| 1 |
<pre>Network Working Group K. Lougheed |
| 2 |
Request for Comments: 1105 cisco Systems |
| 3 |
Y. Rekhter |
| 4 |
T.J. Watson Research Center, IBM Corp. |
| 5 |
June 1989 |
| 6 |
|
| 7 |
<span class="h1">A Border Gateway Protocol (BGP)</span> |
| 8 |
|
| 9 |
Status of this Memo |
| 10 |
|
| 11 |
This RFC outlines a specific approach for the exchange of network |
| 12 |
reachability information between Autonomous Systems. |
| 13 |
|
| 14 |
At the time of this writing, the Border Gateway Protocol |
| 15 |
implementations exist for cisco routers as well as for the NSFNET |
| 16 |
Nodal Switching Systems. A public domain version for "gated" is |
| 17 |
currently being implemented. |
| 18 |
|
| 19 |
Distribution of this memo is unlimited. |
| 20 |
|
| 21 |
<span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Introduction</span> |
| 22 |
|
| 23 |
The Border Gateway Protocol (BGP) is an inter-autonomous system |
| 24 |
routing protocol. It is built on experience gained with EGP as |
| 25 |
defined in <a href="./rfc904">RFC 904</a> [<a href="#ref-1" title=""Exterior Gateway Protocol Formal Specification"">1</a>] and EGP usage in the NSFNET Backbone as |
| 26 |
described in <a href="./rfc1092">RFC 1092</a> [<a href="#ref-2" title=""EGP and Policy Based Routing in the New NSFNET Backbone"">2</a>] and <a href="./rfc1093">RFC 1093</a> [<a href="#ref-3" title=""The NSFNET Routing Architecture"">3</a>]. |
| 27 |
|
| 28 |
The primary function of a BGP speaking system is to exchange network |
| 29 |
reachability information with other BGP systems. This network |
| 30 |
reachability information includes information on the autonomous |
| 31 |
systems (AS's) that traffic must transit to reach these networks. |
| 32 |
This information is sufficient to construct a graph of AS |
| 33 |
connectivity from which routing loops may be pruned and policy |
| 34 |
decisions at an AS level may be enforced. |
| 35 |
|
| 36 |
BGP runs over a reliable transport level protocol. This eliminates |
| 37 |
the need to implement explicit update fragmentation, retransmission, |
| 38 |
acknowledgement, and sequencing. Any authentication scheme used by |
| 39 |
the transport protocol may be used in addition to BGP's own |
| 40 |
authentication mechanisms. |
| 41 |
|
| 42 |
The initial BGP implementation is based on TCP [<a href="#ref-4" title=""Transmission Control Protocol - DARPA Internet Program Protocol Specification"">4</a>], however any |
| 43 |
reliable transport may be used. A message passing protocol such as |
| 44 |
VMTP [<a href="#ref-5" title=""VMTP: Versatile Message Transaction Protocol"">5</a>] might be more natural for BGP. TCP will be used, however, |
| 45 |
since it is present in virtually all commercial routers and hosts. |
| 46 |
In the following descriptions the phrase "transport protocol |
| 47 |
connection" can be understood to refer to a TCP connection. BGP uses |
| 48 |
TCP port 179 for establishing its connections. |
| 49 |
|
| 50 |
|
| 51 |
|
| 52 |
<span class="grey">Lougheed & Rekhter [Page 1]</span></pre> |
| 53 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-2" ></span> |
| 54 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 55 |
|
| 56 |
|
| 57 |
<span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Summary of Operation</span> |
| 58 |
|
| 59 |
Two hosts form a transport protocol connection between one another. |
| 60 |
They exchange messages to open and confirm the connection parameters. |
| 61 |
The initial data flow is the entire BGP routing table. Incremental |
| 62 |
updates are sent as the routing tables change. Keepalive messages |
| 63 |
are sent periodically to ensure the liveness of the connection. |
| 64 |
Notification messages are sent in response to errors or special |
| 65 |
conditions. If a connection encounters an error condition, a |
| 66 |
notification message is sent and the connection is optionally closed. |
| 67 |
|
| 68 |
The hosts executing the Border Gateway Protocol need not be routers. |
| 69 |
A non-routing host could exchange routing information with routers |
| 70 |
via EGP or even an interior routing protocol. That non-routing host |
| 71 |
could then use BGP to exchange routing information with a border |
| 72 |
gateway in another autonomous system. The implications and |
| 73 |
applications of this architecture are for further study. |
| 74 |
|
| 75 |
If a particular AS has more than one BGP gateway, then all these |
| 76 |
gateways should have a consistent view of routing. A consistent view |
| 77 |
of the interior routes of the autonomous system is provided by the |
| 78 |
intra-AS routing protocol. A consistent view of the routes exterior |
| 79 |
to the AS may be provided in a variety of ways. One way is to use |
| 80 |
the BGP protocol to exchange routing information between the BGP |
| 81 |
gateways within a single AS. In this case, in order to maintain |
| 82 |
consist routing information, these gateways MUST have direct BGP |
| 83 |
sessions with each other (the BGP sessions should form a complete |
| 84 |
