Debian Codegraph @

rfc2340.html

raw
1 <pre>Network Working Group B. Jamoussi
2 Request for Comments: 2340 D. Jamieson
3 Category: Informational D. Williston
4 S. Gabe
5 Nortel (Northern Telecom) Ltd.
6 May 1998
7
8
9 <span class="h1">Nortel's Virtual Network Switching (VNS) Overview</span>
10
11 Status of this Memo
12
13 This memo provides information for the Internet community. It does
14 not specify an Internet standard of any kind. Distribution of this
15 memo is unlimited.
16
17 Copyright Notice
18
19 Copyright (C) The Internet Society (1998). All Rights Reserved.
20
21 Abstract
22
23 This document provides an overview of Virtual Network Switching
24 (VNS).
25
26 VNS is a multi-protocol switching architecture that provides COS-
27 sensitive packet switching, reduces the complexity of operating
28 protocols like PPP and frame relay, provides logical networks and
29 traffic segregation for Virtual Private Networks (VPNs), security and
30 traffic engineering, enables efficient WAN broadcasting and
31 multicasting, and reduces address space requirements. VNS reduces the
32 number of routing hops over the WAN by switching packets based on
33 labels.
34
35 VNS has been proven in production networks for several years.
36
37 Table of Contents
38
39 <a href="#section-1">1</a> Introduction ............................................ <a href="#page-2">2</a>
40 <a href="#section-2">2</a> What is VNS? ............................................ <a href="#page-3">3</a>
41 <a href="#section-3">3</a> VNS Header ............................................. <a href="#page-5">5</a>
42 <a href="#section-4">4</a> VNS Label Distribution .................................. <a href="#page-7">7</a>
43 <a href="#section-5">5</a> Logical Networks (LNs) .................................... <a href="#page-7">7</a>
44 <a href="#section-6">6</a> VNS Routing ............................................. <a href="#page-8">8</a>
45 <a href="#section-7">7</a> VNS Forwarding .......................................... <a href="#page-9">9</a>
46 <a href="#section-7.1">7.1</a> Unicast ................................................ <a href="#page-9">9</a>
47 <a href="#section-7.2">7.2</a> Multicast .............................................. <a href="#page-9">9</a>
48 <a href="#section-8">8</a> Traffic Engineering ..................................... <a href="#page-10">10</a>
49
50
51
52 <span class="grey">Jamoussi, et. al. Informational [Page 1]</span></pre>
53 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-2" ></span>
54 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
55
56
57 <a href="#section-8.1">8.1</a> Equal Cost Multipaths .................................. <a href="#page-10">10</a>
58 <a href="#section-8.2">8.2</a> Trunk Load Spreading ................................... <a href="#page-10">10</a>
59 <a href="#section-9">9</a> Class of Service ........................................ <a href="#page-11">11</a>
60 <a href="#section-10">10</a> VNS Migration Strategies ................................ <a href="#page-11">11</a>
61 <a href="#section-11">11</a> Summary ................................................. <a href="#page-11">11</a>
62 <a href="#section-12">12</a> Security Considerations ................................. <a href="#page-12">12</a>
63 <a href="#section-13">13</a> Acknowledgments ......................................... <a href="#page-12">12</a>
64 <a href="#section-14">14</a> Authors' Addresses ...................................... <a href="#page-13">13</a>
65 <a href="#section-15">15</a> Full Copyright Statement ................................ <a href="#page-14">14</a>
66
67 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Introduction</span>
68
69 There are several key problem areas with today's wide area backbone
70 networks that carry LAN traffic: scalability, service
71 differentiation, redundancy, administration, and traffic containment.
72
73 First, scalability is becoming a major concern because of the rapid
74 growth in bandwidth demand and geographical reach. As the size of the
75 WAN network grows traditional point-to-point and NBMA topologies or
76 network models lose their performance.
77
78 Second, the need to provide several Classes of Service (CoS) has
79 never been greater. The days of a single "best effort" service are
80 over and service providers demand ways to differentiate the quality
81 of the service offered to their clients based on several policies.
