Error Subcode is set to Invalid NEXT_HOP Attribute. The minimum Data field
contains the incorrect attribute (type, length and value). Syntactic

Expiration Date September 2003 [Page 32]

RFC DRAFT March 2003

correctness means that the NEXT_HOP attribute represents a valid IP
host address.

The IP address in the NEXT_HOP MUST meet the following criteria to be
considered semantically correct:

a) It MUST NOT be the IP address of the NOTIFICATION message is 21 octets
(including message header).

Expiration Date July 2002 [Page 19]

RFC DRAFT January 2002

5. Path Attributes

This section discusses receiving speaker

b) In the path attributes case of an EBGP where 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 sender and must be included in every
UPDATE message that contains NLRI. Others receiver are discretionary and may
or may not be sent one IP
hop away from each other, either the IP address in a particular UPDATE message.

All well-known attributes must the NEXT_HOP
MUST be passed along (after proper
updating, if necessary) the sender's IP address (that is used to other establish the BGP peers.

In addition to well-known attributes, each path may contain one or
more optional attributes. It is not required
connection), or expected that all the interface associated with the NEXT_HOP IP
address MUST share a common subnet with the receiving BGP
implementations support all optional attributes. The handling of an
unrecognized optional speaker.

If the NEXT_HOP attribute is determined by the setting of semantically incorrect, the
Transitive bit in error SHOULD
be logged, and the attribute flags octet. Paths with unrecognized
transitive optional attributes should route SHOULD be accepted. If ignored. In this case, a NOTIFICA-
TION message SHOULD NOT be sent, and connection SHOULD NOT be closed.

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

If the UPDATE message is received from an external peer, the local
system MAY check whether the leftmost AS in the AS_PATH attribute is accepted and passed
along
equal to other BGP peers, then the unrecognized transitive optional
attribute autonomous system number of the peer that path must be passed along with sent the path to other
BGP peers with mes-
sage. If the Partial bit in check determines that this is not the Attribute Flags octet case, the Error
Subcode is set to 1. Malformed AS_PATH.

If a path with recognized transitive an optional attribute is accepted
and passed along to other BGP peers and recognized, then the Partial bit in value of this
attribute is checked. If an error is detected, the
Attribute Flags octet attribute is set to 1 by some previous AS, it dis-
carded, and the Error Subcode is not set
back to 0 by Optional Attribute Error.
The Data field contains the current AS. Unrecognized non-transitive optional
attributes must be quietly ignored attribute (type, length and not passed along to other BGP
peers.

New transitive optional attributes may be attached to the path by the
originator or by value).

If any other BGP speaker attribute appears more than once in the path. If they are not
attached by the originator, the Partial bit in UPDATE message, then
the Attribute Flags
octet Error Subcode is set to 1. Malformed Attribute List.

The rules NLRI field in the UPDATE message is checked for attaching new non-transitive
optional attributes will depend on syntactic valid-
ity. If the nature of field is syntactically incorrect, then the specific
attribute. The documentation of each new non-transitive optional
attribute will be expected Error Subcode
is set to include such rules. (The description of
the MULTI_EXIT_DISC attribute gives Invalid Network Field.

If a prefix in the NLRI field is semantically incorrect (e.g., an example.) All optional
attributes (both transitive and non-transitive) may
unexpected multicast IP address), an error SHOULD be updated (if
appropriate) by BGP speakers in logged locally,
and the path. prefix SHOULD be ignored.

An UPDATE message that contains correct path attributes, but no NLRI,
SHALL be treated as a valid UPDATE message.

Expiration Date July 2002 September 2003 [Page 20] 33]

RFC DRAFT January 2002

The sender of an UPDATE March 2003

6.4 NOTIFICATION message should order path attributes within error handling.

If a peer sends a NOTIFICATION message, and 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 mes-
sage detects an error in that are out of order.

The same attribute cannot appear more than once within message, the Path
Attributes field of receiver can not use a particular UPDATE message.

The mandatory category refers
NOTIFICATION message to report this error back to an attribute which must be present
in both IBGP and EBGP exchanges if NLRI are contained in the UPDATE
message. Attributes classified peer. Any such
error, such as optional for the purpose of the
protocol extension mechanism may be purely discretionary, or
discretionary, required, an unrecognized Error Code or disallowed in certain contexts.

attribute EBGP IBGP
ORIGIN mandatory mandatory
AS_PATH mandatory mandatory
NEXT_HOP mandatory mandatory
MULTI_EXIT_DISC discretionary discretionary
LOCAL_PREF disallowed required
ATOMIC_AGGREGATE see section 5.1.6 Error Subcode, SHOULD be
noticed, logged locally, and 9.1.4
AGGREGATOR discretionary discretionary

5.1 Path Attribute Usage

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

5.1.1 ORIGIN

ORIGIN is a well-known mandatory attribute. adminis-
tration of the peer. The ORIGIN attribute
shall be generated by means to do this, however, lies outside the autonomous
scope of this document.

6.5 Hold Timer Expired error handling.

If a system that originates does not receive successive KEEPALIVE and/or UPDATE
and/or NOTIFICATION messages within the
associated routing information. It shall be included period specified in the UPDATE
messages Hold
Time field of all the OPEN message, then the NOTIFICATION message with
Hold Timer Expired Error Code is sent and the BGP speakers that choose to propagate this
information to other connection is
closed.

6.6 Finite State Machine error handling.

Any error detected by the BGP speakers.

5.1.2 AS_PATH

AS_PATH Finite State Machine (e.g., receipt of
an unexpected event) is a well-known mandatory attribute. This attribute

Expiration Date July 2002 [Page 21]

RFC DRAFT January 2002

identifies indicated by sending the autonomous systems through which routing information
carried in this UPDATE NOTIFICATION message has passed. The components
with Error Code Finite State Machine Error.

6.7 Cease.

In absence of any fatal errors (that are indicated in this
list can 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 AS_SETs or AS_SEQUENCEs.

When used when a fatal error
indicated by this section does exist.

A BGP speaker propagates MAY support the ability to impose an (locally config-
ured) upper bound on the number of address prefixes the speaker is
willing to accept from a route which it has learned neighbor. When the upper bound is reached,
the speaker (under control of local configuration) either (a) dis-
cards new address prefixes from
another the neighbor (while maintaining BGP speaker's UPDATE message, it shall modify
connection with the route's
AS_PATH attribute based on neighbor), or (b) terminates the location of BGP connection
with the neighbor. If the BGP speaker decides to which
the route will be sent:

a) When a given terminate its BGP speaker advertises
connection with a neighbor because the route to an internal
peer, number of address prefixes
received from the neighbor exceeds the locally configured upper

Expiration Date September 2003 [Page 34]

RFC DRAFT March 2003

bound, then the advertising speaker shall not modify MUST send to the AS_PATH
attribute associated neighbor a NOTIFICATION mes-
sage with the route.

b) When Error Code Cease.

6.8 BGP connection collision detection.

If a given pair of BGP speaker advertises the route speakers try simultaneously to an external
peer, establish a BGP con-
nection to each other, then two parallel connections between this
pair of speakers might well be formed. If the advertising speaker shall update source IP address used
by one of these connections is the AS_PATH
attribute same as follows:

1) if the first path segment of destination IP address
used by the AS_PATH other, and the destination IP address used by the first
connection is of type
AS_SEQUENCE, the local system shall prepend its own AS number same as the last element of source IP address used by the sequence (put it other, we
refer to this situation as connection collision. Clearly in the leftmost
position). If
presence of connection collision, one of these connections MUST be
closed.

Based on the act value of prepending will cause an overflow in the AS_PATH segment, i.e. more than 255 elements, it shall be
legal BGP Identifier a convention is established
for detecting which BGP connection is to prepend be preserved when a new segment colli-
sion does occur. The convention is to compare the BGP Identifiers of type AS_SEQUENCE
the peers involved in the collision and prepend
its own AS number to this new segment.

2) if retain only the first path segment of connection
initiated by the AS_PATH is BGP speaker with the higher-valued BGP Identifier.

Upon receipt of type AS_SET, an OPEN message, the local system shall prepend a new path segment MUST examine all of type
AS_SEQUENCE to the AS_PATH, including
its own AS number in connections that
segment.

When a BGP speaker originates a route then:

a) are in the originating OpenConfirm state. A BGP speaker shall include its own AS number in a
path segment of type AS_SEQUENCE MAY
also examine connections in an OpenSent state if it knows the AS_PATH attribute BGP
Identifier of all
UPDATE messages sent to an external peer. (In this case, the AS
number peer by means outside of the originating speaker's autonomous system will be protocol. If among
these connections there is a connection to a remote BGP speaker whose
BGP Identifier equals the
only entry one in the path segment, OPEN message, and this path segment will be the
only segment in connec-
tion collides with the AS_PATH attribute).

b) connection over which the originating speaker shall include an empty AS_PATH
attribute in all UPDATE messages sent to internal peers. (An
empty AS_PATH attribute OPEN message is one whose length field contains
received then the
value zero).

Whenever local system performs the modification following collision reso-
lution procedure:

1. The BGP Identifier of the AS_PATH attribute calls for
including or prepending local system is compared to the BGP
Identifier of the remote system (as specified in the OPEN mes-
sage). Comparing BGP Identifiers is done by converting them to
host byte order and treating them as (4-octet long) unsigned inte-
gers.

2. If the AS number value of the local system, BGP Identifier is less than the
remote one, the local system may include/prepend more than closes the BGP connection that
already exists (the one instance of its own AS
number that is already in the AS_PATH attribute. This is controlled via OpenConfirm state),
and accepts the 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).

Expiration Date July 2002 September 2003 [Page 22] 35]

RFC DRAFT January 2002

configuration.

5.1.3 NEXT_HOP

The NEXT_HOP path attribute defines the IP address March 2003

Unless allowed via configuration, a connection collision with an
existing BGP connection that is in Established state causes closing
of the border
router newly created connection.

Note that should a connection collision can not be used as the next hop to the destinations listed detected with connections
that are in Idle, or Connect, or Active states.

Closing the UPDATE message. The NEXT_HOP attribute BGP connection (that results from the collision resolu-
tion procedure) is calculated as
follows.

1) When accomplished by sending a the NOTIFICATION message to an internal peer,
with the Error Code Cease.

7. BGP speaker
should not modify Version Negotiation

BGP speakers MAY negotiate the NEXT_HOP attribute, unless it has been
explicitly configured to announce its own IP address as version of the
NEXT_HOP.

2) When sending a message protocol by making mul-
tiple attempts to an external peer X, and the peer is
one IP hop away from open a BGP connection, starting with the speaker:

- highest
version number each supports. If the route being announced was learned from an internal
peer or is locally originated, open attempt fails with an Error
Code OPEN Message Error, and an Error Subcode Unsupported Version
Number, then the BGP speaker can use for has available the
NEXT_HOP attribute an interface address of version number it
tried, the internal version number its peer
router (or the internal router) through which the announced
network is reachable for tried, the speaker, provided that version number passed
by its peer X
shares a common subnet with this address. This is a form of
"third party" NEXT_HOP attribute.

- If the route being announced was learned from an external
peer, the speaker can use in the NEXT_HOP attribute an IP
address of any adjacent router (known from NOTIFICATION message, and the received
NEXT_HOP attribute) version numbers that
it supports. If the speaker itself uses for local
route calculation, provided that peer X shares a two peers do support one or more common subnet
with versions,
then this address. This is a second form of "third party"
NEXT_HOP attribute.

- If the external peer will allow them to which rapidly determine the route is being advertised
shares a highest common subnet with one
version. In order to support BGP version negotiation, future versions
of BGP MUST retain the announcing router's own
interfaces, the router may use the IP address associated with
such an interface in the NEXT_HOP attribute. This is known as a
"first party" NEXT_HOP attribute.

- By default (if none format of the above conditions apply), OPEN and NOTIFICATION messages.

8. BGP Finite State machine

This section specifies the BGP
speaker should use operation in the NEXT_HOP attribute the IP address terms of a Finite State
Machine (FSM). The section falls into 2 parts:

1) Description of Events for the interface that the speaker uses to establish State machine (Section 8.1)
2) Description of the BGP
session to peer X. FSM (Section 8.2)

Session Attributes required for each connection are;

1) State
2) Connect Retry timer
3) When sending a message to an external peer X, and the peer is Hold timer
4) Hold time
5) Keepalive timer
6) Keepalive time
7) Connect Retry Count
8) Connect Retry Initial Value

Expiration Date July 2002 September 2003 [Page 23] 36]

RFC DRAFT January 2002

multiple IP hops away from the speaker (aka "multihop EBGP"):

- March 2003

The speaker optional Session attributes are listed below. These optional
attributes may be configured to propagate supported either per connection or per local sys-
tem:

1) Delay Open flag
2) Open Delay Timer
3) Perform automatic start flag
4) Perform automatic stop flag
5) Passive TCP establishment flag
6) Perform BGP peer oscillation damping flag
(which will be denoted as stop_peer_flap in text)
7) Idle Hold timer
8) Perform Collision detect in Established flag
9) Accept connections from un-configured peers
10) Track TCP state flag
11) Send NOTIFICATION without an OPEN flag

8.1 Events for the NEXT_HOP
attribute. In this case when advertising a route BGP FSM

8.1.1 Administrative Events

Please note that only Event 1 (manual start) and Event 2 (manual
stop) are mandatory administrative events. All other administrative
events are optional. The optional attributes do not have to be sup-
ported. However, if these attributes are supported, the
speaker learned from one of its peers, the NEXT_HOP attribute state of the advertised route is exactly the same
flags should be as indicated.