graph). Note that this requirement does not imply that all BGP |
| 85 |
gateways within a single AS must have direct links to each other; |
| 86 |
other methods may be used to ensure consistent routing information. |
| 87 |
|
| 88 |
<span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. Message Formats</span> |
| 89 |
|
| 90 |
This section describes message formats and actions to be taken when |
| 91 |
errors are detected while processing these messages. |
| 92 |
|
| 93 |
Messages are sent over a reliable transport protocol connection. A |
| 94 |
message is processed after it is entirely received. The maximum |
| 95 |
message size is 1024 bytes. All implementations are required to |
| 96 |
support this maximum message size. The smallest message that may be |
| 97 |
sent consists of a BGP header without a data portion, or 8 bytes. |
| 98 |
|
| 99 |
The phrase "the BGP connection is closed" means that the transport |
| 100 |
protocol connection has been closed and that all resources for that |
| 101 |
BGP connection have been deallocated. Routing table entries |
| 102 |
associated with the remote peer are marked as invalid. This |
| 103 |
information is passed to other BGP peers before being deleted from |
| 104 |
the system. |
| 105 |
|
| 106 |
|
| 107 |
|
| 108 |
<span class="grey">Lougheed & Rekhter [Page 2]</span></pre> |
| 109 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-3" ></span> |
| 110 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 111 |
|
| 112 |
|
| 113 |
<span class="h3"><a class="selflink" id="section-3.1" href="#section-3.1">3.1</a> Message Header Format</span> |
| 114 |
|
| 115 |
Each message has a fixed size header. There may or may not be a data |
| 116 |
portion following the header, depending on the message type. The |
| 117 |
layout of these fields is shown below. |
| 118 |
|
| 119 |
0 1 2 3 |
| 120 |
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 |
| 121 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 122 |
| Marker | Length | |
| 123 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 124 |
| Version | Type | Hold Time | |
| 125 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 126 |
|
| 127 |
Marker: 16 bits |
| 128 |
|
| 129 |
The Marker field is 16 bits of all ones. This field is used to |
| 130 |
mark the start of a message. If the first two bytes of a message |
| 131 |
are not all ones then we have a synchronization error and the BGP |
| 132 |
connection should be closed after sending a notification message |
| 133 |
with opcode 5 (connection not synchronized). No notification data |
| 134 |
is sent. |
| 135 |
|
| 136 |
Length: 16 bits |
| 137 |
|
| 138 |
The Length field is 16 bits. It is the total length of the |
| 139 |
message, incluluding header, in bytes. If an illegal length is |
| 140 |
encountered (more than 1024 bytes or less than 8 bytes), a |
| 141 |
notification message with opcode 6 (bad message length) and two |
| 142 |
data bytes of the bad length should be sent and the BGP connection |
| 143 |
closed. |
| 144 |
|
| 145 |
Version: 8 bits |
| 146 |
|
| 147 |
The Version field is 8 bits of protocol version number. The |
| 148 |
current BGP version number is 1. If a bad version number is |
| 149 |
found, a notification message with opcode 8 (bad version number) |
| 150 |
should be sent and the BGP connection closed. The bad version |
| 151 |
number should be included in one byte of notification data. |
| 152 |
|
| 153 |
Type: 8 bits |
| 154 |
|
| 155 |
The Type field is 8 bits of message type code. The following type |
| 156 |
codes are defined: |
| 157 |
|
| 158 |
|
| 159 |
|
| 160 |
|
| 161 |
|
| 162 |
|
| 163 |
|
| 164 |
<span class="grey">Lougheed & Rekhter [Page 3]</span></pre> |
| 165 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-4" ></span> |
| 166 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 167 |
|
| 168 |
|
| 169 |
1 - OPEN |
| 170 |
2 - UPDATE |
| 171 |
3 - NOTIFICATION |
| 172 |
4 - KEEPALIVE |
| 173 |
5 - OPEN CONFIRM |
| 174 |
|
| 175 |
If an unrecognized type value is found, a notification message |
| 176 |
with opcode 7 (bad type code) and data consisting of the byte of |
| 177 |
type field in question should be sent and the BGP connection |
| 178 |
closed. |
| 179 |
|
| 180 |
Hold Timer: 16 bits. |
| 181 |
|
| 182 |
This field contains the number of seconds that may elapse since |
| 183 |
receiving a BGP KEEPALIVE or BGP UPDATE message from our BGP peer |
| 184 |
before we declare an error and close the BGP connection. |
| 185 |
|
| 186 |
<span class="h3"><a class="selflink" id="section-3.2" href="#section-3.2">3.2</a> OPEN Message Format</span> |
| 187 |
|
| 188 |
After a transport protocol connection is established, the first |
| 189 |
message sent by either side is an OPEN message. If the OPEN message |
| 190 |
is acceptable, an OPEN CONFIRM message confirming the OPEN is sent |
| 191 |
back. Once the OPEN is confirmed, UPDATE, KEEPALIVE, and |
| 192 |
NOTIFICATION messages may be exchanged. |
| 193 |
|
| 194 |
In addition to the fixed size BGP header, the OPEN message contains |
| 195 |