82
83 Third, the WAN is often carrying mission-critical traffic and loss of
84 service is not acceptable. So far, path redundancy has been addressed
85 inefficiently by requiring additional links or VCs.
86
87 Fourth, network operators demand easy and simplified network
88 administration. Large NBMA topologies require extensive PVC
89 provisioning until SVC deployment becomes more ubiquitous. For
90 Point-to-point models, IP address space may be used inefficiently and
91 non-trivial network schemas are required to contain reserved address
92 space.
93
94 Finally, proper segregation of traffic is becoming a must. This
95 requirement is being addressed today by adding leased lines or VCs
96 used to separate traffic flows based on regions or interest or
97 protocol.
98
99 Nortel's Virtual Network Switching (VNS) is a technology that
100 provides efficient solutions to these challenges.
101
102
103
104
105
106
107
108 <span class="grey">Jamoussi, et. al. Informational [Page 2]</span></pre>
109 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-3" ></span>
110 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
111
112
113 <a href="#section-2">Section 2</a> provides an overview of VNS. The VNS header is specified in
114 <a href="#section-3">Section 3</a>. <a href="#section-4">Section 4</a> describes the VNS label distribution mechanism.
115 <a href="#section-5">Section 5</a> defines how a VNS network can be partitioned into Logical
116 Networks (LN). <a href="#section-6">Section 6</a> outlines VNS routing. <a href="#section-7">Section 7</a> defines both
117 unicast and multicast forwarding. <a href="#section-8">Section 8</a> describes the mechanisms
118 used to engineer the traffic. <a href="#section-9">Section 9</a> defines the COS based
119 switching of VNS. <a href="#section-10">Section 10</a> provides network migration scenarios
120 using VNS. A summary of VNS is provided in <a href="#section-11">Section 11</a>.
121
122 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. What is VNS?</span>
123
124 Virtual Network Switching (VNS) is a CoS-sensitive multi-protocol
125 label switching architecture that reduces or eliminates the number of
126 layer 3 hops over the WAN by switching traffic based on labels.
127
128 VNS makes a network of point to point links appear to be a single
129 LAN (broadcast, multiple access) media. The network used by a
130 particular instance of VNS is called a Logical Network (LN) which is
131 described in more detail in <a href="#section-5">Section 5</a>.
132
133 In reference to the ISO Network Layering Model, the Data Link Layer
134 is expanded to include VNS network layer. To the ISO Network Layer,
135 (e.g., IP), VNS is treated as a Data Link Layer.
136
137 ------------------------
138 | Application |
139 ------------------------
140 | Presentation |
141 ------------------------
142 | Session |
143 ------------------------
144 | Transport |
145 ------------------------ -------------------------
146 | Network (e.g., IP) | / Network VNS |
147 ----------------------------- |
148 | Data Link |--------------------------
149 ----------------------------- |
150 | Physical | \ data link (e.g., ATM) |
151 ------------------------ -------------------------
152
153 Figure 1. ISO Network Layering Model for VNS
154
155 In a VNS Network, three separate nodal functions are defined. An
156 ingress node, an egress node, and a tandem node. The ingress and
157 egress nodes define the boundary between an IP network and the VNS
158 network. Therefore, these nodes run both IP routing and VNS routing.
159 However, tandem nodes need only run VNS routing.
160
161
162
163
164 <span class="grey">Jamoussi, et. al. Informational [Page 3]</span></pre>
165 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-4" ></span>
166 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
167
168
169 A LAN packet is encapsulated in a VNS header as it enters the LN. The
170 label in the header is used to switch the packet across the LN. The
171 encapsulation header contains the identifier of the last node (or
172 egress node) that processes the packet as it traverses the LN. It is
173 the first node (or ingress node) that decides to which egress node
174 the packet is sent. All nodes between the ingress and egress nodes
175 (known as tandem nodes) decide independently the best packet
176 forwarding route to the egress node identified in the packet.