Event1: Manual start

Definition: Local system administrator manually starts peer
connection.

Status: Mandatory

Optional
attributes: Passive TCP establishment flag SHOULD not be set.

Event2: Manual stop

Definition: Local system administrator manually
stops the NEXT_HOP
attribute of peer connection.

Status: Mandatory

Expiration Date September 2003 [Page 37]

RFC DRAFT March 2003

Event3: Automatic start

Definition: Local system automatically starts the learned route (the speaker just doesn't modify
BGP connection.

Status: Optional depending on local system.

Optional
attributes: 1) Perform automatic start flag SHOULD be set.
if this event occurs.
2) if the NEXT_HOP attribute).

- By default, passive Passive TCP establishment flag
is supported, it SHOULD not be set if this
event occurs.
3) if bgp peer oscillation damping is supported,
the BGP speaker stop_peer_flap flag should use in the NEXT_HOP
attribute not be set
when this event occurs.

Event4: Manual start with passive TCP flag

Definition: Local system administrator manually starts the IP address of peer
connection, but has the interface passive TCP establishment
enabled. The passive TCP establishment flag indicates
that the speaker uses peer will listen prior to establish
establishing the BGP session to connection.

Status: Optional depending on local system.

Optional
attributes: 1) Passive TCP Establishment flag SHOULD be set.
if this event occurs.
2) If bgp peer X.

Normally the NEXT_HOP attribute oscilation damping is chosen such that supported, the shortest
available path will be taken. A BGP speaker must
stop_peer_flap falg should not be able to support
disabling advertisement of third party NEXT_HOP attributes to handle
imperfectly bridged media.

A BGP speaker must never advertise an address of a peer to that peer
as a NEXT_HOP, for a route that set when
this event occurs.

Event5: Automatic start with passive TCP flag

Definition: Local system automatically starts the speaker is originating. A
BGP
speaker must never install a route connection with itself as the next hop. passive flag
enabled. The NEXT_HOP attribute is used by the BGP speaker to determine the
actual outbound interface and immediate next-hop address passive flag indicates
that should
be used to forward transit packets to the associated destinations.
The immediate next-hop address is determined by performing peer will listen prior to
establishing a
recursive route lookup operation for the IP address in the NEXT_HOP
attribute using the contents connection.

Status: Optional depending on local system use
of the Routing Table (see Section
9.1.2.2). The resolving route will always specify the outbound
interface. a passive connection and automatic start.

Expiration Date September 2003 [Page 38]

RFC DRAFT March 2003

Optional
attributes: 1) Perform Automatic start flag SHOULD be set
2) Passive TCP establishment flag SHOULD be set
3) If the resolving route specifies bgp peer oscillation flag is supported,
the next-hop address,
this address should stop_peer_flap flag SHOULD not be used as the immediate address for packet
forwarding. If the address in set.

Event6: Automatic start with bgp_stop_flap option set

Definition: Local system automatically starts the NEXT_HOP attribute
BGP peer connection with peer oscillation
damping enabled. The exact method of damping
persistent peer oscillations is directly
resolved through a route left up to an attached subnet (such a route will not
specify the next-hop address), the outbound interface should be taken
from the resolving route
implementation, and is outside the address in the NEXT_HOP attribute
should be scope of
this document.

Status: Optional, used as only if the immediate next-hop address.

5.1.4 MULTI_EXIT_DISC

The MULTI_EXIT_DISC bgp peer has enabled
bgp peer oscillation damping enabled with the
optional attribute may settings below.

Optional
attributes: 1) Perform automatic start flag SHOULD be used on external (inter-AS)
links to discriminate among multiple exit or entry points to set
2) stop_peer_flap flag SHOULD be set
3) Passive TCP establishment flag SHOULD not be set
(cleared).

Event 7: Automatic start with bgp_stop_flap option set and passive
TCP establishment option set

Definition: Local system automatically starts the same
neighboring AS.
BGP peer connection with peer oscillation
damping enabled and passive TCP establishment
enabled. The value exact method of the MULTI_EXIT_DISC attribute damping
persistent peer oscillations is a four
octet unsigned number which left up to the
implementation, and is called a metric. All other factors
being equal, outside the exit point with lower metric should be preferred. If
received over external links, scope of
this document.

Status: Optional, used only if the MULTI_EXIT_DISC attribute MAY bgp peer has enabled
bgp peer oscillation damping with following optional
flags settings below.

Optional
attributes: 1) Perform automatic start flag SHOULD be
propagated over internal links to other BGP speakers within the same set
2) stop_peer_flap flag SHOULD be set
3) Passive TCP establishment flag SHOULD be set

Expiration Date July 2002 September 2003 [Page 24] 39]

RFC DRAFT January 2002

AS. The MULTI_EXIT_DISC attribute received from a neighboring AS MUST
NOT be propagated to other neighboring ASs.

A March 2003

Event8: Automatic stop

Definition: Local system automatically stops the
BGP speaker MUST IMPLEMENT a mechanism based on local configuration
which allows connection.

An example of an automatic stop event is
exceeding the MULTI_EXIT_DISC attribute to be removed from number of prefixes for a
route. This MAY be done prior to determining given
peer and the degree of preference
of local system automatically
disconnecting the route and performing route selection (decision process phases
1 and 2).

An implementation MAY also (based peer.

Status: Optional depending on local configuration) alter system

Optional
attributes: 1) Peform automatic stop flag SHOULD Be set

8.1.2 Timer Events

Event9: Connect retry timer expires

Definition: An event generated when the
value of Connect Retry timer
expires.

Status: Mandatory

Event10: Hold timer expires

Definition: An event generated when the MULTI_EXIT_DISC attribute received over an external
link. If it does so, it shall do so prior to determining Hold Timer expires.

Status: Mandatory

Event11: Keepalive timer expires

Definition: An event generated when the degree
of preference of Keepalive timer expires.
Status: Mandatory

Event12: Open Delay timer expires

Definition: An event generated when the route and performing route selection (decision
process phases 1 and 2).

5.1.5 LOCAL_PREF

LOCAL_PREF is a well-known attribute that SHALL Open Delay timer expires.

Status: Optional

Optional
attributes: If this event occurs,

Expiration Date September 2003 [Page 40]

RFC DRAFT March 2003

1) Delay Open flag SHOULD be included in all
UPDATE messages that a given BGP speaker sends to the other internal
peers. A BGP speaker SHALL calculate the degree of preference for
each external route based on set
2) Open Delay timer SHOULD be supported

Event13: Idle hold timer expires

Definition: An event generated when the locally configured policy, and
include Idle Hold Timer
expires indicating that the degree of preference when advertising session has completed
waiting for a route back-off period to its
internal peers. prevent bgp peer
oscillation.

The higher degree of preference MUST be preferred. A
BGP speaker shall use Idle Hold Timer is only used when the degree of preference learned via LOCAL_PREF
in its decision process (see section 9.1.1).

A BGP speaker MUST NOT include this attribute in UPDATE messages that
it sends to external peers, except for persistent
peer oscillation damping function is enabled.

Implementations not implementing the case of BGP Confederations
[13]. presistent peer
oscillation damping function may not have the Idle Hold
Timer.

Status: Optional

Optional
Attributes: If it is contained in an UPDATE message that is received from
an external peer, then this attribute MUST event occurs:
1) stop_peer_flap flag SHOULD be ignored by the
receiving speaker, except set indicating
support for persistent peer oscillation damping
functions,
2) Idle Hold timer should be supported

8.1.3 TCP Connection based Events

Event14: TCP connection valid indication

Definition: Event indicating the case local system reception of BGP Confederations [13].

5.1.6 ATOMIC_AGGREGATE

ATOMIC_AGGREGATE is
a well-known discretionary attribute.

When a router aggregates several routes for the purpose of
advertisement to TCP connection request with a particular peer, valid source
IP address and the AS_PATH of the aggregated
route excludes at least some of the AS numbers present in the AS_PATH TCP port, and valid destination
IP address and TCP Port. The definition of the routes that are aggregated, the aggregated route, when
advertised
invalid source, and invalid destination
IP address is left to the peer, MUST include the ATOMIC_AGGREGATE attribute.

A BGP speaker that receives a route with the ATOMIC_AGGREGATE
attribute MUST NOT remove the attribute from implementation.

BGP's destination port SHOULD be port
179 as defined by IANA.

TCP connection request is denoted by
the route when local system receiving a TCP SYN.

Expiration Date July 2002 September 2003 [Page 25] 41]

RFC DRAFT January 2002

propagating it to other speakers.

A BGP speaker that receives a route with March 2003

Status: Optional

Optional
Attributes: 1) The Track TCP state flag SHOULD be set if
this event occurs.

Event15: RCV TCP invalid indication

Definition: Event indicating the ATOMIC_AGGREGATE
attribute MUST NOT make any NLRI local system reception 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 TCP connection request 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 not be the path specified in the AS_PATH
attribute of the route.

5.1.7 AGGREGATOR

AGGREGATOR is either
an optional transitive attribute which may be included
in updates which are formed by aggregation (see Section 9.2.2.2). A invalid source address or port
number or an invalid destination
address or port number.

BGP speaker which performs route aggregation may add the AGGREGATOR
attribute which shall contain its own AS destination port number and IP address. SHOULD be 179
as defined by IANA.

Again, a TCP connection request
denoted by local system receiving a TCP
SYN.

Status: Optional

Optional
Attributes: 1) The
IP address Track TCP state should be set if this event
occurs.

Event16: TCP connection request Acknowledged

Definition: Event indicating the same as the BGP Identifier of the speaker.

6. BGP Error Handling.

This section describes actions Local system's request
to establish a TCP connection to be taken when errors are detected
while processing BGP messages.

When any of the conditions described here are detected, remote
peer.

The local system's TCP session sent a
NOTIFICATION message with the indicated Error Code, Error Subcode, TCP
SYN, and Data fields is sent, received a TCP SYN, ACK messages,
and the BGP connection is closed. If no
Error Subcode is specified, then Sent a zero must be used.

The phrase "the BGP TCP ACK.

Status: Mandatory

Event17: TCP connection is closed" means confirmed

Definition: Event indicates that the transport
protocol local system receiving
a confirmation that the TCP connection has
been closed, established by the associated Adj-RIB-In has
been cleared, remote site.

The remote peer's TCP engine sent a TCP SYN.
The local peer's TCP engine sent a SYN, ACK

Expiration Date September 2003 [Page 42]

RFC DRAFT March 2003

message, and now has received a final ACK.

Status: Mandatory

Event18: TCP connection fails

Definition: Event indicates that all resources for that BGP the local system has
received a TCP connection failure notice.

The remote BGP peer's TCP machine could have
been deallocated. Entries
sent a FIN. The local peer would respond
with a FIN-ACK. Another alternative is that
the local peer indicated a timeout in the Loc-RIB associated with
TCP session and downed the remote connection.

Status: Mandatory

8.1.4 BGP Messages based Events

Event19: BGPOpen

Definition: An event is generated when a valid OPEN
message has been received.

Status: Mandatory

optional
attributes: 1) Delay Open flag SHOULD not be set
2) Open Delay timer SHOULD not be running

Event20: BGPOpen with Open Delay Timer running

Definition: An event is generated when valid OPEN
message has been received for a peer are marked as invalid. The fact
that the routes have become
invalid has a successfully established
transport connection and is passed to other BGP peers before the routes are deleted
from the system.

Unless specified explicitly, currently
delaying the Data field sending of the NOTIFICATION a BGP open
message.

Status: Optional

Optional
attributes: 1) Delay Open Flag SHOULD be set

Expiration Date September 2003 [Page 43]

RFC DRAFT March 2003

2) Open Delay Timer SHOULD be running.

Event21: BGPHeaderErr

Definition: An event is generated when a received
BGP message that header is sent not valid.

Status: Mandatory

Event22: BGPOpenMsgErr

Definition: An event is generated when an OPEN message
has been received with errors.

Status: Mandatory

Event23: Open collision dump

Definition: An event generated administratively
when a connection collision has been
detected while processing an incoming
OPEN message and this connection has been
selected to indicate disconnected. See Section
6.8 for more information on collision
detection.

Event23 is an error administrative based only
implementation specific policy. This
Event may occur if the FSM is empty. implemented
as two linked state machines.

Status: Optional, depending on local system

Optional
Attributes: If the state machine is to process this
attribute in Established state,
1) Peform Collision detect in Established
flag SHOULD be set.

Please note: The Open collision dump can occur
in Idle, Connect, Active, OpenSent, OpenConfirm
without any optional flags being set.

Expiration Date July 2002 September 2003 [Page 26] 44]

RFC DRAFT January 2002

6.1 Message Header error handling.

All errors detected while processing the Message Header are indicated
by sending the March 2003

Event24: NotifMsgVerErr

Definition: An event is generated when a
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 "version
error" is received.

Status: Mandatory

Event25: NotifMsg

Definition: An event is generated when a
NOTIFICATION messages is received and
the error code is anything but
"version error".