the following fields. |
| 196 |
|
| 197 |
0 1 2 3 |
| 198 |
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 |
| 199 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 200 |
| My Autonomous System | Link Type | Auth. Code | |
| 201 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 202 |
| | |
| 203 |
| Authentication Data | |
| 204 |
| | |
| 205 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 206 |
|
| 207 |
My Autonomous System: 16 bits |
| 208 |
|
| 209 |
This field is our 16 bit autonomous system number. If there is a |
| 210 |
problem with this field, a notification message with opcode 9 |
| 211 |
(invalid AS field) should be sent and the BGP connection closed. |
| 212 |
No notification data is sent. |
| 213 |
|
| 214 |
Link Type: 8 bits |
| 215 |
|
| 216 |
The Link Type field is a single octet containing one of the |
| 217 |
|
| 218 |
|
| 219 |
|
| 220 |
<span class="grey">Lougheed & Rekhter [Page 4]</span></pre> |
| 221 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-5" ></span> |
| 222 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 223 |
|
| 224 |
|
| 225 |
following codes defining our position in the AS graph relative to |
| 226 |
our peer. |
| 227 |
|
| 228 |
0 - INTERNAL |
| 229 |
1 - UP |
| 230 |
2 - DOWN |
| 231 |
3 - H-LINK |
| 232 |
|
| 233 |
UP indicates the peer is higher in the AS hierarchy, DOWN |
| 234 |
indicates lower, and H-LINK indicates at the same level. INTERNAL |
| 235 |
indicates that the peer is another BGP speaking host in our |
| 236 |
autonomous system. INTERNAL links are used to keep AS routing |
| 237 |
information consistent with an AS with multiple border gateways. |
| 238 |
If the Link Type field is unacceptable, a notification message |
| 239 |
with opcode 1 (link type error in open) and data consisting of the |
| 240 |
expected link type should be sent and the BGP connection closed. |
| 241 |
The acceptable values for the Link Type fields of two BGP peers |
| 242 |
are discussed below. |
| 243 |
|
| 244 |
Authentication Code: 8 bits |
| 245 |
|
| 246 |
The Authentication Code field is an octet whose value describes |
| 247 |
the authentication mechanism being used. A value of zero |
| 248 |
indicates no BGP authentication. Note that a separate |
| 249 |
authentication mechanism may be used in establishing the transport |
| 250 |
level connection. If the authentication code is not recognized, a |
| 251 |
notification message with opcode 2 (unknown authentication code) |
| 252 |
and no data is sent and the BGP connection is closed. |
| 253 |
|
| 254 |
Authentication Data: variable length |
| 255 |
|
| 256 |
The Authentication Data field is a variable length field |
| 257 |
containing authentication data. If the value of Authentication |
| 258 |
Code field is zero, the Authentication Data field has zero length. |
| 259 |
If authentication fails, a notification message with opcode 3 |
| 260 |
(authentication failure) and no data is sent and the BGP |
| 261 |
connection is closed. |
| 262 |
|
| 263 |
<span class="h3"><a class="selflink" id="section-3.3" href="#section-3.3">3.3</a> OPEN CONFIRM Message Format</span> |
| 264 |
|
| 265 |
An OPEN CONFIRM message is sent after receiving an OPEN message. |
| 266 |
This completes the BGP connection setup. UPDATE, NOTIFICATION, and |
| 267 |
KEEPALIVE messages may now be exchanged. |
| 268 |
|
| 269 |
An OPEN CONFIRM message consists of a BGP header with an OPEN CONFIRM |
| 270 |
type code. There is no data in an OPEN CONFIRM message. |
| 271 |
|
| 272 |
|
| 273 |
|
| 274 |
|
| 275 |
|
| 276 |
<span class="grey">Lougheed & Rekhter [Page 5]</span></pre> |
| 277 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-6" ></span> |
| 278 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 279 |
|
| 280 |
|
| 281 |
<span class="h3"><a class="selflink" id="section-3.4" href="#section-3.4">3.4</a> UPDATE Message Format</span> |
| 282 |
|
| 283 |
UPDATE messages are used to transfer routing information between BGP |
| 284 |
peers. The information in the UPDATE packet can be used to construct |
| 285 |
a graph describing the relationships of the various autonomous |
| 286 |
systems. By applying rules to be discussed, routing information |
| 287 |
loops and some other anomalies may be detected and removed from the |
| 288 |
inter-AS routing. |
| 289 |
|
| 290 |
Whenever an error in a UPDATE message is detected, a notification |
| 291 |
message is sent with opcode 4 (bad update), a two byte subcode |
| 292 |
describing the nature of the problem, and a data field consisting of |
| 293 |
as much of the UPDATE message data portion as possible. UPDATE |
| 294 |
messages have the following format: |
| 295 |
|
| 296 |
0 1 2 3 |
| 297 |
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 |
| 298 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 299 |
| Gateway | |
| 300 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 301 |
| AS count | Direction | AS Number | |
| 302 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 303 |
| repeat (Direction, AS Number) pairs AS count times | |
| 304 |
/ / |
