Debian Codegraph @

rfc5127.html

raw
1 <pre>Network Working Group K. Chan
2 Request for Comments: 5127 J. Babiarz
3 Category: Informational Nortel
4 F. Baker
5 Cisco Systems
6 February 2008
7
8
9 <span class="h1">Aggregation of Diffserv Service Classes</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 Abstract
18
19 In the core of a high-capacity network, service differentiation may
20 still be needed to support applications' utilization of the network.
21 Applications with similar traffic characteristics and performance
22 requirements are mapped into Diffserv service classes based on end-
23 to-end behavior requirements of the applications. However, some
24 network segments may be configured in such a way that a single
25 forwarding treatment may satisfy the traffic characteristics and
26 performance requirements of two or more service classes. In these
27 cases, it may be desirable to aggregate two or more Diffserv service
28 classes into a single forwarding treatment. This document provides
29 guidelines for the aggregation of Diffserv service classes into
30 forwarding treatments.
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
55
56
57 Table of Contents
58
59 <a href="#section-1">1</a>. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-3">3</a>
60 <a href="#section-1.1">1.1</a>. Requirements Notation . . . . . . . . . . . . . . . . . . <a href="#page-4">4</a>
61 <a href="#section-2">2</a>. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-4">4</a>
62 <a href="#section-3">3</a>. Overview of Service Class Aggregation . . . . . . . . . . . . <a href="#page-5">5</a>
63 <a href="#section-4">4</a>. Service Classes to Treatment Aggregate Mapping . . . . . . . . <a href="#page-6">6</a>
64 <a href="#section-4.1">4.1</a>. Mapping Service Classes into Four Treatment Aggregates . . <a href="#page-7">7</a>
65 <a href="#section-4.1.1">4.1.1</a>. Network Control Treatment Aggregate . . . . . . . . . <a href="#page-9">9</a>
66 <a href="#section-4.1.2">4.1.2</a>. Real-Time Treatment Aggregate . . . . . . . . . . . . <a href="#page-10">10</a>
67 <a href="#section-4.1.3">4.1.3</a>. Assured Elastic Treatment Aggregate . . . . . . . . . <a href="#page-10">10</a>
68 <a href="#section-4.1.4">4.1.4</a>. Elastic Treatment Aggregate . . . . . . . . . . . . . <a href="#page-12">12</a>
69 <a href="#section-5">5</a>. Treatment Aggregates and Inter-Provider Relationships . . . . <a href="#page-12">12</a>
70 <a href="#section-6">6</a>. Security Considerations . . . . . . . . . . . . . . . . . . . <a href="#page-13">13</a>
71 <a href="#section-7">7</a>. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-13">13</a>
72 <a href="#section-8">8</a>. References . . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-13">13</a>
73 <a href="#section-8.1">8.1</a>. Normative References . . . . . . . . . . . . . . . . . . . <a href="#page-13">13</a>
74 <a href="#section-8.2">8.2</a>. Informative References . . . . . . . . . . . . . . . . . . <a href="#page-14">14</a>
75 <a href="#appendix-A">Appendix A</a>. Using MPLS for Treatment Aggregates . . . . . . . . <a href="#page-15">15</a>
76 <a href="#appendix-A.1">A.1</a>. Network Control Treatment Aggregate with E-LSP . . . . . . <a href="#page-17">17</a>
77 <a href="#appendix-A.2">A.2</a>. Real-Time Treatment Aggregate with E-LSP . . . . . . . . . <a href="#page-17">17</a>
78 <a href="#appendix-A.3">A.3</a>. Assured Elastic Treatment Aggregate with E-LSP . . . . . . <a href="#page-17">17</a>
79 <a href="#appendix-A.4">A.4</a>. Elastic Treatment Aggregate with E-LSP . . . . . . . . . . <a href="#page-17">17</a>
80 <a href="#appendix-A.5">A.5</a>. Treatment Aggregates and L-LSP . . . . . . . . . . . . . . <a href="#page-18">18</a>
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
111
112
113 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Introduction</span>
114
115 In the core of a high capacity network, it is common for the network
116 to be engineered in such a way that a major link, switch, or router
117 can fail, and the result will be a routed network that still meets
118 ambient Service Level Agreements (SLAs). The implications are that
119 there is sufficient capacity on any given link such that all SLAs
120 sold can be simultaneously supported at their respective maximum
121 rates, and that this remains true after re-routing (either IP re-
122 routing or Multiprotocol Label Switching (MPLS) protection-mode
123 switching) has occurred.