Status: Mandatory

Event26: KeepAliveMsg

Definition: An event is generated when a KEEPALIVE
message header is all
ones if the received.

Status: Mandatory

Event27: UpdateMsg

Definition: An event is generated when a valid
UPDATE message type is OPEN. The expected value of the Marker
field received.

Status: Mandatory

Event28: UpdateMsgErr

Definition: An event is generated when an invalid
UPDATE message is received.

Status: Mandatory

8.2 Description of FSM

8.2.1 FSM Definition

BGP MUST maintain a separate FSM for all other types of each configured peer, Each BGP messages determined based on the
presence of the Authentication Information Optional Parameter
peer paired in the
BGP OPEN message and the actual authentication mechanism (if the
Authentication Information a potential connection unless configured to remain in

Expiration Date September 2003 [Page 45]

RFC DRAFT March 2003

the BGP OPEN message is present). The
Marker field should be all ones if idle state, or configured to remain passive, will attempt to to
connect to the OPEN message carried no
authentication information. If other. For the Marker field purpose of this discussion, the message header
is not active
or connect side of the expected one, then TCP connection (the side of a synchronization error has occurred
and TCP connection
(the side sending the Error Subcode first TCP SYN packet) is set to Connection Not Synchronized.

If the Length field called outgoing. The
passive or listening side (the sender of the message header first SYN ACK) is less than 19 or greater
than 4096, or if the Length field of called
an OPEN message is less than incoming connection (see Section 8.2.1.1 on the
minimum length terms active and
passive below).

A BGP implementation MUST connect to and listen on TCP port 179 for
incoming connections in addition to trying to connect to peers. For
each incoming connection, a state machine MUST be instantiated.
There exists a period in which the identity of the OPEN message, or if peer on the Length field other
end of an
UPDATE message incoming connection is less than the minimum length of the UPDATE message,
or if known but the Length field of a KEEPALIVE message BGP identifier is not equal to 19, or
if
known. During this time, both an incoming and an outgoing connection
for the Length field of a NOTIFICATION message same configured peering may exist. This is less referred to as a
connection collision (see Section 6.8).

A BGP implementation will have at most one FSM for each configured
peering plus one FSM for each incoming TCP connection for which the
peer has not yet been identified. Each FSM corresponds to exactly one
TCP connection.

There may be more than the
minimum length one connections between a pair of peers if the NOTIFICATION message, then the Error Subcode is
set
connections are configured to Bad Message Length. The Data field contains the erroneous
Length field.

If the Type field use a different pair of the message header IP addresses.
This is not recognized, then referred to as multiple "configured peerings" to the
Error Subcode same
peer.

8.2.1.1 Terms "active" and "passive"

The terms active and passive have been in our vocabulary for almost a
decade and have proven useful. The words active and passive have
slightly different meanings applied to a TCP connection or applied to
a peer. There is set only one active side and one passive side to Bad Message Type. The Data field contains the
erroneous Type field.

6.2 OPEN message error handling.

All errors detected while processing any
one TCP connection per the OPEN message are indicated
by sending definition above and the NOTIFICATION message with Error Code OPEN Message
Error. The Error Subcode elaborates state machine
below. When a BGP speaker is configured active it may end up on
either the specific nature active or passive side of the
error.

If the version number contained in the Version field of connection that eventually
gets established. Once the received
OPEN message TCP connection is not supported, then completed, it doesn't
matter which end was active and which end was passive and the Error Subcode is set to
Unsupported Version Number. The Data field only
difference is a 2-octets unsigned
integer, which indicates side of the largest locally supported version TCP connection has port number
less than the version the remote 179.

8.2.1.2 FSM and collision detection

There is one FSM per BGP peer bid (as indicated in the
received OPEN message), or if the smallest locally supported version connection. Prior to determining what peer
a connection is associated with there may be two connections for a

Expiration Date July 2002 September 2003 [Page 27] 46]

RFC DRAFT January 2002

number is greater March 2003

given peer. There SHOULD be no more than one connection per peer.
The collision detection identifies the version the remote BGP case where there is more than
one connection per peer bid, then the
smallest locally supported version number.

If and provides guidance for which connection to
get rid of. When this occurs, the Autonomous System field of corresponding FSM for the OPEN message connec-
tion that is unacceptable,
then closed SHOULD be disposed of.

8.2.1.3 FSM and Optional Attributes

Optional Attributes specify either flags that augment the Error Subcode is set to Bad Peer AS. The determination normal pro-
cessing of
acceptable Autonomous System numbers is outside the scope of this
protocol. BGP FSM, or optional timers. If the Hold Time field of the OPEN message is unacceptable, then the
Error Subcode MUST a Optional attribute
can 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 on a Hold Time MUST use the negotiated
value for system, the Hold Time.

If Events and the BGP Identifier field of FSM actions must be
support. For example, if the OPEN message is syntactically
incorrect, following options can be set in a BGP
implementation: AutoStart and Passive TCP connection Establishment
flag, then the Error Subcode events 3, 4 and 5 must be supported.

If an Optional attribute is cannot be set to Bad BGP Identifier.
Syntactic correctness means that (that is declared always
off logically), the BGP Identifier field represents
a valid IP host address.

If one events supporting that set of the Optional Parameters options do not have
to be supported.

8.2.1.4 FSM Event numbers

The Event numbers (1-28) utilized in this state machine description
aid in specifying the OPEN message is not
recognized, then the Error Subcode is set to Unsupported Optional
Parameters.

If one behavior of the Optional Parameters in the OPEN message is recognized,
but BGP state machine. Implementa-
tions MAY use these numbers to provide network management informa-
tion.

8.2.2 Finite State Machine

Idle state:

Initially BGP is malformed, then in the Error Subcode is set Idle state.

In this state BGP refuses all incoming BGP connections. No
resources are allocated to 0 (Unspecific).

If the OPEN message carries Authentication Information (as peer. In response to a
manual start event(Event1) or an
Optional Parameter), then the corresponding authentication procedure
is invoked. If automatic start
event(Event3), the authentication procedure (based on Authentication
Code and Authentication Data) fails, then local system:
- initializes all BGP resources,
- sets ConnectRetryCnt (the connect retry counter) to zero
- starts the Error Subcode is set connect retry timer with initial value,
- initiates a TCP connection to
Authentication Failure.

6.3 UPDATE message error handling.

All errors detected while processing the UPDATE message are indicated other BGP peer,
- listens for a connection that may be initiated by sending
the NOTIFICATION message remote BGP peer, and
- changes its state to Connect.

Expiration Date September 2003 [Page 47]

RFC DRAFT March 2003

An manual stop event (Event2) and Auto stop (Event 8) events are
are ignored in the Idle state.

In response to a manual start event with Error Code UPDATE Message
Error. The error subcode elaborates on the specific nature of passive TCP connection
flag (Event 4) or automatic start with the
error.

Error checking of an UPDATE message begins passive TCP connection
flag (Event 5), the local system:
- initializes all BGP resources,
- sets ConnectRetryCnt (the connect retry counter) to zero,
- starts the connect retry timer with initial value,
- listens for a connection that may be initiated by examining
the path
attributes. If remote peer, and
- changes its state to Active.

The exact value of the Withdrawn Routes Length or Total Attribute Length ConnectRetry timer is too a local
matter, but it SHOULD be sufficiently large (i.e., if Withdrawn Routes Length + Total Attribute
Length + 23 exceeds the message Length), then the Error Subcode is

Expiration Date July 2002 [Page 28]

RFC DRAFT January 2002

set to Malformed Attribute List. allow TCP
initialization.

If any recognized attribute has Attribute Flags that conflict with
the Attribute Type Code, then the Error Subcode persistent peer oscillation damping function is
enabled, three additional events may occur within Idle state:
- Automatic start with peer_stop_flap set to Attribute
Flags Error. [Event6],
- Automatic start with peer_stop_flag set [Event7],
- Idle Hold Timer expired [Event 13].

The Data field contains method of preventing persistent peer oscillation is
outside the erroneous attribute (type,
length and value).

If any recognized attribute has Attribute Length that conflicts with scope of this document.

Any other events [Events 9-12, 15-28] received in the expected length (based on Idle state does
not cause change in the attribute type code), then state of the
Error Subcode local system.

Connect State:

In this state, BGP is set waiting for the TCP connection to Attribute Length Error.
be completed.

The Data field
contains start events [Event 1, 3-7] are ignored in connect
state.

In response to a manual stop event [Event2], the erroneous attribute (type, length and value).

If any of local system:
- drops the mandatory well-known attributes are not present, then TCP connection,
- releases all BGP resources,
- sets ConnectRetryCnt (the connect retry count) to zero
- resets the Error Subcode is set connect retry timer (sets to zero), and
- changes its state to Idle.

Expiration Date September 2003 [Page 48]

RFC DRAFT March 2003

In response to Missing Well-known Attribute. The Data
field contains the Attribute Type Code of connect retry timer expires event [Event
9], the missing well-known
attribute.

If any of local system:
- drops the mandatory well-known attributes are not recognized,
then TCP connection,
- restarts the Error Subcode is set to Unrecognized Well-known Attribute.
The Data field contains connect retry timer,
- stops the unrecognized attribute (type, length Open Delay timer and
value).

If resets the ORIGIN attribute has an undefined value, then timer to zero,
- initiates a TCP connection to the Error
Subcode is set other BGP peer,
- continues to Invalid Origin Attribute. The Data field contains listen for a connection that may be
initiated by the unrecognized attribute (type, length remote BGP peer, and value).
- stays in Connect state.

If the NEXT_HOP attribute field is syntactically incorrect, then Open Delay timer expires [Event12] in the
Error Subcode is set to Invalid NEXT_HOP Attribute. The Data field
contains connect
state, the incorrect attribute (type, length and value). Syntactic
correctness means that local system:
- sends an OPEN message to its peer,
- sets the NEXT_HOP attribute represents hold timer to a valid IP
host address. Semantic correctness applies only large value, and
- changes its state to OpenSent.

If the external BGP
links, and only when port receives a valid TCP connection indication
[Event 14], the sender TCP connection is processed and
the receiving speaker are one IP
hop away from each other. To be semantically correct, the IP address connection remains in the NEXT_HOP must not be the IP address of the receiving speaker,
and the NEXT_HOP IP address must either be the sender's IP address
(used to establish Connect state.

If the BGP session), or TCP connection receives an invalid indication [Event 15]:
the interface associated with local system rejects the NEXT_HOP IP address must share a common subnet with TCP connection and the receiving
BGP speaker. connection
remains in the Connect state.

If the NEXT_HOP attribute is semantically incorrect, TCP connection succeeds [Event 16 or
Event 17], the
error should be logged, and local system checks the route should be ignored. In this
case, no NOTIFICATION message should be sent.

The AS_PATH attribute is checked for syntactic correctness. Delay Open flag prior
to processing. If the
path Delay Open flag is syntactically incorrect, then set, the Error Subcode is local system:
- clears the connect retry timer,
- set the Open Delay timer to
Malformed AS_PATH.

The information carried by the AS_PATH attribute is checked for AS
loops. AS loop detection initial value, and
- stays in the Connect state.
If the Delay Open flag is done by scanning not set, the full AS path (as

Expiration Date July 2002 [Page 29]

RFC DRAFT January 2002

specified in local system:
- clears the AS_PATH attribute), connect retry timer,
- completes BGP initialization
- sends an OPEN message to its peer,
- sets hold timer to a large value, and checking that the autonomous
system number
- changes its state to OpenSent.

A hold timer value of 4 minutes is suggested.

If the TCP connection fails [Event18], the local system does not appear in checks
the AS path. Open Delay Timer. If the autonomous system number appears in Open Delay timer is running,
the AS path local system:
- restarts the route connect retry time with initial value,
- stops the Open Delay timer and resets value to zero,
- continues to listen for a connection that may be
stored in
initiated by the Adj-RIB-In, but unless remote BGP peer, and
- changes its state to Active.
If the router open Delay timer is configured to
accept routes with its own autonomous system in not running, the AS path, local system:

Expiration Date September 2003 [Page 49]

RFC DRAFT March 2003

- resets the
route shall not be passed connect retry timer (sets to zero), and
- Drops the TCP connection,
- Releases all BGP Decision Process. Operations of
a router that is configured resources,
- and goes to accept routes with its own autonomous
system number in the AS path are outside the scope of this document. Idle State.

If an optional attribute OPEN message is recognized, then received with the value of this
attribute is checked. If an error Open Delay timer is detected,
running [Event 20], the attribute is
discarded, and local system:
- clears the Error Subcode is set connect retry timer (cleared to Optional Attribute Error.
The Data field contains zero),
- completes the attribute (type, length BGP initialization,
- stops and value).

If any attribute appears more than once in clears the UPDATE Open Delay timer,
- sends an OPEN message, then
- sends a Keepalive message,
- If the Error Subcode hold timer value is set non-zero,
- start the keepalive timer to Malformed Attribute List.

The NLRI field in inital value,
- reset the UPDATE message hold timer to the negotiated value,
else if hold timer value is checked for syntactic
validity. zero,
- reset the keepalive timer. and
- reset the hold timer value to zero.
- and changes its state to OpenConfirm.

If the value of the autonomous system field is syntactically incorrect, then the Error
Subcode is same as the local
Autonomous System number, set the connection status to Invalid Network Field. an internal
connection; otherwise it is "external".