| 305 |
/ / |
| 306 |
| | |
| 307 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 308 |
| Net Count | |
| 309 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 310 |
| Network | |
| 311 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 312 |
| Metric | | |
| 313 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |
| 314 |
| repeat (Network, Metric) pairs Net Count times | |
| 315 |
/ / |
| 316 |
/ / |
| 317 |
| | |
| 318 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 319 |
|
| 320 |
Gateway: 32 bits. |
| 321 |
|
| 322 |
The Gateway field is the address of a gateway that has routes to |
| 323 |
the Internet networks listed in the rest of the UPDATE message. |
| 324 |
This gateway MUST belong to the same AS as the BGP peer who |
| 325 |
advertises it. If there is a problem with the gateway field, a |
| 326 |
notification message with subcode 6 (invalid gateway field) is |
| 327 |
sent. |
| 328 |
|
| 329 |
|
| 330 |
|
| 331 |
|
| 332 |
<span class="grey">Lougheed & Rekhter [Page 6]</span></pre> |
| 333 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-7" ></span> |
| 334 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 335 |
|
| 336 |
|
| 337 |
AS count: 8 bits. |
| 338 |
|
| 339 |
This field is the count of Direction and AS Number pairs in this |
| 340 |
UPDATE message. If an incorrect AS count field is detected, |
| 341 |
subcode 1 (invalid AS count) is specified in the notification |
| 342 |
message. |
| 343 |
|
| 344 |
Direction: 8 bits |
| 345 |
|
| 346 |
The Direction field is an octet containing the direction taken by |
| 347 |
the routing information when exiting the AS defined by the |
| 348 |
succeeding AS Number field. The following values are defined. |
| 349 |
|
| 350 |
1 - UP (went up a link in the graph) |
| 351 |
2 - DOWN (went down a link in the graph) |
| 352 |
3 - H_LINK (horizontal link in the graph) |
| 353 |
4 - EGP_LINK (EGP derived information) |
| 354 |
5 - INCOMPLETE (incomplete information) |
| 355 |
|
| 356 |
There is a special provision to pass exterior learned (non-BGP) |
| 357 |
routes over BGP. If an EGP learned route is passed over BGP, then |
| 358 |
the Direction field is set to EGP-LINK and the AS Number field is |
| 359 |
set to the AS number of the EGP peer that advertised this route. |
| 360 |
All other exterior-learned routes (non-BGP and non-EGP) may be |
| 361 |
passed by setting AS Number field to zero and Direction field to |
| 362 |
INCOMPLETE. If the direction code is not recognized, a |
| 363 |
notification message with subcode 2 (invalid direction code) is |
| 364 |
sent. |
| 365 |
|
| 366 |
AS Number: 16 bits |
| 367 |
|
| 368 |
This field is the AS number that transmitted the routing |
| 369 |
information. If there is a problem with this AS number, a |
| 370 |
notification message with subcode 3 (invalid autonomous system) is |
| 371 |
sent. |
| 372 |
|
| 373 |
Net Count: 16 bits. |
| 374 |
|
| 375 |
The Net Count field is the number of Metric and Network field |
| 376 |
pairs which follow this field. If there is a problem with this |
| 377 |
field, a notification with subcode 7 (invalid net count field) is |
| 378 |
sent. |
| 379 |
|
| 380 |
Network: 32 bits |
| 381 |
|
| 382 |
The Network field is four bytes of Internet network number. If |
| 383 |
there is a problem with the network field, a notification message |
| 384 |
with subcode 8 (invalid network field) is sent. |
| 385 |
|
| 386 |
|
| 387 |
|
| 388 |
<span class="grey">Lougheed & Rekhter [Page 7]</span></pre> |
| 389 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-8" ></span> |
| 390 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 391 |
|
| 392 |
|
| 393 |
Metric: 16 bits |
| 394 |
|
| 395 |
The Metric field is 16 bits of an unspecified metric. BGP metrics |
| 396 |
are comparable ONLY if routes have exactly the same AS path. A |
| 397 |
metric of all ones indicates the network is unreachable. In all |
| 398 |
other cases the metric field is MEANINGLESS and MUST BE IGNORED. |
| 399 |
There are no illegal metric values. |
| 400 |
|
| 401 |
<span class="h3"><a class="selflink" id="section-3.5" href="#section-3.5">3.5</a> NOTIFICATION Message Format</span> |
| 402 |
|
| 403 |
NOTIFICATION messages are sent when an error condition is detected. |
| 404 |
The BGP connection is closed shortly after sending the notification |
| 405 |
message. |
| 406 |
|
| 407 |
0 1 2 3 |
| 408 |
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 |
| 409 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 410 |
| Opcode | Data | |
| 411 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |
| 412 |
| | |
| 413 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 414 |
|
| 415 |
Opcode: 16 bits |
| 416 |
|
| 417 |
The Opcode field describes the type of NOTIFICATION. The |
| 418 |
following opcodes have been defined. |
| 419 |
|
| 420 |
1 (*) - link type error in open. Data is one byte of proper |
| 421 |
link type. |
| 422 |
2 (*) - unknown authentication code. No data. |
| 423 |
3 (*) - authentication failure. No data. |
| 424 |
4 - update error. See below for data description. |