124
125 Over-provisioning is generally considered to meet the requirements of
126 all traffic without further quality of service (QoS) treatment, and
127 in the general case, that is true in high-capacity backbones.
128 However, as the process of network convergence continues, and with
129 the increasing speed of the access networks, certain services may
130 still have issues. Delay, jitter, and occasional loss are perfectly
131 acceptable for elastic applications. However, sub-second surges that
132 occur in the best-designed of networks [<a href="#ref-12" title="&quot;Analysis of Point-To-Point Packet Delay in an Operational Network&quot;">12</a>] affect real-time
133 applications. Moreover, denial of service (DoS) loads, worms, and
134 network disruptions such as that of 11 September 2001 affect routing
135 [<a href="#ref-13" title="&quot;Internet Routing Behavior on 9/11&quot;">13</a>]. Our objective is to prevent disruption to routing (which in
136 turn affects all services) and to protect real-time jitter-sensitive
137 services, while minimizing loss and delay of sensitive elastic
138 traffic.
139
140 <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>] defines a set of basic Diffserv classes from the points
141 of view of the application requiring specific end-to-end behaviors
142 from the network. The service classes are differentiated based on
143 the application payload's tolerance to packet loss, delay, and delay
144 variation (jitter). Different degrees of these criteria form the
145 foundation for supporting the needs of real-time and elastic traffic.
146 <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>] also provides recommendations for the treatment method
147 of these service classes. But, at some network segments of the end-
148 to-end path, the number of levels of network treatment
149 differentiation may be less than the number of service classes that
150 the network segment needs to support. In such a situation, that
151 network segment may use the same treatment to support more than one
152 service class. In this document, we provide guidelines on how
153 multiple service classes may be aggregated into a forwarding
154 treatment aggregate. This entails having the IP traffic belonging to
155 service classes, expressed using the DSCP (Differentiated Services
156 Code Point), as described by <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>]. Note that in a given
157 domain, we may recommend that the supported service classes be
158 aggregated into forwarding treatment aggregates; however, this does
159 not mean all service classes need to be supported, and hence not all
160 forwarding treatment aggregates need to be supported. A domain may
161
162
163
164 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
167
168
169 support a fewer or greater number of forwarding treatment aggregates
170 than recommended by this document. Which service classes and which
171 forwarding treatment aggregates are supported by a domain is up to
172 the domain administration and may be influenced by business reasons
173 or other reasons (e.g., operational considerations).
174
175 In this document, we've provided:
176
177 o definitions for terminology we use in this document,
178
179 o requirements for performing this aggregation,
180
181 o an example of performing the aggregation when four treatment
182 aggregates are used, and
183
184 o an example (in the appendix) of performing this aggregation over
185 MPLS using E-LSP, EXP Inferred PHB Scheduling Class (PSC) Label
186 Switched Path (LSP).
187
188 The treatment aggregate recommendations are designed to aggregate the
189 service classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>] in such a manner as to protect real-time traffic
190 and routing, on the assumption that real-time sessions are protected
191 from each other by admission at the edge. The recommendation given
192 is one possible way of performing the aggregation; there may be other
193 ways of aggregation, for example, into fewer treatment aggregates or
194 more treatment aggregates.
195
196 In the appendix, an example of aggregation over MPLS networks using
197 E-LSP to realize the treatment aggregates is provided. Note that the
198 MPLS E-LSP is just an example; this document does not exclude the use
199 of other methods. This example only considers aggregation of IP
200 traffic into E-LSP. The use of E-LSP by non-IP traffic is not
201 discussed.
202
203 <span class="h3"><a class="selflink" id="section-1.1" href="#section-1.1">1.1</a>. Requirements Notation</span>
204
205 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
206 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
207 document are to be interpreted as described in <a href="./rfc2119">RFC 2119</a> [<a href="#ref-1" title="&quot;Key words for use in RFCs to Indicate Requirement Levels&quot;">1</a>].
208
209 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Terminology</span>
210
211 This document assumes the reader is familiar with the terms used in
212 differentiated services. This document provides the definitions for
213 new terms introduced by this document and references information
214 defined in RFCs for existing terms not commonly used in
215 differentiated services.