177
178 The network layer protocols view VNS as a shared broadcast media,
179 where the speed to reach any node on the media is the same for all
180 nodes. VNS ensures that traffic destined to other nodes is forwarded
181 optimally. This transparent view of the VNS means that all the
182 details of the network (for example, topology and link states) can be
183 hidden from the Upper Layer Protocols (e.g. Layer 3 routing
184 protocols) and their applications. VNS also ensures that changes to
185 topology and link state are hidden.
186
187 The network layer protocol on the ingress node views the network
188 layer protocol on the egress node as its logical and directly
189 connected neighbor. This is significant because the network layer
190 protocols always decide which directly connected neighbor should
191 receive a forwarded packet. The details of the actual topology
192 supporting the connectionless network are managed entirely by the
193 Virtual Network Switching and are hidden from the network layer
194 protocols. To the network layer, VNS simply appears to be another
195 Data Link Layer (or media), even though VNS is a network layer itself
196 running on top of the actual Data Link Layer (for example, ATM
197 trunks).
198
199 For the ingress node to choose the egress node that provides the best
200 path to the packet's final destination, it must have knowledge of the
201 following:
202
203 - the nodes that can be reached in the network
204 - the topology of the network that is using the VNS services for
205 transport across the network (but not necessarily the topology
206 of the full network)
207
208 This knowledge is obtained through the network layer routing
209 mechanisms such as, IP's Open Shortest Path First (OSPF) and Address
210 Resolution Protocol (ARP).
211
212 Once the network layer protocol on the ingress node has decided which
213 neighbor to transmit the packet to, it is the responsibility of VNS
214 forwarding, a part of VNS, to deliver the packet to that node. Once
215 the packet arrives at the egress node, the packet is delivered to the
216 network layer protocol, which then forwards it to its ultimate
217
218
219
220 <span class="grey">Jamoussi, et. al. Informational [Page 4]</span></pre>
221 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-5" ></span>
222 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
223
224
225 destination.
226
227 Tandem nodes have no interaction with the network layer protocols.
228 They only require knowledge of the VNS network topology. They make
229 their packet forwarding decision on the egress node identifier and
230 LN identifier carried in the VNS header of the packet.
231
232 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. VNS Header</span>
233
234 VNS defines a unicast header shown in Figure 2 and a multicast header
235 shown in Figure 3.
236
237 3 2 1 0
238 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
239 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
240 | TTL | LNN |x|LS-Key |x|DP | CmnHdr |
241 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
242 | Protocol Type | Destination Node Identifier |
243 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
244 | COS |x x x x| Source Node Identifier |
245 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
246 | Network Layer Header (e.g. IP) |
247 / /
248 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
249 | Data |
250 / /
251 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
252
253 Figure 2. Unicast VNS Header
254
255 The unicast header includes the following fields:
256
257 - Common Header (CmnHdr): The common header identifies the packet to
258 be a VNS encapsulated packet.
259
260 - Discard Priority: Indicates the level of congestion at which the
261 packet should be discarded. The value of this field is assigned on
262 the originating node based on policy information (see <a href="#section-9">Section 9</a>).
263
264 - Load Spreading Key: indicates the stream to which the packet
265 belongs for the purposes of equal cost multipath and trunk load
266 spreading (see <a href="#section-8">Section 8</a>).
267
268 - LNN: The Logical Network Number defines the logical network the
269 packet belongs to. This field in is used in conjunction with the
270 destination node identifier as the VNS switching label (see <a href="#section-5">Section</a>
271 <a href="#section-5">5</a>).
272
273
274
275
276 <span class="grey">Jamoussi, et. al. Informational [Page 5]</span></pre>
277 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-6" ></span>
278 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
279
280
281 - TTL: The Time To Live field is used to detect and discard packets
282 caught in temporary routing loops.
283
284 - Destination Node Identifier: This field contains an ID which
285 uniquely identifies the destination node. This ID is unique to the
286 physical network not just the LN. In conjunction with the LNN, this
287 forms a global VNS switching label.
288
289 - Protocol Type: indicates the type of Network layer protocol being
290 carried in the packet. Examples include IP, IPX, and Bridging. If the
291 packet is a multicast packet then this is indicated in this field.
292
293 - Source Node Identifier: This field contains an ID which uniquely
294 identifies the source node (ingress node).