| 425 |
5 (*) - connection out of sync. No data. |
| 426 |
6 (*) - invalid message length. Data is two bytes of |
| 427 |
bad length. |
| 428 |
7 (*) - invalid message type. Data is one byte of bad |
| 429 |
message type. |
| 430 |
8 (*) - invalid version number. Data is one byte of |
| 431 |
bad version. |
| 432 |
9 (*) - invalid AS field in OPEN. No data. |
| 433 |
10 (*) - BGP Cease. No data. |
| 434 |
|
| 435 |
The starred opcodes in the list above are considered fatal errors |
| 436 |
and cause transport connection termination. |
| 437 |
|
| 438 |
The update error (opcode 4) has as data 16 bits of subcode |
| 439 |
followed by the last UPDATE message in question. After the |
| 440 |
subcode comes as much of the data portion of the UPDATE in |
| 441 |
|
| 442 |
|
| 443 |
|
| 444 |
<span class="grey">Lougheed & Rekhter [Page 8]</span></pre> |
| 445 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-9" ></span> |
| 446 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 447 |
|
| 448 |
|
| 449 |
question as possible. The following subcodes are defined: |
| 450 |
|
| 451 |
1 - invalid AS count |
| 452 |
2 - invalid direction code |
| 453 |
3 - invalid autonomous system |
| 454 |
4 - EGP_LINK or INCOMPLETE_LINK link type at other than |
| 455 |
the end of the AS path list |
| 456 |
5 - routing loop |
| 457 |
6 - invalid gateway field |
| 458 |
7 - invalid Net Count field |
| 459 |
8 - invalid network field |
| 460 |
|
| 461 |
Data: variable |
| 462 |
|
| 463 |
The Data field contains zero or more bytes of data to be used in |
| 464 |
diagnosing the reason for the NOTIFICATION. The contents of the |
| 465 |
Data field depend upon the opcode. See the opcode descriptions |
| 466 |
above for more details. |
| 467 |
|
| 468 |
<span class="h3"><a class="selflink" id="section-3.6" href="#section-3.6">3.6</a> KEEPALIVE Message Format</span> |
| 469 |
|
| 470 |
BGP does not use any transport protocol based keepalive mechanism to |
| 471 |
determine if peers are reachable. Instead KEEPALIVE messages are |
| 472 |
exchanged between peers often enough as not to cause the hold time |
| 473 |
(as advertised in the BGP header) to expire. A reasonable minimum |
| 474 |
frequency of KEEPALIVE exchange would be one third of the Hold Time |
| 475 |
interval. |
| 476 |
|
| 477 |
As soon as the Hold Time associated with BGP peer has expired, the |
| 478 |
BGP connection is closed and BGP deallocates all resources associated |
| 479 |
with this peer. |
| 480 |
|
| 481 |
The KEEPALIVE message is a BGP header without any data. |
| 482 |
|
| 483 |
<span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a>. BGP Finite State machine.</span> |
| 484 |
|
| 485 |
This section specifies BGP operation in terms of a Finite State |
| 486 |
Machine (FSM). Following is a brief summary and overview of BGP |
| 487 |
operations by state as determined by this FSM. A condensed version |
| 488 |
of the BGP FSM is found in Appendix 1. |
| 489 |
|
| 490 |
Initially BGP is in the BGP_Idle state. |
| 491 |
|
| 492 |
BGP_Idle state: |
| 493 |
|
| 494 |
In this state BGP refuses all incoming BGP connections. No |
| 495 |
resources are allocated to the BGP neighbor. In response to the |
| 496 |
Start event (initiated by either system or operator) the local |
| 497 |
|
| 498 |
|
| 499 |
|
| 500 |
<span class="grey">Lougheed & Rekhter [Page 9]</span></pre> |
| 501 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-10" ></span> |
| 502 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 503 |
|
| 504 |
|
| 505 |
system initializes all BGP resources and changes its state to |
| 506 |
BGP_Active. |
| 507 |
|
| 508 |
BGP_Active state: |
| 509 |
|
| 510 |
In this state BGP is trying to acquire a BGP neighbor by opening a |
| 511 |
transport protocol connection. If the transport protocol open |
| 512 |
fails (for example, retransmission timeout), BGP stays in the |
| 513 |
BGP_Active state. |
| 514 |
|
| 515 |
Otherwise, the local system sends an OPEN message to its peer, |
| 516 |
and changes its state to BGP_OpenSent. Since the hold time of the |
| 517 |
peer is still undetermined, the hold time is initialized to some |
| 518 |
large value. |
| 519 |
|
| 520 |
In response to the Stop event (initiated by either system or |
| 521 |
operator) the local system releases all BGP resources and changes |
| 522 |
its state to BGP_Idle. |
| 523 |
|
| 524 |
BGP_OpenSent state: |
| 525 |
|
| 526 |
In this state BGP waits for an OPEN message from its peer. When |
| 527 |
an OPEN message is received, all fields are checked for |
| 528 |
correctness. If the initial BGP header checking detects an error, |
| 529 |
BGP deallocates all resources associated with this peer and |
| 530 |
returns to the BGP_Active state. Otherwise, the Link Type, |
| 531 |
Authentication Code, and Authentication Data fields are checked |
| 532 |
for correctness. |
| 533 |
|
| 534 |
If the link type is incorrect, a NOTIFICATION message with opcode |
| 535 |
1 (link type error in open) is sent. The following combination of |
| 536 |