216
217
218
219
220 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
223
224
225 For new terms introduced by this document, we provide the definition
226 here:
227
228 o Treatment Aggregate. This term is defined as the aggregate of
229 Diffserv service classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>]. A treatment aggregate is concerned
230 only with the forwarding treatment of the aggregated traffic,
231 which may be marked with multiple DSCPs. A treatment aggregate
232 differs from Behavior Aggregate [<a href="#ref-2" title="&quot;Definition of the Differentiated Services Field (DS Field) in the IPv4 and IPv6 Headers&quot;">2</a>] and Traffic Aggregate [<a href="#ref-14" title="&quot;Definition of Differentiated Services Per Domain Behaviors and Rules for their Specification&quot;">14</a>],
233 each of which indicate the aggregated traffic having a single
234 Diffserv codepoint and utilizing a single Per Hop Behavior (PHB).
235
236 For terms from existing RFCs, we provide the reference to the
237 appropriate section of the relevant RFC that contain the definition:
238
239 o Real-Time and Elastic Applications and their traffic. <a href="./rfc1633#section-3.1">Section&nbsp;3.1
240 of RFC 1633</a> [<a href="#ref-4" title="&quot;Integrated Services in the Internet Architecture: an Overview&quot;">4</a>].
241
242 o Diffserv Service Class. <a href="./rfc4594#section-1.3">Section&nbsp;1.3 of RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>].
243
244 o MPLS E-LSP, EXP Inferred PHB Scheduling Class (PSC) Label Switched
245 Path (LSP). <a href="./rfc3270#section-1.2">Section&nbsp;1.2 of RFC 3270</a> [<a href="#ref-6" title="&quot;Multi-Protocol Label Switching (MPLS) Support of Differentiated Services&quot;">6</a>].
246
247 o MPLS L-LSP, Label Only Inferred PHB Scheduling Class (PSC) Label
248 Switched Path (LSP). <a href="./rfc3270#section-1.3">Section&nbsp;1.3 of RFC 3270</a> [<a href="#ref-6" title="&quot;Multi-Protocol Label Switching (MPLS) Support of Differentiated Services&quot;">6</a>].
249
250 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. Overview of Service Class Aggregation</span>
251
252 In Diffserv domains where less fine-grained traffic treatment
253 differentiation is provided, aggregation of the different service
254 classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>] may be required.
255
256 These aggregations have the following requirements:
257
258 1. The end-to-end network performance characteristic required by the
259 application MUST be supported. This performance characteristic
260 is represented by the use of Diffserv service classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>].
261
262 2. The treatment aggregate MUST meet the strictest requirements of
263 its member service classes.
264
265 3. The treatment aggregate SHOULD only contain member service
266 classes with similar traffic characteristic and performance
267 requirements.
268
269 4. The notion of the individual end-to-end service classes MUST NOT
270 be destroyed when aggregation is performed. Each domain along
271 the end-to-end path may perform aggregation differently, based on
272 the original end-to-end service classes. We recommend an easy
273
274
275
276 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
279
280
281 way to accomplish this by not altering the DSCP used to indicate
282 the end-to-end service class. But some administrative domains
283 may require the use of their own marking; when this is needed,
284 the original end-to-end service class indication must be restored
285 upon exiting such administrative domains. One possible way of
286 achieving this is with the use of tunnels to encapsulate the end-
287 to-end traffic.
288
289 5. Each treatment aggregate has limited resources; hence, traffic
290 conditioning and/or admission control SHOULD be performed for
291 each service class aggregated into the treatment aggregate.
292 Additional admission control and policing may be used on the sum
293 of all traffic aggregated into the treatment aggregate.
294
295 In addition to the above requirements, we have the following
296 suggestions:
297
298 1. The treatment aggregate and assigned resources may consider
299 historical traffic patterns and the variability of these
300 patterns. For example, a point-point service (e.g., pseudowire)
301 may have a very predictable pattern, while a multipoint service
302 (e.g., VPLS, Virtual Private LAN Service) may have a much less
303 predictable pattern.
304
305 2. In addition to Diffserv, other controls are available to
306 influence the traffic level offered to a particular traffic
307 aggregate. These include adjustment of routing metrics, and
308 usage of MPLS-based traffic engineering techniques.