295
296 - CoS: The Class of Service field is used to provide routing class of
297 service. The COS field also affects the Emission Priority of the
298 packet in the scheduler (see <a href="#section-9">Section 9</a>).
299
300 - Reserved Fields: All the fields marked with "x" are Reserved.
301
302 3 2 1 0
303 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
304 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
305 | TTL | LNN |x|LS-Key |x|DP | CmnHdr |
306 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
307 | PT = Multicast| Destination Node Identifier |
308 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
309 | COS |x x x x| Source Node Identifier |
310 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
311 | Protocol Type |x x x x x x x x| Multicast Group |
312 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
313 | Network Layer Header (e.g. IP) |
314 / /
315 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
316 / Data /
317 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
318
319 Figure 3. Multicast VNS Header
320
321 The multicast header shown in Figure 3, includes all the fields of
322 the unicast header. In addition, the multicast header includes the
323 following fields:
324
325 - Multicast Group: this field is used to identify a sub-group within
326 the logical network that receives the multicast packets.
327
328
329
330
331
332 <span class="grey">Jamoussi, et. al. Informational [Page 6]</span></pre>
333 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-7" ></span>
334 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
335
336
337 - Protocol Type: indicates the type of Network layer protocol being
338 carried in the packet. Examples include IP, IPX, and Bridging.
339
340 <span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a>. VNS Label Distribution</span>
341
342 Label distribution in VNS is based on a distributed serverless
343 topology driven approach. Standard ARP or address gleaning is used to
344 distribute and map network layer addresses to VNS addresses.
345
346 A VNS Label is an 6 byte encoding of the LNN and the node ID. VNS
347 Labels are treated as MAC addresses by the network layer. This means
348 that labels are distributed by the same means network layers use to
349 distribute MAC addresses. Thus, VNS leverages existing L2/L3 mapping
350 techniques and doesn't require a separate Label Distribution
351 Protocol.
352
353 <span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a>. Logical Networks (LNs)</span>
354
355 A logical network consists of a subset of the nodes in a network
356 together with a subset of the trunking facilities that link those
357 nodes. Logical networks partition the network into subnetworks that
358 serve a subset of the overall topology.
359
360 Each of the logical networks supported on any given node has a
361 separate routing and forwarding table (built by VNS). Therefore,
362 routing decisions are based on the resources available to the logical
363 network, not the entire network.
364
365 Each instance of VNS will discover all the trunks which are connected
366 to neighbors which support a matching LNN. This provides a huge
367 administrative saving, since VNS provisioning is on a per-node basis,
368 not on a per-link basis. VNS provisioning requires only a unique
369 node ID and an LNN. Discovery of which trunks support which LNNs is
370 done at run time, relieving administrative effort, and allowing the
371 LN to dynamically adapt to topology changes.
372
373 Multiple Logical Networks provide the following benefits to the
374 network system:
375
376 - Logical networks allow service providers to service multiple
377 private networks or (Virtual Private Internets) easily over one
378 network.
379
380 - Logical networks can be used to limit the impact of one network
381 layer protocol on the others. This is particularly true for
382 protocols that broadcast or multicast a large percentage of either
383 their control or data packets. This increases the effective
384 bandwidth of the trunks and allows the overall network to scale
385
386
387
388 <span class="grey">Jamoussi, et. al. Informational [Page 7]</span></pre>
389 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-8" ></span>
390 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
391
392
393 better.
394
395 - Logical networks allow for the configuration of the network to
396 meet individual community of interest and geographical
397 subnetworking needs.
398
399 - Routing control traffic has significance only in the local
400 subnetwork that is isolated to that subnetwork.
401
402 - Logical networks allow different instances of the same protocol
403 to share trunk facilities.
404
405 <span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a>. VNS Routing</span>
406
407 VNS routing is a link state routing system which uses many concepts
408 similar to OSPF and PNNI. One of the most significant departures from
409 the others is its ability to calculate shortest path trees for
410 routing unicast traffic and spanning trees for routing multicast
411 traffic within a Logical Network.