If BGP message header checking detects an error [Event 21] or
OPEN message checking detects an error [Event 22] (see section
6.2), the local system:
- (optionally) If a prefix in the NLRI field Send Notification without Open flag is semantically incorrect (e.g., an
unexpected multicast IP address), an error should be logged locally,
and set,
then the prefix should be ignored.

An UPDATE message that contains correct path attributes, but no NLRI,
shall be treated as local system first sends a valid UPDATE message.

6.4 NOTIFICATION message
with the appropriate error handling.

If a code, and then

- resets the connect retry timer (sets to zero),
- releases all BGP resources,
- drops the TCP connection,
- increments the ConnectRetryCnt (connect retry count) by 1,
- [optionally] performs peer sends oscillation damping,
- and goes to Idle.

If a NOTIFICATION message, and there is an error in that
message, there message is unfortunately no means of reporting this error via received with a subsequent NOTIFICATION message. Any such error, such as an
unrecognized Error Code or Error Subcode, should be noticed, logged
locally, and brought to version
error[Event24], the attention of local system checks the administration of Open Delay timer.
If the
peer. The means Open Delay timer is running, the local system:
- resets the connect retry timer (sets to do this, however, lies outside zero),
- stops and reset the scope of this
document.

6.5 Hold Timer Expired error handling. Open Delay timer (sets to zero),
- releases all BGP resources,
- drops the TCP connection,
- changes its state to Idle.
If a system does the Open Delay timer is not receive successive KEEPALIVE and/or UPDATE
and/or NOTIFICATION messages within running, the local system:
- resets the connect retry timer (sets to zero),

Expiration Date September 2003 [Page 50]

RFC DRAFT March 2003

- releases all BGP resources,
- drops the TCP connection,
- increments the ConnectRetryCnt (connect retry count) by 1,
- optionally performs peer oscillation damping, and
- changes its state to Idle.

In response to any other events [Events 8,10-11,13,19,23,
25-28] the period specified in local system:
- if the Hold
Time field of connect retry timer is running,
stop and reset the OPEN message, then connect retry timer (sets to zero),
- if the NOTIFICATION message with
Hold Timer Expired Error Code must be sent Delay Open timer is running,
stop and reset the Delay Open timer (sets to zero),
- releases all BGP connection

Expiration Date July 2002 [Page 30]

RFC DRAFT January 2002

closed.

6.6 Finite State Machine error handling.

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

6.7 Cease. ConnectRetryCnt (connect retry count) by 1,
- optionally performs peer oscillation damping, and
- changes its state to Idle.

Active State:

In absence of any fatal errors (that are indicated in this section),
a state BGP peer may choose at any given time is trying to close its BGP connection acquire a peer by sending listening
for and accepting a TCP connection.

The start events [Event1, 3-7] are ignored in the Active
state.

A manual stop event[Event2], the local system:
- If the Delay Open timer is running and the
Send NOTIFICATION message with Error Code Cease. However, without Open flag is set,
the Cease local system Sends a NOTIFICATION message must not be used when with a fatal error
indicated by this section does exist.

A Cease,
- releases all BGP speaker may support resources including
- stopping the ability to impose an (locally
configured) upper bound on Open delay timer
- drops the number of address prefixes TCP connection,
- sets ConnectRetryCnt (connect retry count) to zero
- resets the speaker
is willing connect retry timer (sets to zero),
- changes its state to accept from a neighbor. When Idle.

In response the upper bound is
reached, ConnectRetry timer expires event[Event9],
the speaker (under control of local configuration) may
either (a) discard new address prefixes from the neighbor, or (b)
terminate the BGP peering with system:
- restarts the neighbor. If connect retry timer (with initial value),
- initiates a TCP connection to the other BGP speaker
decides peer,
- Continues to terminate listen for TCP connection that may be
initiated by remote BGP peer,
- and changes its peering with a neighbor because state to Connect.

Expiration Date September 2003 [Page 51]

RFC DRAFT March 2003

If the number
of address prefixes received from local system has the neighbor exceeds Open Delay timer expired
[Event12], the locally
configured upper bound, then local system:
- clears the speaker must send connect retry timer (set to zero),
- stops and clears the neighbor a
NOTIFICATION message with Open Delay timer (set to zero),
- completes the Error Code Cease.

6.8 Connection collision detection.

If a pair of BGP speakers try simultaneously initialization,
- sends the OPEN message to it's remote peer,
- sets its hold timer to establish a BGP
connection large value, and
- changes its state to each other, then two parallel connections between this
pair of speakers might well be formed. If the source IP address used
by one OpenSent.

A hold timer value of these connections 4 minutes is also suggested for this
state transition.

If the same as the destination IP address
used by the other, and local system receives a valid TCP indication
[Event 14], the destination IP address used by local system processes the first TCP connection is
flags, and stays in Active state.

If the same as local system receives an invalid TCP indication [Event 15]:
the source IP address used by local system rejects the other, we
refer to this situation as connection collision. Clearly TCP connection, and stays in
the
presence of Active State.

A TCP connection collision, one of these connections must be
closed.

Based on succeeds [Event 16 or Event 17], the value of
local system checks the BGP Identifier a convention is established
for detecting which BGP connection is "Delay Open Flag" prior to be preserved when a
collision does occur. The convention
processing. If the Delay Open flag is to compare set, the BGP

Expiration Date July 2002 [Page 31]

RFC DRAFT January 2002

Identifiers of local system
o clears the peers involved in connect retry timer,
o sets the collision and BGP Open Delay timer to retain only the connection initiated by initial value, and
o stays in the BGP speaker with Active state.

-If the higher-valued
BGP Identifier.

Upon receipt of an OPEN message, Delay Open flag is not set, the local system must examine all of
its connections that are in
o clears the OpenConfirm state. A BGP speaker may
also examine connections in an OpenSent state if it knows connect retry timer,
o completes the BGP
Identifier of initialization,
o sends the peer by means outside OPEN message to it's peer,
o sets its hold timer to a large value, and
o changes its state to OpenSent.

A hold timer value of 4 minutes is suggested as a "large value" for
the protocol. hold timer.

If among
these connections there is the local system receives a TCP connection fails event [Event 18],
the local system will:
- restart connect retry timer (with initial value),
- stops and clears Open Delay Timer (sets the value to a remote BGP speaker whose zero),
- release all BGP Identifier equals the one in resources
- Acknowledge the OPEN message, drop of TCP connection if
TCP disconnect (send a FIN ACK),
- Increment ConnectRetryCnt (connect retry count) by 1, and
- optionally perform peer oscillation damping,

Expiration Date September 2003 [Page 52]

RFC DRAFT March 2003

- and this
connection collides with the connection over which the go to to Idle.

If an OPEN message is received then the local system performs with the following collision
resolution procedure:

1. The BGP Identifier of Open Delay timer is
running [Event 20], the local system is compared to
- clears the BGP
Identifier of connect retry timer (cleared to zero),
- stops and clears the remote system (as specified in Open Delay timer
- completes the BGP initialization,
- sends an OPEN
message).

2. If message,
- send a Keepalive message, and
- if the hold timer value of the local BGP Identifier is less than non-zero,
- starts the
remote one, keepalive timer to initial value,
- resets the local system closes BGP connection that already
exists (the one that is already in hold timer to the OpenConfirm state), and
accepts BGP connection initiated by negotiated value,
else if the remote system.

3. Otherwise, hold timer is zero
- resets the local system closes newly created BGP connection
(the one associated with keepalive timer (set to zero),
- resets the newly received OPEN message), and
continues hold timer to use zero.
- changes its state to OpenConfirm.

If the existing one (the one that is already in value of the
OpenConfirm state).

Comparing BGP Identifiers autonomous system field is done by treating them the same as (4-octet
long) unsigned integers.

Unless allowed via configuration, a connection collision with an
existing BGP connection that is in Established state causes
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 local
Autonomous System number, set the BGP connection (that results from the collision
resolution procedure) is accomplished by sending status to an internal
connection; otherwise it is "external".

If BGP message header checking detects an error [Event 21] or OPEN
message checking detects an error [Event 22] (see section 6.2), the
local system:
- (optionally) sends NOTIFICATION message with the Error Code Cease.

7. BGP Version Negotiation.
appropriate error code,
- resets the connect retry timer (sets to zero),
- releases all BGP speakers may negotiate resources,
- drops the version of TCP connection,
- increments the protocol ConnectRetryCnt (connect retry count) by making

Expiration Date July 2002 [Page 32]

RFC DRAFT January 2002

multiple attempts 1,
- [optionally] performs peer oscillation damping,
- and goes to open Idle.

If a BGP connection, starting NOTIFICATION message is received with the highest a version number each supports. If an open attempt fails with an Error
Code OPEN Message Error, and an Error Subcode Unsupported Version
Number, then
error[Event24], the BGP speaker has available local system checks the version number it
tried, Open Delay timer.
If the version number its peer tried, Open Delay timer is running, the version number passed
by its peer in local system:
- resets the NOTIFICATION message, connect retry timer (sets to zero),
- stops and reset the version numbers that
it supports. If Open Delay timer (sets to zero,
- releases all BGP resources,
- drops the two peers do support one or more common versions,
then this will allow them TCP connection,
- changes its state to rapidly determine Idle.
If the highest common
version. In order Open Delay timer is not running, the local system:
- resets the connect retry timer (sets to support BGP version negotiation, future versions
of zero),
- releases all BGP must retain resources,
- drops the format of TCP connection,
- increments 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 ConnectRetryCnt (connect retry count) by 1,
- optionally performs peer oscillation damping, and

Expiration Date September 2003 [Page 53]

RFC DRAFT March 2003

- changes its state as determined by this FSM.

Initially BGP is in the Idle state. to Idle state:

A manual start event is a start event initiated by an operator.
An automatic start event is a start

In response to any other event generated by [Events 8,10-11,13,19,23,25-28],
the
system. local system:
- resets the connect retry timer (sets to zero),
- drops the TCP connection,
- releases all BGP resources,
- increments the ConnectRetryCnt (connect retry count) by one,
- optionally performs peer oscillation damping, and
- changes its state to Idle.

OpenSent:

In this state BGP refuses all incoming BGP connections. No
resources waits for an OPEN message from its peer.

The Start events [Event1, 3-7] are allocated to ignored in the peer. In response to OpenSent
state.

If a Start manual stop event (manual or automatic), [Event 2] is issued in Open sent
state, the local system:
- initializes sends the NOTIFICATION with a cease,
- release all BGP resources,
- starts drops the ConnectRetry timer, TCP connection,
- initiates a transport connection set ConnectRetryCnt (connect retry count) to the other BGP peer, zero,
- listens for a connection that may be initiated by resets the
remote BGP peer, Connect Retry timer (set to zero), and
- changes its state to connect.

The exact value of the ConnectRetry timer is a local matter,
but it should be sufficiently large to allow TCP
initialization.

Any other Idle.

If an automatic stop event received [Event 8] is issued in the IDLE OpenSent
state, is ignored.

Expiration Date July 2002 [Page 33]

RFC DRAFT January 2002

IdleHold state:

The IdleHold state keeps the system in "Idle" mode until a
certain time period has passed or an operator intervenes to
manually restart local system:
- sends the connection. This "IdleHold timeout"
prevents persistent flapping of NOTIFICATION with a BGP peering session.

Upon entering cease,
- release all the Idle Hold state, if BGP resources
- drops the IdleHoldTimer exceeds TCP connection,
- increments the local limit ConnectRetryCnt (connect retry count) by 1,
- optionally performs peer oscillation damping, and
- changes its state to Idle.

If the "Keep Idle" flag is set.

Upon receiving a Manual start, Hold Timer expires[Event 10], the local system:
- clears the IdleHoldtimer, send a NOTIFICATION message with error code Hold
Timer Expired,
- clears "keep Idle" flag reset the connect retry timer (sets to zero),
- initializes releases all BGP resources,
- starts drops the ConnectRetry timer, TCP connection,
- initiates a transport connection to increments the other BGP peer,

- listens for a connection that may be initiated ConnectRetryCnt (connect retry count) by the
remote BGPPeer, 1, and
- changes its state to connect.

Upon receiving Idle.

If a IdleHoldtimer expired event, the local system
checks to see that the Keep Idle flag TCP indication is set. If the Keep Idle
flag received for valid connection
[Event 14] or TCP request aknowledgement [Event 16]

Expiration Date September 2003 [Page 54]

RFC DRAFT March 2003

is set, the system stays received, or a TCP connect confirm [Event 17] is
received a second TCP session may be in progress. This
second TCP session is tracked per the "Idle Hold" state. Connection Collision
processing (Section 6.8) until an OPEN message is received.

A TCP connection for an invalid port [Event 15] is ignored.

If the Keep Idle flag a TCP connection fails event [Event18] indication is not set, received
the local system:
- clears closes the IdleHoldtimer, BGP connection,
- and transitions restarts the state Connect Retry timer,
- continues to Idle.

Getting out of the IdleHoldstate requires either operator
intervention via listen for a manual start or the IdleHoldtimer to expire
with the "Keep Idle" flag to connection that may be clear.

Any other event received in
initiated by the IdleHold state is ignored.

Connect State:

In this state, remote BGP is waiting for the transport protocol
connection peer, and
- changes its state to be completed.