309
310 This document only describes the aggregation of IP traffic based on
311 the use of Diffserv service classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>].
312
313 <span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a>. Service Classes to Treatment Aggregate Mapping</span>
314
315 The service class and DSCP selection in <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>] has been defined
316 to allow, in many instances, mapping of two or possibly more service
317 classes into a single forwarding treatment aggregate. Notice that
318 there is a relationship/trade-off between link speed, queue depth,
319 delay, and jitter. The degree of aggregation and hence the number of
320 treatment aggregates will depend on the aggregation's impacts on
321 loss, delay, and jitter. This depends on whether the speed of the
322 links and scheduler behavior, being used to implement the
323 aggregation, can minimize the effects of mixing traffic with
324 different packet sizes and transmit rates on queue depth. A general
325 rule-of-thumb is that higher link speeds allow for more aggregation/
326 smaller number of treatment aggregates, assuming link utilization is
327 within the engineered level.
328
329
330
331
332 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
335
336
337 <span class="h3"><a class="selflink" id="section-4.1" href="#section-4.1">4.1</a>. Mapping Service Classes into Four Treatment Aggregates</span>
338
339 This section provides an example of mapping all the service classes
340 defined in <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>] into four treatment aggregates. The use of
341 four treatment aggregates assumes that the resources allocated to
342 each treatment aggregate are sufficient to honor the required
343 behavior of each service class [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>]. We use the performance
344 requirement (tolerance to loss, delay, and jitter) from the
345 application/end-user as a guide on how to map the service classes
346 into treatment aggregates. We have also used <a href="./rfc1633#section-3.1">section&nbsp;3.1 of RFC 1633</a>
347 [<a href="#ref-4" title="&quot;Integrated Services in the Internet Architecture: an Overview&quot;">4</a>] to provide us with guidance on the definition of Real-Time and
348 Elastic applications. An overview of the mapping between service
349 classes and the four treatment aggregates is provided by Figure 1,
350 with the mapping being based on performance requirements. In Figure
351 1, the right side columns of "Service Class" and "Tolerance to Loss/
352 Delay/Jitter" are from Figure 2 of <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>].
353
354 It is recommended that certain service classes be mapped into
355 specific treatment aggregates. But this does not mean that all the
356 service classes recommended for that treatment aggregate need to be
357 supported. Hence, for a given domain, a treatment aggregate may
358 contain only a subset of the service classes recommended in this
359 document, i.e., the service classes supported by that domain. A
360 domain's treatment of non-supported service classes should be based
361 on the domain's local policy. This local policy may be influenced by
362 its agreement with its customers. Such treatment may use the Elastic
363 Treatment Aggregate, dropping the packets, or some other
364 arrangements.
365
366 Our example of four treatment aggregates is based on the basic
367 differences in performance requirement from the application/end-user
368 perspective. A domain may choose to support more or fewer treatment
369 aggregates than the four recommended. For example, a domain may
370 support only three treatment aggregates and map any network control
371 traffic into the Assured Elastic treatment aggregate. This is a
372 choice the administrative domain has. Hence, this example of four
373 treatment aggregates does not represent a minimum required set of
374 treatment aggregates one must implement; nor does it represent the
375 maximum set of treatment aggregates one can implement.