412
413 There is only one type of interface that VNS routing supports and
414 this is known as a VNS link. A link is a set of trunks that join two
415 VNS neighbor nodes. Each node in a VNS network maintains information
416 about the state of locally attached links. This information is
417 flooded throughout the network whenever there is a significant change
418 to the link's state or attributes (i.e. up/down, speed change,
419 available bandwidth change).
420
421 Each node stores and forwards the link state information received
422 from all other nodes. This allows each node to have the same view of
423 all of the nodes in the network together with all of their link state
424 information. This data is used to compute both the shortest path to
425 reach each node in the Logical Network and a spanning tree for the
426 Logical Network.
427
428 Logical networks are not bound to a particular trunk or link. They
429 are configured on a node. By default, a link will support a specific
430 logical network if the two nodes which it connects both are
431 configured to support the logical network number. This provides a
432 significant savings in operations over having to configure logical
433 networks on links or trunks.
434
435 When a link first comes into service, a protocol is run which allows
436 the two neighboring nodes to exchange information about the logical
437 networks they support. This allows the two nodes to determine if the
438 links are to be considered as a locally attached link for a logical
439 network.
440
441
442
443
444 <span class="grey">Jamoussi, et. al. Informational [Page 8]</span></pre>
445 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-9" ></span>
446 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
447
448
449 <span class="h2"><a class="selflink" id="section-7" href="#section-7">7</a>. VNS Forwarding</span>
450
451 VNS supports two types of forwarding: unicasting and multicasting. In
452 the first type, the data packet arrives on the ingress node and
453 unicasting forwards the data packet to a single destination (egress
454 node). In the second type, the data packet arrives on the ingress
455 node and multicasting forwards the data packet to all other nodes in
456 the logical network.
457
458 <span class="h3"><a class="selflink" id="section-7.1" href="#section-7.1">7.1</a> Unicast</span>
459
460 When a packet first enters the LAN internetwork, the network layer
461 routing protocol determines the next hop of the best route for the
462 packet to reach its final destination. If the best route is through a
463 VNS Logical Network, the network layer routing protocol relies on VNS
464 forwarding to get the packet to the egress node. A VNS packet header
465 containing the node ID (the unique ID assigned to each node) of the
466 egress node is added to the front of the packet and VNS forwarding is
467 invoked to deliver the packet. The network layer routing protocol
468 learns the egress node ID through an Address Resolution Protocol
469 (ARP) for IP and Source Address learning for bridging.
470
471 As the packet traverses the LN, routing decisions are made to
472 determine the next hop in the route to reach the destination node ID
473 specified in the VNS header. A forwarding table is built on each node
474 that assists in making the routing decision.
475
476 Each VNS instance on each node builds and maintains a forwarding
477 table for its LN. Each forwarding table has an entry for every node
478 that is a member of the logical network.
479
480 <span class="h3"><a class="selflink" id="section-7.2" href="#section-7.2">7.2</a> Multicast</span>
481
482 In addition to the unicast forwarding function, VNS also supports a
483 multicast forwarding service for traffic within an LN at the VNS
484 layer. Multicast packets are delivered to all nodes supporting the
485 logical network to which the multicast packet belongs. The packets
486 are sent along the branches of a spanning tree that is built by each
487 node supporting the logical network and is based on a common root
488 node (so that each node's view of the tree is the same as other
489 nodes). In other words, multicast packets are sent intelligently,
490 consuming a minimum of network bandwidth. If the network topology is
491 stable, each node receives each multicast packet only once.
492
493 Multicast packets received at any node are not acknowledged. They are
494 simply forwarded to the specified network layer interface and sent to
495 any other neighbor nodes on the spanning tree.
496
497
498
499
500 <span class="grey">Jamoussi, et. al. Informational [Page 9]</span></pre>
501 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-10" ></span>
502 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
503
504
505 <span class="h2"><a class="selflink" id="section-8" href="#section-8">8</a>. Traffic Engineering</span>
506
507 VNS forwarding supports two types of traffic engineering mechanisms:
508 equal cost multipaths and trunk load spreading.