link type fields are correct; all other combinations are invalid. |
| 537 |
|
| 538 |
Our view Peer view |
| 539 |
UP DOWN |
| 540 |
DOWN UP |
| 541 |
INTERNAL INTERNAL |
| 542 |
H-LINK H-LINK |
| 543 |
|
| 544 |
If the link between two peers is INTERNAL, then AS number of both |
| 545 |
peers must be the same. Otherwise, a NOTIFICATION message with |
| 546 |
opcode 1 (link type error in open) is sent. |
| 547 |
|
| 548 |
If both peers have the same AS number and the link type between |
| 549 |
these peers is not INTERNAL, then a NOTIFICATION message with |
| 550 |
opcode 1 (link type error in open) is sent. |
| 551 |
|
| 552 |
If the value of the Authentication Code field is zero, any |
| 553 |
|
| 554 |
|
| 555 |
|
| 556 |
<span class="grey">Lougheed & Rekhter [Page 10]</span></pre> |
| 557 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-11" ></span> |
| 558 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 559 |
|
| 560 |
|
| 561 |
information in the Authentication Data field (if present) is |
| 562 |
ignored. If the Authentication Code field is non-zero it is |
| 563 |
checked for known authentication codes. If authentication code is |
| 564 |
unknown, then the BGP NOTIFICATION message with opcode 2 (unknown |
| 565 |
authentication code) is sent. |
| 566 |
|
| 567 |
If the Authentication Code value is non-zero, then the |
| 568 |
corresponding authentication procedure is invoked. The default |
| 569 |
values are a zero Authentication Code and no Authentication Data. |
| 570 |
|
| 571 |
If any of the above tests detect an error, the local system closes |
| 572 |
the BGP connection and changes its state to BGP_Idle. |
| 573 |
|
| 574 |
If there are no errors in the BGP OPEN message, BGP sends an OPEN |
| 575 |
CONFIRM message and goes into the BGP_OpenConfirm state. At this |
| 576 |
point the hold timer which was originally set to some arbitrary |
| 577 |
large value (see above) is replaced with the value indicated in |
| 578 |
the OPEN message. |
| 579 |
|
| 580 |
If disconnect notification is received from the underlying |
| 581 |
transport protocol or if the hold time expires, the local system |
| 582 |
closes the BGP connection and changes its state to BGP_Idle. |
| 583 |
|
| 584 |
BGP_OpenConfirm state: |
| 585 |
|
| 586 |
In this state BGP waits for an OPEN CONFIRM message. As soon as |
| 587 |
this message is received, BGP changes its state to |
| 588 |
BGP_Established. If the hold timer expires before an OPEN CONFIRM |
| 589 |
message is received, the local system closes the BGP connection |
| 590 |
and changes its state to BGP_Idle. |
| 591 |
|
| 592 |
BGP_Established state: |
| 593 |
|
| 594 |
In the BGP_Established state BGP can exchange UPDATE, |
| 595 |
NOTIFICATION, and KEEPALIVE messages with its peer. |
| 596 |
|
| 597 |
If disconnect notification is received from the underlying |
| 598 |
transport protocol or if the hold time expires, the local system |
| 599 |
closes the BGP connection and changes its state to BGP_Idle. |
| 600 |
|
| 601 |
In response to the Stop event initiated by either the system or |
| 602 |
operator, the local system sends a NOTIFICATION message with |
| 603 |
opcode 10 (BGP Cease), closes the BGP connection, and changes its |
| 604 |
state to BGP_Idle. |
| 605 |
|
| 606 |
|
| 607 |
|
| 608 |
|
| 609 |
|
| 610 |
|
| 611 |
|
| 612 |
<span class="grey">Lougheed & Rekhter [Page 11]</span></pre> |
| 613 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-12" ></span> |
| 614 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 615 |
|
| 616 |
|
| 617 |
<span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a>. UPDATE Message Handling</span> |
| 618 |
|
| 619 |
A BGP UPDATE message may be received only in the BGP_Established |
| 620 |
state. When a BGP UPDATE message is received, each field is checked |
| 621 |
for validity. When a NOTIFICATION message is sent regarding an |
| 622 |
UPDATE, the opcode is always 4 (update error), the subcode depends on |
| 623 |
the type of error, and the rest of the data field is as much as |
| 624 |
possible of the data portion of the UPDATE that caused the error. |
| 625 |
|
| 626 |
If the Gateway field is incorrect, a BGP NOTIFICATION message is sent |
| 627 |
with subcode 6 (invalid gateway field). All information in this |
| 628 |
UPDATE message is discarded. |
| 629 |
|
| 630 |
If the AS Count field is less than or equal to zero, a BGP |
| 631 |
NOTIFICATION is sent with subcode 1 (invalid AS count). Otherwise, |
| 632 |
the complete AS path is extracted and checked as described below. |
| 633 |
|
| 634 |
If one of the Direction fields in the AS route list is not defined, a |
| 635 |
BGP NOTIFICATION message is with subcode 2 (invalid direction code). |
| 636 |
|
| 637 |
If one of the AS Number fields in the AS route list is incorrect, a |
| 638 |
BGP NOTIFICATION message is sent with subcode 3 (invalid autonomous |
| 639 |
system). |
| 640 |
|
| 641 |
If either a EGP_LINK or a INCOMPLETE_LINK link type occurs at other |
| 642 |
than the end of the AS path, a BGP NOTIFICATION message is sent with |
| 643 |