Expiration Date July 2002 [Page 34]

RFC DRAFT January 2002 Active.

When an OPEN message is received, all fields are checked
for correctness. If there are no errors in the transport connection succeeds, OPEN message
[Event 19] the local system:
- clears resets the ConnectRetry timer, Open Delay timer to zero,
- completes initialization, reset BGP Connect Timer to zero,
- send an Open sends a KEEPALIVE message to its peer, and
- set Hold sets a KeepAlive timer (via the text below)
- sets the hold timer according to a large value, the negotiated value
(see Section 4.2), and
- changes its state to Open Sent.

A OpenConfirm.

If the negotiated hold timer time value of 4 minutes is suggested. zero, then the Hold and
KeepAlive timers are not started. If the transport protocol value of the Autonomous
System field is the same as the local Autonomous System number,
then the connection fails (e.g.,
retransmission timeout), is an "internal" connection; otherwise, it
is an "external" connection. (This will impact UPDATE processing
as described below.)

If the BGP message header checking [Event21] or OPEN message
check detects an error (see Section 6.2)[Event22], the local system:
- restarts the ConnectRetry timer, sends a NOTIFICATION message with appropriate error
code,
- continues resets the connect retry timer (sets to listen for a zero),
- releases all BGP resources,
- drops the TCP connection that may be initiated
by
- increments the remote BGP peer, ConnectRetryCnt (connect retry cout) by 1,
- optionally performs peer oscillation damping, and
- changes its state to Active.

In response Idle.

Collision detection mechanisms (Section 6.8) need to be
applied when a valid BGP OPEN message is received [Event 19 or

Expiration Date September 2003 [Page 55]

RFC DRAFT March 2003

Event 20]. Please refer to Section 6.8 for the ConnectRetry timer expired event, details of
the local
system:

- restarts comparison. An administrative collision detect is when
BGP implementation determines my means outside the ConnectRetry timer,

- initiates scope of
this document that a transport connection collision has occurred.

If a connection in OpenSent is determined to be the other BGP peer,

- continues to listen for a
connection that may must be initiated
by closed, an open collision dump [Event 23]
is signaled to the remote BGP peer, and

- stays in Connect state.

The start state machine. If such an event (manual or automatic) is ignored
received in the Connect
state.

In response to any other event (initiated by the system or
operator), OpenSent, the local system:
- IdleHoldtimer = 2**(ConnectRetryCnt)*60

- Increment ConnectRetryCnt by 1, sends a NOTIFICATION with a Cease
- Set resets the connect retry timer to zero,

Expiration Date July 2002 [Page 35]

RFC DRAFT January 2002

- Drops TCP connection, timer,
- Releases releases all BGP resources,
- drops the TCP connection,
- increments ConnectRetryCnt (connect rery count) by 1,
- optionally performs peer oscillation damping, and
- Goes to IdleHoldstate

Active State:

In this changes its state BGP is trying to acquire Idle.

If a peer by listening for
and accepting NOTIFICATION message is received with a transport protocol connection.

If the transport connection succeeds, version
error[Event24], the local system:
- clears resets the ConnectRetry timer, connect retry timer (sets to zero)
- completes the initialization, releases all BGP resources,
- sends drops the Open message to it's peer,

- sets its Hold timer to a large value, TCP connection,
- and changes its state to OpenSent.

A Hold timer value of 4 minutes is suggested. Idle.

In response the ConnectRetry timer expired event, to any other event [Events 9, 11-13,20,25-28],
the local system:
- restarts sends the ConnectRetry timer,

- initiates a transport connection to NOTIFICATION with the other BGP peer, Error Code Finite
state machine error,
- continues resets the connect retry timer (sets to listen for connection that may be initiated
by remote BGP peer, zero),
- and changes its state to Connect.

If the local system does not allow releases all BGP connections with
unconfigured peers, then the local system: resources
- rejects connections from IP addresses that are not
configured peers, drops the TCP connection,
- and remains in increments the Active state.

The start events (initiated ConnectRetryCnt (connect retry count) by the system 1,
- optionally performs peer oscillation damping, and
- changes its state to Idle.

OpenConfirm State:

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

Any start event [Event1, 3-7] is ignored in the Active OpenConfirm
state.

Expiration Date July 2002 September 2003 [Page 36] 56]

RFC DRAFT January 2002 March 2003

In response to any other event (initiated a manual stop event[Event 2] initiated by
the system or
operator), operator, the local system:
- IdleHoldtimer = 2**(ConnectRetryCnt)*60 sends the NOTIFICATION message with Cease,
- Increment releases all BGP resources,
- drop the TCP connection,
- sets the ConnectRetryCnt by 1, (connect retry count) to zero
- Set sets the connect retry timer to zero, and
- Drops TCP connection,

- Releases all BGP resources,

- Goes changes its state to IdleHold state.

Open Sent: Idle.

In this state BGP waits for an Open Message from its peer.
When an OPEN message is received, all fields are check for
correctness. If response to the BGP message header checking or OPEN
message check detects an error (see Section 6.2), or a
connection collision (see Section 6.8) Automatic stop event initiated by the
system[Event 8], the local system:
- sends a the NOTIFICATION message with Cease,
- IdleHoldtimer = 2**(ConnectRetryCnt)*60 connect retry timer reset (set to zero)
- Increment release all BGP resources,
- drops the TCP connection,
- increments the ConnectRetryCnt (connect retry count)
by 1,
- Set optionally performs peer oscillation damping,
- changes its state to Idle.

If the Hold Timer expires before a KEEPALIVE message is
received [Event 10], the local system:
- send the NOTIFICATION message with the error code
set to Hold Time Expired,
- resets the connect retry timer (sets the timer to zero, and

- Drops TCP connection, to
zero),
- Releases releases all BGP resources,
- Goes to IdleHold state.

If there are no errors in drops the OPEN message, TCP connection,
- increments the ConnectRetryCnt (connect retry count)
by 1,
- optionally performs peer oscillation damping,
and
- changes its state to Idle.

If the local system receives a KEEPALIVE timer expires
event [Event 11], the system:
- sends a KEEPALIVE message and message,
- sets a KeepAlive timer (via restarts the text below) Keepalive timer, and
- set remains in OpenConfirmed state.

In the Hold timer according to event of TCP connection valid indication [Event 14], or TCP
connection succeeding [Event 16 or Event 17] while in OpenConfirm,
the negotiated value (see
section 4.2),

- set local system needs to track the state 2nd connection.

If a TCP connection is attempted to Open Confirm. an invalid port [Event
15], the local system will ignore the second connection
attempt.

Expiration Date July 2002 September 2003 [Page 37] 57]

RFC DRAFT January 2002

If the negotiated Hold time value is zero, then the Hold Time
timer and KeepAlive timers are not started. March 2003

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 an "external" connection. (This will impact
UPDATE processing as described below.)

If system receives a disconnect NOTIFICATION is received TCP connection fails event
[Event 18] from the underlying
transport protocol, TCP. or a NOTIFICATION
message [Event 25] the local system:
- closes resets the connect retry timer (sets the timer to
zero),
- releases all BGP resources,
- drops the TCP connection,
- restarts increments the Connect Retry timer, ConnectRetryCnt (connect retry count)
by 1,
- and continues optionally performs peer oscillation damping,
- changes its state to listen for a connection that may be
initiated by the remote BGP peer, and goes into Active
state. Idle.

If the Hold Timer expires, the local system:

- send system receives a NOTIFICATION message [Event 24] with error code Hold Timer
Expired,

- IdleHoldtimer = 2**(ConnectRetryCnt)*60

- Increment ConnectRetryCnt by 1,
a version error, the local system:
- Set resets the connect retry timer (sets the timer to zero, and

- Drops TCP connection, zero),
- Releases releases all BGP resources, and
- Goes to IdleHold state.

The Start event (manual and automatic) drops the TCP connection,
- changes its state to Idle. [Verify this/or above]

If the OPEN message is ignored in valid [Event 19], the
OpenSent state. collision
detect function is processed per Section 6.8. If a NOTIFICATION message this
connection is received with a version error, to be dropped due to connection collision, the
local system:
- Closes sends a NOTIFICATION with a Cease
- resets the transport connection Connect timer (set to zero),
- Releases releases all BGP resources,
- ConnectRetryCnt = 0, drops the TCP connection (send TCP FIN),
- Connect increments the ConnectRetryCnt by 1 (connect retry timer = 0, count), and

Expiration Date July 2002 [Page 38]

RFC DRAFT January 2002
- transition to Idle state. optionally performs peer oscillation damping.

If any other NOTIFICATION an OPEN message is received, all fields are check for
correctness. If the BGP message header checking [Event21]
or OPEN message check detects an error (see Section
6.2)[Event22], the local system:
- IdleHoldtimer = 2**(ConnectRetryCnt)*60

- Increment ConnectRetryCnt by 1, sends a NOTIFICATION message with appropriate error
code,
- Set resets the connect retry timer (sets the timer to zero, and

- Drops TCP connection,
zero),
- Releases releases all BGP resources,
- Goes to IdleHold state.

In response to any other event, drops the local system: TCP connection,
- sends increments the NOTFICATION message with Error Code Finite State
Machine Error,

- IdleHoldtimer = 2**(ConnectRetryCnt)*60

- Increment ConnectRetryCnt (connect retry count) by 1,
- Set connect retry timer to zero,

- Drops TCP connection,

- Releases all BGP resources, optionally performs peer oscillation damping, and
- Goes to IdleHold state.

Open Confirm State

In this changes its state to Idle.

If during the processing of another OPEN message, the BGP waits for

Expiration Date September 2003 [Page 58]

RFC DRAFT March 2003

implementation determines my means outside the scope of
this document that a KEEPALIVE or NOTIFICATION
message.

If connection collision has occurred and
this connection is to be closed, the local system receives will
issue a KEEPALIVE message, it changes
its state to Established.

If open collision dump [Event 23]. When the Hold Timer expires before local
system receives a KEEPALIVE message is
received, open collision dump event [Event 23], the
local system:
- send the a NOTIFICATION message with the error code Hold
Timer Expired,

- sets IdleHoldTimer = 2**(ConnectRetryCnt)*60

Expiration Date July 2002 [Page 39]

RFC DRAFT January 2002

- Increments ConnectRetryCnt by 1, a Cease
- Sets resets the connect retry timer to zero, timer,
- Drop the releases all BGP resources
- drops all TCP connection,
- Releases all BGP resources, increments the ConnectRetryCnt (connect retry count) by 1,
- optionally performs peer oscillation damping, and
- Goes changes its state to IdleHoldState. Idle.

If the local system receives a NOTIFICATION message or receives
a disconnect NOTIFICATION from KEEPALIVE message[Event 26],
- restarts the underlying transport
protocol, Hold timer, and
- changes its state to Established.

In response to any other event [Events 9, 12-13, 27-28],
the local system:
- Sets IdleHold Timer = 2**(ConnectRetryCnt)*60

- Increments ConnectRetryCnt by 1, sends a NOTIFICATION with a code of Finite State
Machine Error,
- Sets resets the connect retry timer (sets to zero, zero)
- Drops releases all BGP resources,
- drops the TCP connection,
- Releases all BGP resources, increments the ConnectRetryCnt (connect retrycount) by 1,
- optionally performs peer oscillation damping, and
- Goes changes its state to IdleHoldstate. Idle.

Established State:

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

Any start event (Event 1, 3-7) is ignored in the
Established state.

In response to the Stop a manual stop event initiated (initiated by the system, an
operator)[Event2], the local system: sytem:
- sends the NOTIFICATION message with Cease,
- sets IdleHoldtimer = 2**(ConnectRetryCnt)*60

- Increments ConnectRetryCnt by 1,

- Sets resets the Connect connect retry timer to zero, zero (0),
- Drops the TCP delete all routes associated with this connection,
- Releases all release BGP resources,
- Goes drops TCP connection,
- sets ConnectRetryCnt (connect retry count)

Expiration Date September 2003 [Page 59]

RFC DRAFT March 2003

to zero (0), and
- changes its state to IdleHoldstate. Idle.

In response to a Stop an automatic stop event initiated by the operator,
system (automatic) [Event8], the local system:
- sends the a NOTIFICATION message with Cease,

Expiration Date July 2002 [Page 40]

RFC DRAFT January 2002
- resets the connect retry timer (sets to zero)
- deletes all routes associated with this connection,
- releases all BGP resources resources,
- sets drops the ConnectRetryCnt to zero TCP connection,
- sets increments the connect retry timer to 0 ConnectRetryCnt (connect retry count)
by 1,
- optionally performs peer oscillation damping, and
- transitions changes its state to Idle state.

The Start Idle.

An example automatic stop event is ignored in exceeding the OpenConfirm state.

In response to any other event, number of
prefixes for a given peer and the local system
automatically disconnecting the peer.

If the Hold timer expires [Event10], the local system:
- sends a NOTIFICATION message with a code of Finite State Machine
Error,

- sets IdleHoldtimer = 2**(ConnectRetryCnt)*60

- Increments ConnectRetryCnt by 1, Error Code Hold
Timer Expired,
- Sets resets the Connect connect retry timer (sets to zero,

- Drops the TCP connection, zero),
- Releases releases all BGP resources,
- Goes to IdleHoldstate.