509
510 Equal cost multipaths allows different streams (unique network layer
511 source and destination address pairings) to be load spread between
512 multiple relatively equal cost paths, through the Logical Network to
513 the egress node.
514
515 Trunk load spreading between two neighbors can take place when
516 multiple VNS trunks are defined between neighbors. Again, the load
517 spreading is based on network layer streams.
518
519 <span class="h3"><a class="selflink" id="section-8.1" href="#section-8.1">8.1</a> Equal Cost Multipaths</span>
520
521 From any point in a logical network, there may be multiple paths to
522 reach a specific egress node. If VNS routing determines that more
523 than one of these paths are of equal cost, VNS packets will be load
524 spread between two of them.
525
526 Equal cost multipath forwarding is supported not only on ingress
527 nodes but on tandem nodes as well. Each packet on an ingress node is
528 tagged with an equal cost multipath key. This key is acted upon at
529 the ingress node and stored in the VNS header to be used on tandem
530 nodes.
531
532 The equal cost multipath key is calculated by running an algorithm
533 over the source and destination network layer addresses. This means
534 that, in a stable network, any given stream will always take the same
535 path through a Logical Network avoiding the problems that misordering
536 would otherwise cause.
537
538 <span class="h3"><a class="selflink" id="section-8.2" href="#section-8.2">8.2</a> Trunk Load Spreading Between Neighbors</span>
539
540 VNS allows multiple trunks to be configured between neighboring VNS
541 nodes. VNS routing considers the aggregate bandwidth of those trunks
542 to determine the metric between the nodes. Also, VNS load spreads its
543 traffic amongst those trunks.
544
545 As is the case with equal cost multipaths, the trunk load spreading
546 key is calculated on the ingress node from an algorithm run over the
547 source and destination network layer addresses. The key is then
548 stored in the VNS header to be used on all tandem nodes through the
549 Logical Network.
550
551
552
553
554
555
556 <span class="grey">Jamoussi, et. al. Informational [Page 10]</span></pre>
557 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-11" ></span>
558 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
559
560
561 <span class="h2"><a class="selflink" id="section-9" href="#section-9">9</a>. Class of Service</span>
562
563 At the ingress to a VNS Network, packets are classified according to
564 the Class of Service (Cos) policy settings. The CoS differentiation
565 is achieved through different Emission and Discard priorities. The
566 semantics of the classification is carried in the VNS label (DP and
567 COS Fields described in <a href="#section-3">Section 3</a>) to be used at the ingress node as
568 well as all tandem points in the VNS network to affect queuing and
569 scheduling decisions.
570
571 <span class="h2"><a class="selflink" id="section-10" href="#section-10">10</a>. VNS Migration Strategies</span>
572
573 VNS supports several upper layer protocols such as IP, IPX, and
574 Bridging. Therefore, it is a multiprotocol label switching
575 architecture. In addition, VNS is not tied to a particular L2
576 technology. It runs on cell (e.g., ATM) trunks, frame trunks, or a
577 mixture of both.
578
579 VNS can be gradually introduced in a network. It can be implemented
580 between switching elements interconnected by point to point links.
581 Each of the switching nodes can run layer 3 routing simultaneously
582 with packet switching. VNS also allows for the interconnection of VNS
583 clouds through an ATM VC.
584
585 Since VNS can run on a mixture of Frame and Cell trunks, it allows
586 for the graceful migration of the frame links to ATM without
587 requiring a complete immediate overhaul.
588
589 <span class="h2"><a class="selflink" id="section-11" href="#section-11">11</a>. Summary</span>
590
591 VNS addresses scalability problems in several ways:
592
593 1. By a generally distributed design which doesn't
594 require a Label Distribution Protocol, or servers of any kind.
595 2. By providing an efficient, distributed multicast mechanism.
596 3. By allowing administrators to control the size of a
597 Logical Network, limiting traffic to a subset of the physical
598 topology.
599 4. By reducing layer 3 address space/subnet requirements in the
600 WAN which reduces the routing table size.
601
602 VNS provides redundancy transparent to the network layer protocol by
603 managing the network of trunks independently of the network layer.