subcode 4 (EGP_LINK or INCOMPLETE_LINK link type at other than the |
| 644 |
end of the AS path list). |
| 645 |
|
| 646 |
If none of the above tests failed, the full AS route is checked for |
| 647 |
AS loops. |
| 648 |
|
| 649 |
AS loop detection is done by scanning the full AS route and checking |
| 650 |
that each AS in this route occurs only once. If an AS loop is |
| 651 |
detected, a BGP NOTIFICATION message is sent with subcode 5 (routing |
| 652 |
loop). |
| 653 |
|
| 654 |
If any of the above errors are detected, no further processing is |
| 655 |
done. Otherwise, the complete AS path is correct and the rest of the |
| 656 |
UPDATE message is processed. |
| 657 |
|
| 658 |
If the Net Count field is incorrect, a BGP NOTIFICATION message is |
| 659 |
sent with subcode 7 (invalid Net Count field). |
| 660 |
|
| 661 |
Each network and metric pair listed in the BGP UPDATE message is |
| 662 |
checked for a valid network number. If the Network field is |
| 663 |
incorrect, a BGP Notification message is sent with subcode 8 (invalid |
| 664 |
network field). No checking is done on the metric field. It is up |
| 665 |
|
| 666 |
|
| 667 |
|
| 668 |
<span class="grey">Lougheed & Rekhter [Page 12]</span></pre> |
| 669 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-13" ></span> |
| 670 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 671 |
|
| 672 |
|
| 673 |
to a particular implementation to decide whether to continue |
| 674 |
processing or terminate it upon the first incorrect network. |
| 675 |
|
| 676 |
If the network, its complete AS path, and the gateway are correct, |
| 677 |
then the route is compared with other routes to the same network. If |
| 678 |
the new route is better than the current one, then it is flooded to |
| 679 |
other BGP peers as follows: |
| 680 |
|
| 681 |
- If the BGP UPDATE was received over the INTERNAL link, it is not |
| 682 |
propagated over any other INTERNAL link. This restriction is |
| 683 |
due to the fact that all BGP gateways within a single AS |
| 684 |
form a completely connected graph (see above). |
| 685 |
|
| 686 |
- Before sending a BGP UPDATE message over the non-INTERNAL links, |
| 687 |
check the AS path to insure that doing so would not cause a |
| 688 |
routing loop. The BGP UPDATE message is then propagated (subject |
| 689 |
to the local policy restrictions) over any of the non-INTERNAL |
| 690 |
link of a routing loop would not result. |
| 691 |
|
| 692 |
- If the BGP UPDATE message is propagated over a non-INTERNAL link, |
| 693 |
then the current AS number and link type of the link over which |
| 694 |
it is going to be propagated is prepended to the full AS path |
| 695 |
and the AS count field is incremented by 1. If the BGP UPDATE |
| 696 |
message is propagated over an INTERNAL link, then the full AS |
| 697 |
path passed unmodified and the AS count stays the same. The |
| 698 |
Gateway field is replaced with the sender's own address. |
| 699 |
|
| 700 |
<span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a>. Acknowledgements</span> |
| 701 |
|
| 702 |
We would like to express our thanks to Len Bosack (cisco Systems), |
| 703 |
Jeff Honig (Cornell University) and all members of the IWG task force |
| 704 |
for their contributions to this document. |
| 705 |
|
| 706 |
|
| 707 |
|
| 708 |
|
| 709 |
|
| 710 |
|
| 711 |
|
| 712 |
|
| 713 |
|
| 714 |
|
| 715 |
|
| 716 |
|
| 717 |
|
| 718 |
|
| 719 |
|
| 720 |
|
| 721 |
|
| 722 |
|
| 723 |
|
| 724 |
<span class="grey">Lougheed & Rekhter [Page 13]</span></pre> |
| 725 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-14" ></span> |
| 726 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 727 |
|
| 728 |
|
| 729 |
Appendix 1 |
| 730 |
|
| 731 |
BGP FSM State Transitions and Actions. |
| 732 |
|
| 733 |
This Appendix discusses the transitions between states in the BGP FSM |
| 734 |
in response to BGP events. The following is the list of these states |
| 735 |
and events. |
| 736 |
|
| 737 |
BGP States: |
| 738 |
|
| 739 |
1 - BGP_Idle |
| 740 |
2 - BGP_Active |
| 741 |
3 - BGP_OpenSent |
| 742 |
4 - BGP_OpenConfirm |
| 743 |
5 - BGP_Established |
| 744 |
|
| 745 |
BGP Events: |
| 746 |
|
| 747 |
1 - BGP Start |
| 748 |
2 - BGP Transport connection open |
| 749 |
3 - BGP Transport connection closed |
| 750 |
4 - BGP Transport connection open failed |
| 751 |
5 - Receive OPEN message |
| 752 |
6 - Receive OPEN CONFIRM message |
| 753 |
7 - Receive KEEPALIVE message |
| 754 |
8 - Receive UPDATE messages |
| 755 |
9 - Receive NOTIFICATION message |
| 756 |
10 - Holdtime timer expired |
| 757 |
11 - KeepAlive timer expired |
| 758 |
12 - Receive CEASE message |
| 759 |
13 - BGP Stop |
| 760 |
|
| 761 |
The following table describes the state transitions of the BGP FSM |
| 762 |
and the actions triggered by these transitions. |
| 763 |
|
| 764 |
|
| 765 |
|
| 766 |
|
| 767 |
|
| 768 |
|
| 769 |
|
| 770 |
|
| 771 |
|
| 772 |
|
| 773 |
|
| 774 |
|
| 775 |
|
| 776 |
|
| 777 |
|
| 778 |
|
| 779 |
|
| 780 |
<span class="grey">Lougheed & Rekhter [Page 14]</span></pre> |