Established State:

In drops the Established state BGP can exchange UPDATE, NOTFICATION, TCP connection,
- increments the ConnectRetryCnt (connect retry count)
by 1,
- optionally performs peer oscillation damping, and KEEPALIVE messages with
- changes its peer. state to Idle.

If the KeepAlive timer expires [Event11], the local system receives an UPDATE or
sends a KEEPALIVE message, it restarts its Hold Timer, if KeepAlive timer,
unless the negotiated Hold Time value is
non-zero.

If zero.

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

A TCP connection indication [Event 14] received
for a valid port will cause the 2nd connection to be
tracked.

A TCP connection indications for invalid port [Event 15],
will be ignored.

In response to a TCP connection succeeds [Event 16

Expiration Date September 2003 [Page 60]

RFC DRAFT March 2003

or Event 17], the 2nd connection SHALL be tracked until
it sends an OPEN message.

If a valid OPEN message [Event 19] is received, it will be
checked to see if it collides (Section 6.8) with any other
session. If the BGP implementation determines that this
connection needs to be terminated, it will process an open
collision dump event[Event 23]. If this session needs to be
terminated, the connection will be terminated by:

- send a NOTIFICATION message or with a
disconnect from the underlying transport protocol, it:

- sets IdleHoldtimer = 2**(ConnectRetryCnt)*60,

- Increments ConnectRetryCnt by 1, Cease,
- Sets resets the Connect connect retry timer time (sets to zero, zero),
- Drops the TCP deletes all routes associated with this connection,
- Releases release all BGP resources, and

Expiration Date July 2002 [Page 41]

RFC DRAFT January 2002
- Goes to IdleHoldstate.

If the local system receives an UPDATE message, and the Update
message error handling procedure (see Section 6.3) detecs an
error, drops the local system:

- sends a NOTIFICATION message with Update error,

- sets IdleHoldtimer = 2**(ConnectRetryCnt)*60 TCP connection,
- Increments increments ConnectRetryCnt (connect retry count)
by 1,
- Sets the Connect retry timer to zero,

- Drops the TCP connection,

- Releases all BGP resources, optionally performs peer oscillation damping, and
- Goes changes its state to IdleHoldstate. Idle.

If the Hold timer expires, the local system:

- sends system receives a NOTIFICATION message with Error Code Hold Timer
Expired,

- sets IdleHoldtimer = 2**(ConnectRetryCnt)*60

- Increments ConnectRetryCnt by 1,
[Event24 or Event 25] or a TCP connections fails [Event18]
from the underlying TCP, it:
- Sets resets the connect retry timer (sets to zero, zero),
- Drops the TCP delete all routes associated with this connection,
- Releases releases all the BGP resources,
- Goes drops the TCP connection,
- increments the ConnectRetryCnt (connect retry count)
by 1, and
- changes its state to IdleHold state. Idle.

If the KeepAlive timer expires, the local system sends receives a KEEPALIVE message, it message
[Event 26], the local system will:
- restarts its KeepAlive timer, unless Hold Timer, if the negotiated Hold Time
value is zero.

Each time time non-zero, and
- remain in the Established state.

If the local system sends a KEEPALIVE or receives an UPDATE
message, it message [Event27],
the local system will:
- process the update packet
- restarts its KeepAlive Hold timer, unless if the negotiated Hold Time
value is zero.

In response to non-zero, and
- remain in the Stop event initiated by Established state.

Expiration Date September 2003 [Page 61]

RFC DRAFT March 2003

If the local system
(automatic), receives an UPDATE message, and the
UPDATE message error handling procedure (see Section 6.3)
detects an error [Event28], the local system:

Expiration Date July 2002 [Page 42]

RFC DRAFT January 2002
- sends a NOTIFICATION message with Cease,

- sets IdleHoldtimer = 2**(ConnectRetryCnt)*60

- increments ConnectRetryCnt by 1, Update error,
- sets resets the connect retry timer (sets to zero, zero),
- drops the TCP delets all routes associated with this connection,
- releases all BGP resources,
- goes to IdleHold state, and

- deletes all routes.

An example automatic stop event is exceeding the number of
prefixes for a given peer and the local system automatically
disconnecting drops the peer.

In response to a stop event initiated by an operator:

- release all resources (including deleting all routes), TCP connection,
- set increments the ConnectRetryCnt to zero (0),

- set connect (connect retry timer to zero (0), count)
by 1,
- optionally performs peer oscillation damping, and
- transition changes its state to the Idle.

The Start event is ignored in the Established state.

In response to any other event, event [Events 9, 12-13, 20-22] the
local system:
- sends a NOTIFICATION message with Error Code Finite
State Machine Error,
- sets IdleHoldtimer = 2**(ConnectRetryCnt)*60

- increments ConnectRetryCnt by 1, deletes all routes associated with this connection,
- sets resets the connect retry timer (sets to zero, zero)
- releases all BGP resources,
- drops the TCP connection,
- releases all BGP resources increments the ConnectRetryCnt (connect retry count)
by 1,
- goes to IdleHoldstate, optionally performs peer oscillation damping, and

Expiration Date July 2002 [Page 43]

RFC DRAFT January 2002
- deletes all routes. changes its state to Idle.

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 valid-
ity as specified in Section 6.3.

If an optional non-transitive attribute is unrecognized, it is
quietly qui-
etly 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 propaga-
tion to other BGP speakers.

If the UPDATE message contains a non-empty WITHDRAWN ROUTES field,

Expiration Date September 2003 [Page 62]

RFC DRAFT March 2003

the previously advertised routes whose destinations (expressed as IP
prefixes) contained in this field shall SHALL be removed from the Adj-RIB-
In. This BGP speaker shall SHALL run its Decision Process since the
previously previ-
ously advertised route is no longer available for use.

If the UPDATE message contains a feasible route, the Adj-RIB-In will
be updated with this route as follows: if the NLRI of the new route
is identical to the one of the route currently stored in the Adj-RIB-
In, then the new route shall SHALL replace the older route in the Adj-RIB-
In, thus implicitly withdrawing the older route from service.
Otherwise, Other-
wise, if the Adj-RIB-In has no route with NLRI identical to the new
route, the new route shall SHALL be placed in the Adj-RIB-In.

Once the BGP speaker updates the Adj-RIB-In, the speaker shall SHALL run
its Decision Process.

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-RIBs-In. The output of the Decision
Process Pro-
cess 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.

Expiration Date July 2002 [Page 44]

RFC DRAFT January 2002 Adj-RIB-Out
according to policy.

The selection process is formalized by defining a function that takes
the attribute of a given route as an argument and returns either (a)
a non-negative integer denoting the degree of preference for the
route, or (b) a value denoting that this route is ineligible to be
installed in LocRib and will be excluded from the next phase of route
selection.

The function that calculates the degree of preference for a given
route shall not 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 individ-
ual application of the degree of preference function to each feasible
route, followed by the choice of the one with the highest degree of
preference.

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

- selection of routes to be used locally by the speaker

- selection of routes to be advertised to other BGP peers

Expiration Date September 2003 [Page 63]

RFC DRAFT March 2003

- route aggregation and route information reduction

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

a) Phase 1 is responsible for calculating the degree of preference
for each route received from a peer.

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 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,
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 is invoked whenever the local BGP
speaker receives from a peer an UPDATE message that advertises a new
route, a replacement route, or withdrawn routes.

Expiration Date July 2002 [Page 45]

RFC DRAFT January 2002

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 locks an Adj-RIB-In prior to operating
on any route contained within it, and shall unlock unlocks 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 determines a degree of preference as follows:

If the route is learned from an internal peer, either the value of
the LOCAL_PREF attribute shall be is taken as the degree of preference, or
the local system may compute computes the degree of preference of the route
based on preconfigured policy information. Note that the latter
(computing the degree of preference based on preconfigured policy
information) may result in formation of persistent routing loops.

If the route is learned from an external peer, then the local BGP
speaker computes the degree of preference based on preconfigured
policy information. If the return value indicates that the route
is ineligible, the route may not MAY NOT serve as an input to the next

Expiration Date September 2003 [Page 64]

RFC DRAFT March 2003

phase of route selection; otherwise the return value is used as
the LOCAL_PREF value in any IBGP readvertisement.

The exact nature of this policy information and the computation
involved is a local matter.

9.1.2 Phase 2: Route Selection

The Phase 2 decision function shall be is 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 considers all routes
that are eligible in the Adj-RIBs-In.

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

If the NEXT_HOP attribute of a BGP route depicts an address that is
not resolvable, or it would become unresolvable if the route was
installed in the routing table was
installed in the routing table the BGP route MUST be excluded from
the Phase 2 decision function.

If the AS_PATH attribute of a BGP route contains an AS loop, the BGP
route should be excluded from the Phase 2 decision function. AS loop
detection is done by scanning the full AS path (as specified in the
AS_PATH attribute), and checking that the autonomous system number of
the local system does not appear in the AS path. Operations of a BGP
speaker that is configured to accept routes with its own autonomous
system number in the AS path are outside the scope of this document.

It is critical that routers BGP speakers within an AS do not make conflicting

Expiration Date July 2002 [Page 46]

RFC DRAFT January 2002
decisions regarding route selection that would cause forwarding loops
to occur.

For each set of destinations for which a feasible route exists in the
Adj-RIBs-In, the local BGP speaker shall identify identifies 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 spec-
ified in 9.1.2.2.

Expiration Date September 2003 [Page 65]

RFC DRAFT March 2003

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. If When the new BGP route is installed in the Routing
Table (as a result of the local policy decision), Rout-
ing Table, care must be taken to ensure that invalid BGP existing routes to the
same destination that are now considered invalid are removed from the
Routing Table. Whether or not the new BGP route replaces an
already existing
non-BGP route in the routing table Routing Table depends on the policy configured
on the BGP speaker.

The local speaker MUST determine the immediate next hop to the next-hop address depicted by from
the NEXT_HOP attribute of the selected route by
performing a best matching route lookup in the Routing Table and
selecting one of the possible paths (if multiple best paths to the
same prefix are available). (see Section 5.1.3). If the route to the address depicted by
the NEXT_HOP attribute changes such that
either the immediate next hop or the IGP cost to the NEXT_HOP (if (where
the NEXT_HOP is resolved through an IGP route) changes, route selection should Phase 2 Route
Selection MUST be recalculated as
specified above. performed again.

Notice that even though BGP routes do not have to be installed in the
Routing Table with the immediate next hop(s), implementations must MUST
take care that before any packets are forwarded along a BGP route,
its associated NEXT_HOP address is resolved to the immediate
(directly connected) next-hop address and this address (or multiple
addresses) is finally used for actual packet forwarding.

Unresolvable routes SHALL be removed from the Loc-RIB and the routing
table. However, corresponding unresolvable routes SHOULD be kept in
the Adj-RIBs-In.

Expiration Date July 2002 [Page 47]

RFC DRAFT January 2002 Adj-RIBs-In (in case they become resolvable).

9.1.2.1 Route Resolvability Condition

As indicated in Section 9.1.2, BGP routers should speakers SHOULD exclude
unresolvable unresolv-
able routes from the Phase 2 decision. This ensures that only valid
routes are installed in Loc-RIB and the Routing Table.

The route resolvability condition is defined as follows.

1. A route Rte1, referencing only the intermediate network
address, is considered resolvable if the Routing Table contains at
least one resolvable route Rte2 that matches Rte1's intermediate
network address and is not recursively resolved (directly or
indirectly) indi-
rectly) through Rte1. If multiple matching routes are available,
only the longest matching route should SHOULD be considered.

2. Routes referencing interfaces (with or without intermediate
addresses) are considered resolvable if the state of the
referenced refer-
enced interface is up and IP processing is enabled on this
interface. inter-
face.

Expiration Date September 2003 [Page 66]

RFC DRAFT March 2003

BGP routes do not refer to interfaces, but can be resolved through
the routes in the Routing Table that can be of both types. types (those that
specify interfaces or those that do not). IGP routes and routes to
directly connected networks are expected to specify the outbound
interface. Static routes can specify the outbound interface, or the
intermediate address, or both.

Note that a BGP route is considered unresolvable not only in
situations situa-
tions where the router's BGP speaker's Routing Table contains no route
matching match-
ing the BGP route's NEXT_HOP. Mutually recursive routes (routes
resolving each other or themselves), also fail the resolvability
check.

It is also important that implementations do not consider feasible
routes that would become unresolvable if they were installed in the
Routing Table even if their NEXT_HOPs are resolvable using the
current cur-
rent contents of the Routing Table (an example of such routes would
be mutually recursive routes). This check ensures that a BGP speaker
does not install in the Routing Table routes that will be removed and
not used by the speaker. Therefore, in addition to local Routing
Table stability, this check also improves behavior of the protocol in
the network.

Whenever a BGP speaker identifies a route that fails the
resolvability resolvabil-
ity check because of mutual recursion, an error message
should SHOULD be
logged.

Expiration Date July 2002 [Page 48]

RFC DRAFT January 2002

9.1.2.2 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 routes with the
same degrees of preference, both those received from internal peers,
and those received from external peers.

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, and follow the same route selection
algorithm.

The tie-breaking algorithm begins by considering all equally
preferable prefer-
able routes to the same destination, and then selects routes to be
removed from consideration. The algorithm terminates as soon as only
one route remains in consideration. The criteria must MUST be applied in

Expiration Date September 2003 [Page 67]

RFC DRAFT March 2003

the order specified.