604 VNS will automatically discover any topology changes and re-route
605 traffic accordingly.
606
607
608
609
610
611
612 <span class="grey">Jamoussi, et. al. Informational [Page 11]</span></pre>
613 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-12" ></span>
614 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
615
616
617 VNS eases network administration by dynamically keeping track of
618 which trunks are available for each LNN. Network administrators
619 don't have to configure VNS or network layer addresses on a per link
620 basis. Network layer addresses only have to be assigned on a per
621 Logical Network basis. For nodes which will only be tandem VNS
622 nodes, network layer addresses aren't required at all.
623
624 Since VNS traffic is constrained within an LNN, administrators have
625 control of where VNS traffic is allowed to flow.
626
627 Finally, VNS supports switching of several Upper Layer Protocols and
628 supports several media (cell and Frame) or a mixture thereof.
629 Switching in the core of the WAN removes the need for routers and
630 improves the performance due to a reduction in the number of fields
631 that need to processed.
632
633 <span class="h2"><a class="selflink" id="section-12" href="#section-12">12</a>. Security Considerations</span>
634
635 Logical networks provide a means of restricting traffic flow for
636 security purposes. VNS also relies on the inherent security of the L2
637 media such as an ATM Virtual Circuit.
638
639 <span class="h2"><a class="selflink" id="section-13" href="#section-13">13</a>. Acknowledgments</span>
640
641 The authors would like to acknowledge the valuable comments of Terry
642 Boland, Pierre Cousineau, Robert Eros, Robert Tomkins, and John
643 Whatman.
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668 <span class="grey">Jamoussi, et. al. Informational [Page 12]</span></pre>
669 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-13" ></span>
670 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
671
672
673 <span class="h2"><a class="selflink" id="section-14" href="#section-14">14</a>. Authors' Addresses</span>
674
675 Bilel Jamoussi
676 Nortel (Northern Telecom), Ltd.
677 PO Box 3511 Station C
678 Ottawa ON K1Y 4H7
679 Canada
680
681 EMail: [email protected]
682
683
684 Dwight Jamieson
685 Nortel (Northern Telecom), Ltd.
686 PO Box 3511 Station C
687 Ottawa ON K1Y 4H7
688 Canada
689
690 EMail: [email protected]
691
692
693 Dan Williston
694 Nortel (Northern Telecom), Ltd.
695 PO Box 3511 Station C
696 Ottawa ON K1Y 4H7
697 Canada
698
699 EMail: [email protected]
700
701
702 Stephen Gabe
703 Nortel (Northern Telecom), Ltd.
704 PO Box 3511 Station C
705 Ottawa ON K1Y 4H7
706 Canada
707
708 EMail: [email protected]
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724 <span class="grey">Jamoussi, et. al. Informational [Page 13]</span></pre>
725 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-14" ></span>
726 <span class="grey"><a href="./rfc2340">RFC 2340</a> Nortel's Virtual Network Switching (VNS) May 1998</span>
727
728
729 <span class="h2"><a class="selflink" id="section-15" href="#section-15">15</a>. Full Copyright Statement</span>
730
731 Copyright (C) The Internet Society (1998). All Rights Reserved.
732
733 This document and translations of it may be copied and furnished to
734 others, and derivative works that comment on or otherwise explain it
735 or assist in its implementation may be prepared, copied, published
736 and distributed, in whole or in part, without restriction of any
737 kind, provided that the above copyright notice and this paragraph are
738 included on all such copies and derivative works. However, this
739 document itself may not be modified in any way, such as by removing
740 the copyright notice or references to the Internet Society or other
741 Internet organizations, except as needed for the purpose of
742 developing Internet standards in which case the procedures for
743 copyrights defined in the Internet Standards process must be
744 followed, or as required to translate it into languages other than
745 English.
746
747 The limited permissions granted above are perpetual and will not be
748 revoked by the Internet Society or its successors or assigns.
749
750 This document and the information contained herein is provided on an
751 "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
752 TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
753 BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
754 HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
755 MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780 Jamoussi, et. al. Informational [Page 14]
781 </pre>