| 781 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-15" ></span> |
| 782 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 783 |
|
| 784 |
|
| 785 |
Event Actions Message Sent Next State |
| 786 |
-------------------------------------------------------------------- |
| 787 |
BGP_Idle (1) |
| 788 |
1 Initialize resources none 2 |
| 789 |
BGP_Active (2) |
| 790 |
2 Initialize resources OPEN 3 |
| 791 |
4 none none 2 |
| 792 |
13 Release resources none 1 |
| 793 |
|
| 794 |
BGP_OpenSent(3) |
| 795 |
3 none none 1 |
| 796 |
5 Process OPEN is OK OPEN CONFIRM 4 |
| 797 |
Process OPEN Message failed NOTIFICATION 1 |
| 798 |
11 Restart KeepAlive timer KEEPALIVE 3 |
| 799 |
13 Release resources none 1 |
| 800 |
|
| 801 |
BGP_OpenConfirm (4) |
| 802 |
6 Complete initialization none 5 |
| 803 |
3 none none 1 |
| 804 |
10 Close transport connection none 1 |
| 805 |
11 Restart KeepAlive timer KEEPALIVE 4 |
| 806 |
13 Release resources none 1 |
| 807 |
|
| 808 |
BGP_Established (5) |
| 809 |
7 Process KEEPALIVE none 5 |
| 810 |
8 Process UPDATE is OK UPDATE 5 |
| 811 |
Process UPDATE failed NOTIFICATION 5 |
| 812 |
9 Process NOTIFICATION none 5 |
| 813 |
10 Close transport connection none 1 |
| 814 |
11 Restart KeepAlive timer KEEPALIVE 5 |
| 815 |
12 Close transport connection NOTIFICATION 1 |
| 816 |
13 Release resources none 1 |
| 817 |
-------------------------------------------------------------------- |
| 818 |
|
| 819 |
All other state-event combinations are considered fatal errors and |
| 820 |
cause the termination of the BGP transport connection (if necessary) |
| 821 |
and a transition to the BGP_Idle state. |
| 822 |
|
| 823 |
|
| 824 |
|
| 825 |
|
| 826 |
|
| 827 |
|
| 828 |
|
| 829 |
|
| 830 |
|
| 831 |
|
| 832 |
|
| 833 |
|
| 834 |
|
| 835 |
|
| 836 |
<span class="grey">Lougheed & Rekhter [Page 15]</span></pre> |
| 837 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-16" ></span> |
| 838 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 839 |
|
| 840 |
|
| 841 |
The following is a condensed version of the above state transition |
| 842 |
table. |
| 843 |
|
| 844 |
Events|BGP_Idle BGP_Active BGP_OpenSent BGP_OpenConfirm BGP_Estab |
| 845 |
| (1) | (2) | (3) | (4) | (5) |
| 846 |
|------------------------------------------------------------- |
| 847 |
1 | 2 | | | | |
| 848 |
| | | | | |
| 849 |
2 | | 3 | | | |
| 850 |
| | | | | |
| 851 |
3 | | | 1 | 1 | |
| 852 |
| | | | | |
| 853 |
4 | | 2 | | | |
| 854 |
| | | | | |
| 855 |
5 | | | 4 or 1 | | |
| 856 |
| | | | | |
| 857 |
6 | | | | 5 | |
| 858 |
| | | | | |
| 859 |
7 | | | | | 5 |
| 860 |
| | | | | |
| 861 |
8 | | | | | 5 |
| 862 |
| | | | | |
| 863 |
9 | | | | | 5 |
| 864 |
| | | | | |
| 865 |
10 | | | | 1 | 1 |
| 866 |
| | | | | |
| 867 |
11 | | | 3 | 4 | 5 |
| 868 |
| | | | | |
| 869 |
12 | | | | | 1 |
| 870 |
| | | | | |
| 871 |
13 | | 1 | 1 | 1 | 1 |
| 872 |
| | | | | |
| 873 |
-------------------------------------------------------------- |
| 874 |
|
| 875 |
|
| 876 |
|
| 877 |
|
| 878 |
|
| 879 |
|
| 880 |
|
| 881 |
|
| 882 |
|
| 883 |
|
| 884 |
|
| 885 |
|
| 886 |
|
| 887 |
|
| 888 |
|
| 889 |
|
| 890 |
|
| 891 |
|
| 892 |
<span class="grey">Lougheed & Rekhter [Page 16]</span></pre> |
| 893 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-17" ></span> |
| 894 |
<span class="grey"><a href="./rfc1105">RFC 1105</a> BGP June 1989</span> |
| 895 |
|
| 896 |
|
| 897 |
References |
| 898 |
|
| 899 |
[<a id="ref-1">1</a>] Mills, D., "Exterior Gateway Protocol Formal Specification", <a href="./rfc904">RFC</a> |
| 900 |
<a href="./rfc904">904</a>, BBN, April 1984. |
| 901 |
|
| 902 |
[<a id="ref-2">2</a>] Rekhter, Y., "EGP and Policy Based Routing in the New NSFNET |
| 903 |
Backbone", <a href="./rfc1092">RFC 1092</a>, T. J. Watson Research Center, February 1989. |
| 904 |
|
| 905 |
[<a id="ref-3">3</a>] Braun, H-W., "The NSFNET Routing Architecture", <a href="./rfc1093">RFC 1093</a>, |
| 906 |
MERIT/NSFNET Project, February 1989. |
| 907 |
|
| 908 |
[<a id="ref-4">4</a>] Postel, J., "Transmission Control Protocol - DARPA Internet |
| 909 |
Program Protocol Specification", <a href="./rfc793">RFC 793</a>, DARPA, September 1981. |
| 910 |
|
| 911 |
[<a id="ref-5">5</a>] Cheriton, D., "VMTP: Versatile Message Transaction Protocol", <a href="./rfc1045">RFC</a> |
| 912 |
<a href="./rfc1045">1045</a>, Stanford University, February 1988. |
| 913 |
|
| 914 |
Authors' Addresses |
| 915 |
|
| 916 |
Kirk Lougheed |
| 917 |
cisco Systems, Inc. |
| 918 |
1360 Willow Road, Suite 201 |
| 919 |
Menlo Park, CA 94025 |
| 920 |
|
| 921 |
Phone: (415) 326-1941 |
| 922 |
|
| 923 |
Email: [email protected] |
| 924 |
|
| 925 |
|
| 926 |
Jacob Rekhter |
| 927 |
T.J. Watson Research Center |
| 928 |
IBM Corporation |
| 929 |
P.O. Box 218 |
| 930 |
Yorktown Heights, NY 10598 |
| 931 |
|
| 932 |
Phone: (914) 945-3896 |
| 933 |
|
| 934 |
Email: [email protected] |
| 935 |
|
| 936 |
|
| 937 |
|
| 938 |
|
| 939 |
|
| 940 |
|
| 941 |
|
| 942 |
|
| 943 |
|
| 944 |
|
| 945 |
|
| 946 |
|
| 947 |
|
| 948 |
Lougheed & Rekhter [Page 17] |
| 949 |
</pre> |