Several of the criteria are described using pseudo-code. Note that
the pseudo-code shown was chosen for clarity, not efficiency. It is
not intended to specify any particular implementation. BGP
implementations implemen-
tations MAY use any algorithm which produces the same results as
those described here.

a) Remove from consideration all routes which are not tied for
having the smallest number of AS numbers present in their AS_PATH
attributes. Note, that when counting this number, an AS_SET counts
as 1, no matter how many ASs are in the set, and that, if the
implementation supports [13], then AS numbers present in segments
of type AS_CONFED_SEQUENCE or AS_CONFED_SET are not included in
the count of AS numbers present in the AS_PATH. set.

b) Remove from consideration all routes which are not tied for
having the lowest Origin number in their Origin attribute.

c) Remove from consideration routes with less-preferred less-preferred
MULTI_EXIT_DISC attributes. MULTI_EXIT_DISC is only comparable
between routes learned from the same neighboring AS (the neighbor-
ing AS is determined from the AS_PATH attribute). Routes which do
not have the MULTI_EXIT_DISC attribute are considered to have the
lowest possible MULTI_EXIT_DISC value.

This is also described in the following procedure:

for m = all routes still under consideration
for n = all routes still under consideration
if (neighborAS(m) == neighborAS(n)) and (MED(n) < MED(m))
remove route m from consideration

In the pseudo-code above, MED(n) is a function which returns the
value of route n's MULTI_EXIT_DISC attribute. If route n has no
MULTI_EXIT_DISC attribute, the function returns the lowest possi-
ble MULTI_EXIT_DISC value, i.e. 0.

Similarly, neighborAS(n) is a function which returns the neighbor
AS from which the route was received. If the route is learned via
IBGP, and the other IBGP speaker didn't originate the route, it is
the neighbor AS from which the other IBGP speaker learned the
route. If the route is learned via IBGP, and the other IBGP
speaker originated the route, it is the local AS.

If a MULTI_EXIT_DISC attribute is removed before re-advertising a
route into IBGP, then comparison based on the received EBGP
MULTI_EXIT_DISC attribute MAY still be performed. If an implemen-
tation chooses to remove MULTI_EXIT_DISC, then the optional com-
parison on MULTI_EXIT_DISC if performed at all MUST be performed
only among EBGP learned routes. The best EBGP learned route may

Expiration Date September 2003 [Page 68]

RFC DRAFT March 2003

then be compared with IBGP learned routes after the removal of the
MULTI_EXIT_DISC attributes. attribute. If MULTI_EXIT_DISC is only comparable
between routes learned removed from a
subset of EBGP learned routes and the same neighboring AS. Routes which
do selected "best" EBGP learned
route will not have the MULTI_EXIT_DISC attribute are considered to have removed, then the lowest possible
MULTI_EXIT_DISC value.

This is also described must be used in the following procedure:

for m = all routes still under consideration

Expiration Date July 2002 [Page 49]

RFC DRAFT January 2002

for n = all comparison with IBGP learned
routes. For IBGP learned routes still under consideration
if (neighborAS(m) == neighborAS(n)) and (MED(n) < MED(m))
remove route m from consideration

In the pseudo-code above, MED(n) is a function MULTI_EXIT_DISC MUST be used
in route comparisons which returns reach this step in the decision pro-
cess. Including the
value of route n's MULTI_EXIT_DISC attribute. If of an EBGP learned route n has no
MULTI_EXIT_DISC attribute, in
the function returns comparison with an IBGP learned route, then removing the lowest
possible
MULTI_EXIT_DISC value, i.e. 0.

Similarly, neighborAS(n) is a function which returns the neighbor
AS from which atribute and advertising the route was received. has been proven
to cause route loops.

d) If at least one of the candidate routes was received from an
external peer in a neighboring autonomous system, via EBGP,
remove from consideration all routes which were received from internal peers. via IBGP.

e) Remove from consideration any routes with less-preferred
interior inte-
rior cost. The interior cost of a route is determined by
calculating calcu-
lating the metric to the next hop NEXT_HOP for the route using the Routing
Table. If the next NEXT_HOP hop for a route is reachable, but no cost
can be determined, then this step should be skipped (equivalently,
consider all routes to have equal costs).

This is also described in the following procedure.

for m = all routes still under consideration
for n = all routes in still under consideration
if (cost(n) is better lower than cost(m))
remove m from consideration

In the pseudo-code above, cost(n) is a function which returns the
cost of the path (interior distance) to the address given in the
NEXT_HOP attribute of the route.

f) Remove from consideration all routes other than the route that
was advertised by the BGP speaker whose BGP Identifier has the
lowest value.

g) Prefer the route received from the lowest neighbor peer address.

9.1.3 Phase 3: Route Dissemination

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

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

Expiration Date July 2002 September 2003 [Page 50] 69]

RFC DRAFT January 2002 March 2003

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 estab-
lished

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 is
blocked from running while the Phase 2 decision function is in process. pro-
cess.

All routes in the Loc-RIB shall be are processed into Adj-RIBs-Out according
to configured policy. This policy may MAY exclude a route in the Loc-RIB
from being installed in a particular Adj-RIB-Out. A route shall not SHALL NOT
be installed in the Adj-Rib-Out unless the destination and NEXT_HOP
described by this route may be forwarded appropriately by the Routing
Table. If a route in Loc-RIB is excluded from a particular Adj-RIB-Out Adj-RIB-
Out the previously advertised route in that Adj-RIB-Out must MUST be withdrawn with-
drawn from service by means of an UPDATE message (see 9.2).

Route aggregation and information reduction techniques (see 9.2.2.1)
may optionally be applied.

Any local policy which results in routes being added to an Adj-RIB-
Out without also being added to the local BGP speaker's forwarding
table, is outside the scope of this document.

When the updating of the Adj-RIBs-Out and the Routing Table is
complete, com-
plete, the local BGP speaker shall run runs the Update-Send process of 9.2.

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 mul-
tiple routes. Since BGP encodes NLRI using IP prefixes, overlap 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); shorter pre-
fix); similarly, a route describing a larger set of
destinations (a shorter prefix) destinations is
said to be less specific than a route describing a smaller set of destinations (a longer prefix).
destinations.

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

Expiration Date September 2003 [Page 70]

RFC DRAFT March 2003

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

Expiration Date July 2002 [Page 51]

RFC DRAFT January 2002

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

When overlapping routes are present in the same Adj-RIB-In, the more
specific route shall take takes 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 desti-
nations described by the overlap will still be reachable using the
less specific route.

If a BGP speaker receives overlapping routes, the Decision Process
MUST consider both routes based on the configured acceptance policy.
If both a less and a more specific route are accepted, then the
Decision Deci-
sion Process MUST either install both the less and the more specific
routes or it MUST aggregate the two routes and install the aggregated
route, provided that both routes have the same value of the NEXT_HOP
attribute.

If a BGP speaker chooses to aggregate, then it MUST SHOULD either include
all AS used to form the aggreagate in an AS_SET or add the
ATOMIC_AGGREGATE attribute to the route. This attribute is now pri-
marily informational. With the elimination of IP routing protocols
that do not support classless routing and the elimination of router
and host implementations that do not support classless routing, there
is no longer a need to deaggregate. Routes SHOULD NOT be de-aggre-
gated. A route that carries ATOMIC_AGGREGATE attribute can not in particular
MUST 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.

9.2 Update-Send Process

The Update-Send process is responsible for advertising UPDATE
messages mes-
sages 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.

When a BGP speaker receives an UPDATE message from an internal peer,

Expiration Date September 2003 [Page 71]

RFC DRAFT March 2003

the receiving BGP speaker shall not SHALL NOT re-distribute the routing
information infor-
mation contained in that UPDATE message to other internal peers,
unless the speaker acts as a BGP Route Reflector [11]. [RFC2796].

As part of Phase 3 of the route selection process, the BGP speaker
has updated its Adj-RIBs-Out. All newly installed routes and all
newly unfeasible routes for which there is no replacement route shall SHALL
be advertised to its peers by means of an UPDATE message.

Expiration Date July 2002 [Page 52]

RFC DRAFT January 2002

A BGP speaker should not SHOULT NOT advertise a given feasible BGP route from
its Adj-RIB-Out if it would produce an UPDATE message containing the
same BGP route as was previously advertised.

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

If due to the limits on the maximum size of an UPDATE message (see
Section 4) a single route doesn't fit into the message, the BGP
speaker MUST not advertise the route to its peers and MAY choose to
log an error locally.

9.2.1 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.

9.2.1.1 Frequency of Route Advertisement

The parameter MinRouteAdvertisementInterval determines the minimum
amount of time that must elapse between advertisement and/or with-
drawal of routes to a particular destination from by a single BGP speaker. speaker to a
peer. 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 by a single BGP speaker to a peer that advertise
feasible routes and/or withdrawal of unfeasible routes to some common
set of destinations received from
external peers must MUST be separated by at least

Expiration Date September 2003 [Page 72]

RFC DRAFT March 2003

MinRouteAdvertisementInterval. Clearly, this can only be achieved
precisely by keeping a separate timer for each common set of
destinations. destina-
tions. This would be unwarranted overhead. Any technique which
ensures that the interval between two UPDATE messages sent from a
single BGP
speaker to a peer that advertise feasible routes and/or withdrawal of
unfeasible routes to some common set of destinations received from external peers will be at least
MinRouteAdvertisementInterval, and will also ensure a constant upper
bound on the interval is acceptable.

Since fast convergence is needed within an autonomous system, this
procedure does not apply either
(a) the MinRouteAdvertisementInterval used for routes received from other internal
peers. To avoid long-lived black holes, peers SHOULD
be shorter than the MinRouteAdvertisementInterval used for external
peers, or (b) the procedure does not describe in this section SHOULD NOT apply
to the explicit withdrawal of unfeasible routes (that is,
for routes
whose destinations (expressed as IP prefixes) are listed in the
WITHDRAWN ROUTES field of an UPDATE message). sent to internal peers.

This procedure does not limit the rate of route selection, but only

Expiration Date July 2002 [Page 53]

RFC DRAFT January 2002
the rate of route advertisement. If new routes are selected multiple
times while awaiting the expiration of MinRouteAdvertisementInterval,
the last route selected shall SHALL be advertised at the end of
MinRouteAdvertisementInterval. MinRouteAd-
vertisementInterval.

9.2.1.2 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.

9.2.1.3 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.

9.2.2 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.

9.2.2.1 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

Expiration Date July 2002 September 2003 [Page 54] 73]

RFC DRAFT January 2002 March 2003

methods:

a) Network Layer Reachability Information (NLRI):

Destination IP addresses can be represented as IP address
prefixes. pre-
fixes. 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 algo-
rithm described in 9.2.2.2. 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 con-
stituent autonomous systems. AS_SETs provide sufficient
information informa-
tion 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 infor-
mation and can distinguish individual paths to destinations.

9.2.2.2 Aggregating Routing Information

Aggregation is the process of combining the characteristics of
several sev-
eral different routes in such a way that a single route can be
advertised. adver-
tised. 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 different MULTI_EXIT_DISC attribute SHALL NOT be aggregated
unless the corresponding attributes of each route are identical:
MULTI_EXIT_DISC, NEXT_HOP.
aggregated.

Expiration Date July 2002 September 2003 [Page 55] 74]

RFC DRAFT January 2002

If the aggregation occurs as part of the update process, routes with
different NEXT_HOP values can be aggregated when announced through an
external BGP session. March 2003

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

NEXT_HOP:
When aggregating routes that have different NEXT_HOP attribute,
the NEXT_HOP attribute of the aggregated route SHALL identify
an interface on the BGP speaker that performs the aggregation.

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

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 SHALL satisfy all of the
following conditions:

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

- all tuples of type AS_SET in the aggregated AS_PATH shall 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 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 of type AS_SET with the same value shall SHALL appear
more than once in the aggregated AS_PATH.

Expiration Date September 2003 [Page 75]

RFC DRAFT March 2003

- Multiple tuples of type AS_SEQUENCE with the same value
may appear in the aggregated AS_PATH only when adjacent to
another tuple of the same type and value.

Expiration Date July 2002 [Page 56]

RFC DRAFT January 2002

An implementation may choose any algorithm which conforms to
these rules. At a minimum a conformant implementation shall 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 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.

- for each pair of adjacent tuples in the aggregated
AS_PATH, if both tuples have the same type, merge them
together, as long as doing so will not cause a segment with
length greater than 255 to be generated.

Appendix 6, section 6.8 F, Section F.6 presents another algorithm that satisfies satis-
fies the conditions and allows for more complex policy configurations. configu-
rations.

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

AGGREGATOR: All
Any AGGREGATOR attributes of all from the routes to be aggregated should MUST
NOT be ignored. included in the aggregated route. The BGP speaker performing per-
forming the route aggregation may MAY attach a new AGGREGATOR
attribute (see Section 5.1.7).

9.3 Route Selection Criteria

Generally speaking, additional rules for comparing routes among

Expiration Date September 2003 [Page 76]

RFC DRAFT March 2003

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 can not be viewed as better than
any other route (provided that the speaker is configured to accept
such routes). If such a route were ever used, a routing loop could
result (see Section 6.3).
result.

- In order to achieve successful distributed operation, only
routes with a likelihood of stability can be chosen. Thus, an AS
must
SHOULD avoid using unstable routes, and it must not SHOULD NOT make rapid

Expiration Date July 2002 [Page 57]

RFC DRAFT January 2002
spontaneous changes to its choice of route. Quantifying the terms
"unstable" and "rapid" in the previous sentence will require
experience, expe-
rience, but the principle is clear.

Care must principle is clear.

Care must be taken to ensure that BGP speakers in the same AS do not
make inconsistent decisions.

9.4 Originating BGP routes

A BGP speaker may originate BGP routes by injecting routing informa-
tion acquired by some other means (e.g. via an IGP) into BGP. A BGP
speaker that originates BGP routes assigns 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 update process (see Sec-
tion 9.2). 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.

10 BGP Timers

BGP employs five timers: ConnectRetry (see Section 8), Hold Time (see
Section 4.2), KeepAlive (see Section 8), MinASOriginationInterval
(see Section 9.2.1.2), and MinRouteAdvertisementInterval (see Section
9.2.1.1).

The suggested default value for the ConnectRetry timer is 120 sec-
onds.

The suggested default value for the Hold Time is 90 seconds.

The suggested default value for the KeepAlive timer is 1/3 of the
Hold Time.

Expiration Date September 2003 [Page 77]

RFC DRAFT March 2003

The suggested default value for the MinASOriginationInterval is 15
seconds.

The suggested default value for the MinRouteAdvertisementInterval is
30 seconds.

An implementation of BGP MUST allow the Hold Time timer to be taken config-
urable on a per peer basis, and MAY allow the other timers to ensure be con-
figurable.

To minimize the likelihood that BGP speakers in the same AS do
not make inconsistent decisions.

9.4 Originating distribution of BGP routes

A messages by a
given BGP speaker may originate BGP routes by injecting routing
information acquired by some other means (e.g. via an IGP) into BGP. will contain peaks, jitter SHOULD be applied to the
timers associated with MinASOriginationInterval, KeepAlive, Min-
RouteAdvertisementInterval, and ConnectRetry. A given BGP speaker that originates BGP routes shall assign MAY
apply the degree of
preference same jitter to each of 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 quantities regardless of the update process
(see Section 9.2). The decision whether
destinations to distribute non-BGP
acquired routes within an AS via BGP or which the updates are being sent; that is, jitter
need not depends be configured on a "per peer" basis.

The suggested default amount of jitter SHALL be determined by multi-
plying the
environment within base value of the AS (e.g. type appropriate timer by a random factor
which is uniformly distributed in the range from 0.75 to 1.0. A new
random value SHOULD be picked each time the timer is set. The range
of IGP) and should the jitter random value MAY be controlled
via configuration. configurable.

Appendix 1. A. Comparison with RFC1771

There are numerous editorial changes (too many to list here).

The following list the technical changes:

Changes to reflect the usages of such features as TCP MD5 [10],
[RFC2385], BGP Route Reflectors [11], [RFC2796], BGP Confederations [13],
[RFC3065], and BGP Route Refresh [12]. [RFC2918].

Clarification on the use of the BGP Identifier in the AGGREGATOR
attribute.

Procedures for imposing an upper bound on the number of prefixes
that a BGP speaker would accept from a peer.

The ability of a BGP speaker to include more than one instance of
its own AS in the AS_PATH attribute for the purpose of inter-AS
traffic engineering.

Clarifications on the various types of NEXT_HOPs.

Expiration Date July 2002 September 2003 [Page 58] 78]

RFC DRAFT January 2002 March 2003

Clarifications to the use of the ATOMIC_AGGREGATE attribute.

The relationship between the immediate next hop, and the next hop
as specified in the NEXT_HOP path attribute.

Clarifications on the tie-breaking procedures. tie-breaking procedures.

Clarifications on the frequency of route advertisements.

Optional Parameter Type 1 (Authentication Information) has been
deprecated.

UPDATE Message Error subcode 7 (AS Routing Loop) has been depre-
cated.

OPEN Message Error subcode 5 (Authentication Failure) has been
deprecated.

Use of the Marker field for authentication has been deprecated.

Use of TCP MD5 [RFC2385] for authentication is mandatory.

Appendix 2. B. Comparison with RFC1267

All the changes listed in Appendix 1, A, plus the following.

BGP-4 is capable of operating in an environment where a set of
reachable reach-
able destinations may be expressed via a single IP prefix. The
concept con-
cept 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 nego-
tiated on a per-connection basis. Hold Times of zero are now
supported. sup-
ported.

Expiration Date September 2003 [Page 79]

RFC DRAFT March 2003

Appendix 3. C. Comparison with RFC 1163

All of the changes listed in Appendices 1 A and 2, B, 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 recov-
ering 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.

Expiration Date July 2002 [Page 59]

RFC DRAFT January 2002

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

Appendix 4. D. Comparison with RFC 1105

All of the changes listed in Appendices 1, 2 A, B and 3, C, plus the
following. follow-
ing.

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.

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. signifi-
cantly. 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

Expiration Date September 2003 [Page 80]

RFC DRAFT March 2003

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. E. TCP options that may be used with BGP

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

Expiration Date July 2002 [Page 60]

RFC DRAFT January 2002
receiver.

If a local system TCP user interface supports setting precedence of the DSCP
field [RFC2474] for TCP connection, connections, then the BGP transport TCP connection should used by
BGP SHOULD be opened with precedence bits 0-2 of the DSCP field set to Internetwork Control (110) value (see also
[6]).

A local system may protect its BGP sessions by using the TCP MD5
Signature Option [10]. 110
(binary).

Appendix 6. F. Implementation Recommendations

This section presents some implementation recommendations.

6.1

Appendix F.1 Multiple Networks Per Message

The BGP protocol allows for multiple address prefixes with the same
path attributes to be specified in one message. Making use of this
capability is highly recommended. With one address prefix 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
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 address prefixes per
a path attribute set from a routing table that is not organized on a
per path attribute set basis is to build many messages as the routing
table is scanned. As each address prefix is processed, a message for
the associated set of path attributes is allocated, if it does not
exist, and the new address prefix is added to it. If such a message
exists, the new address prefix is just appended to it. If the message
lacks the space to hold the new address prefix, it is transmitted, a

Expiration Date September 2003 [Page 81]

RFC DRAFT March 2003

new message is allocated, and the new address prefix 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 the destinations covered by the
address prefixes share a common set of path attributes making it
possible pos-
sible to send many address prefixes in one 4096-byte message.

When peering with a BGP implementation that does not compress
multiple multi-
ple address prefixes 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

Expiration Date July 2002 [Page 61]

RFC DRAFT January 2002 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 disadvan-
tage 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 multi-
ple 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 better method would "hang" the routing information process while trying
be 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
received.

6.3 all received messages before sending updates.

Appendix F.2 Reducing route flapping

To avoid excessive route flapping a BGP speaker which needs to
withdraw with-
draw a destination and send an update about a more specific or less
specific route SHOULD 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 1/3 of the Hold Time. The suggested value

Expiration Date July 2002 [Page 62]

RFC DRAFT January 2002

for the MinASOriginationInterval is 15 seconds. The suggested value
for the MinRouteAdvertisementInterval is 30 seconds.

An implementation of BGP MUST allow the Hold Time timer to be
configurable, and MAY allow the other timers to be configurable.

6.5

Appendix F.3 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 differ-
ent 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

Appendix F.4 AS_SET sorting

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

6.7

Expiration Date September 2003 [Page 82]

RFC DRAFT March 2003

Appendix F.5 Control over version negotiation

Since BGP-4 is capable of carrying aggregated routes which cannot can not 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

Appendix F.6 Complex AS_PATH aggregation

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

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

Expiration Date July 2002 [Page 63]

RFC DRAFT January 2002
same if their corresponding <type, value> tuples are the same.

The algorithm to aggregate two AS_PATH attributes works as
follows: fol-
lows:

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 seg-
ment that consists of the intervening ASs from both AS_PATH
attributes; this segment is then placed in between the two
consecutive con-
secutive 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 consec-
utive ASs identified in (a) of the aggregated attribute.

Expiration Date September 2003 [Page 83]

RFC DRAFT March 2003

c) For each pair of adjacent tuples in the aggregated AS_PATH,
if both tuples have the same type, merge them together, as long
as doing so will not cause a segment with length greater than
255 to be generated.

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 SHOULD
be removed from the aggregated AS_PATH attribute.

Security Considerations

BGP supports the ability to authenticate BGP messages by using BGP
authentication.

The authentication could be done on a per peer basis.
In addition, mechanism that an implementation of BGP supports the ability to authenticate its data stream
by using [10]. This MUST sup-
port is specified in [RFC2385]. The authentication provided by this
mechanism could be done on a per peer basis.
Finally,

Security issues with BGP could also use IPSec routing information dissemination are dis-
cussed in [XXX].

IANA Considerations

All extensions to authenticate its data stream.
Among this protocol, including new message types and Path
Attributes MUST only be made using the mechanisms mentioned Standards Action process
defined in this paragraph, [10] is [RFC2434].

Normative References

[RFC791] Postel, J., "Internet Protocol - DARPA Internet Program Pro-
tocol Specification", RFC791, September 1981.

[RFC793] Postel, J., "Transmission Control Protocol - DARPA Internet
Program Protocol Specification", RFC793, September 1981.

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

[RFC2385] Heffernan, A., "Protection of BGP Sessions via the most
widely deployed. TCP MD5
Signature Option", RFC2385, August 1998.

[RFC2434] Narten, T., Alvestrand, H., "Guidelines for Writing an IANA
Considerations Section in RFCs", RFC2434, October 1998

Expiration Date July 2002 September 2003 [Page 64] 84]

RFC DRAFT January 2002 March 2003

[RFC2474] Nichols, K., et al.,"Definition of the Differentiated Ser-
vices Field (DS Field) in the IPv4 and IPv6 Headers", RFC2474, Decem-
ber 1998

Non-normative References

[1]

[RFC904] Mills, D., "Exterior Gateway Protocol Formal Specification",
RFC904, April 1984.

[2]

[RFC1092] Rekhter, Y., "EGP and Policy Based Routing in the New
NSFNET Backbone", RFC1092, February 1989.

[3]

[RFC1093] Braun, H-W., "The NSFNET Routing Architecture", RFC1093,
February 1989.

[4] Postel, J., "Transmission Control Protocol - DARPA Internet
Program Protocol Specification", RFC793, September 1981.

[5]

[RFC1772] Rekhter, Y., and P. Gross, "Application of the Border Gateway Gate-
way Protocol in the Internet", RFC1772, March 1995.

[6] Postel, J., "Internet Protocol - DARPA Internet Program Protocol
Specification", RFC791, September 1981.

[7] "Information Processing Systems - Telecommunications and
Information Exchange between Systems - Protocol

[RFC1518] Rekhter, Y., Li, T., "An Architecture for Exchange of
Inter-domain Routeing Information among Intermediate Systems to
Support Forwarding of ISO 8473 PDUs", ISO/IEC IS10747, 1993

[8] IP Address Allo-
cation with CIDR", RFC 1518, September 1993.

[RFC1519] Fuller, V., Li, T., Yu, J., and Varadhan, K., ""Classless Inter-
Domain
Inter-Domain Routing (CIDR): an Address Assignment and Aggregation
Strategy", RFC1519, September 1993.

[9] Rekhter, Y.,

[RFC1997] R. Chandra, P. Traina, T. Li, T., "An Architecture for IP Address Allocation
with CIDR", "BGP Communities Attribute",
RFC 1518, September 1993.

[10] Heffernan, A., "Protection of BGP Sessions via the TCP MD5
Signature Option", RFC2385, 1997, August 1996.

[RFC2439] C. Villamizar, R. Chandra, R. Govindan, "BGP Route Flap
Damping", RFC2439, November 1998.

[11]

[RFC2796] Bates, T., Chandra, R., Chen, E., "BGP Route Reflection -
An Alternative to Full Mesh IBGP", RFC2796, April 2000.

[12]

[RFC2842] R. Chandra, J. Scudder, "Capabilities Advertisement with
BGP-4", RFC2842.

[RFC2858] T. Bates, R. Chandra, D. Katz, Y. Rekhter, "Multiprotocol
Extensions for BGP-4", RFC2858.

[RFC2918] Chen, E., "Route Refresh Capability for BGP-4", RFC2918,
September 2000.

[13]

[RFC3065] Traina, P, McPherson, D., Scudder, J., "Autonomous System
Confederations for BGP", RFC3065, February 2001.

Expiration Date July 2002 September 2003 [Page 65] 85]

RFC DRAFT January 2002 March 2003

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

[XXX] Murphy, S., "BGP Security Vulnerabilities Analysis", draft-
ietf-idr-bgp-vuln-00.txt, work in progress

Editors' Addresses

Yakov Rekhter
Juniper Networks
1194 N. Mathilda Avenue
Sunnyvale, CA 94089
email: yakov@juniper.net

Tony Li
Procket Networks
1100 Cadillac Ct.
Milpitas, CA 95035
Email: Networks, Inc.
email: tli@procket.com

Susan Hares

NextHop Technologies, Inc.
email: skh@nexthop.com

Expiration Date July 2002 September 2003 [Page 66] 86]

----