376
377
378
379
380
381
382
383
384
385
386
387
388 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
391
392
393 ---------------------------------------------------------------------
394 |Treatment | Tolerance to ||Service Class | Tolerance to |
395 |Aggregate | Loss |Delay |Jitter|| | Loss |Delay |Jitter|
396 |==========+======+======+======++===============+======+======+======|
397 | Network | Low | Low | Yes || Network | Low | Low | Yes |
398 | Control | | | || Control | | | |
399 |==========+======+======+======++===============+======+======+======|
400 | Real- | Very | Very | Very || Telephony | VLow | VLow | VLow |
401 | Time | Low | Low | Low ||---------------+------+------+------|
402 | | | | || Signaling | Low | Low | Yes |
403 | | | | ||---------------+------+------+------|
404 | | | | || Multimedia |Low - | Very | Low |
405 | | | | || Conferencing |Medium| Low | |
406 | | | | ||---------------+------+------+------|
407 | | | | || Real-time | Low | Very | Low |
408 | | | | || Interactive | | Low | |
409 | | | | ||---------------+------+------+------|
410 | | | | || Broadcast | Very |Medium| Low |
411 | | | | || Video | Low | | |
412 |==========+======+======+======++===============+======+======+======|
413 | Assured | Low |Low - | Yes || Multimedia |Low - |Medium| Yes |
414 | Elastic | |Medium| || Streaming |Medium| | |
415 | | | | ||---------------+------+------+------|
416 | | | | || Low-Latency | Low |Low - | Yes |
417 | | | | || Data | |Medium| |
418 | | | | ||---------------+------+------+------|
419 | | | | || OAM | Low |Medium| Yes |
420 | | | | ||---------------+------+------+------|
421 | | | | ||High-Throughput| Low |Medium| Yes |
422 | | | | || Data | |- High| |
423 |==========+======+======+======++===============+======+======+======|
424 | Elastic | Not Specified || Standard | Not Specified |
425 | | | | ||---------------+------+------+------|
426 | | | | || Low-Priority | High | High | Yes |
427 | | | | || Data | | | |
428 ---------------------------------------------------------------------
429
430 Figure 1: Treatment Aggregate and Service Class Performance
431 Requirements
432
433 As we are recommending to preserve the notion of the individual end-
434 to-end service classes, we also recommend that the original DSCP
435 field marking not be changed when treatment aggregates are used.
436 Instead, classifiers that select packets based on the contents of the
437 DSCP field should be used to direct packets from the member Diffserv
438 service classes into the queue that handles each of the treatment
439 aggregates, without remarking the DSCP field of the packets. This is
440
441
442
443
444 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
447
448
449 summarized in Figure 2, which shows the behavior each treatment
450 aggregate should have, and the DSCP field marking of the packets that
451 should be classified into each of the treatment aggregates.
452
453 ------------------------------------------------------------
454 |Treatment |Treatment || DSCP |
455 |Aggregate |Aggregate || |
456 | |Behavior || |
457 |==========+==========++=====================================|
458 | Network | CS || CS6 |
459 | Control |(<a href="./rfc2474">RFC 2474</a>)|| |
460 |==========+==========++=====================================|
461 | Real- | EF || EF, CS5, AF41, AF42, AF43, CS4, CS3 |
462 | Time |(<a href="./rfc3246">RFC 3246</a>)|| |
463 |==========+==========++=====================================|
464 | Assured | AF || CS2, AF31, AF21, AF11 |
465 | Elastic |(<a href="./rfc2597">RFC 2597</a>)||-------------------------------------|
466 | | || AF32, AF22, AF12 |
467 | | ||-------------------------------------|
468 | | || AF33, AF23, AF13 |
469 |==========+==========++=====================================|
470 | Elastic | Default || Default, (CS0) |
471 | |(<a href="./rfc2474">RFC 2474</a>)||-------------------------------------|
472 | | || CS1 |
473 ------------------------------------------------------------
474
475 Figure 2: Treatment Aggregate Behavior
476
477 Notes for Figure 2: For Assured Elastic and Elastic Treatment
478 Aggregates, please see sections <a href="#section-4.1.3">4.1.3</a> and <a href="#section-4.1.4">4.1.4</a>, respectively, for
479 details on additional priority within the treatment aggregate.
480
481 <span class="h4"><a class="selflink" id="section-4.1.1" href="#section-4.1.1">4.1.1</a>. Network Control Treatment Aggregate</span>
482
483 The Network Control Treatment Aggregate aggregates all service
484 classes that are functionally necessary for the survival of a network
485 during a DoS attack or other high-traffic load interval. The theory
486 is that whatever else is true, the network must protect itself. This
487 includes the traffic that <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>] characterizes as being
488 included in the Network Control service class.
489
490 Traffic in the Network Control Treatment Aggregate should be carried
491 in a common queue or class with a PHB as described in <a href="./rfc2474">RFC 2474</a> [<a href="#ref-2" title="&quot;Definition of the Differentiated Services Field (DS Field) in the IPv4 and IPv6 Headers&quot;">2</a>],
492 section 4.2.2.2 for Class Selector (CS). This treatment aggregate
493 should have a lower probability of packet loss and bear a relatively
494 deep target mean queue depth (min-threshold if RED (Random Early
495 Detection) is being used).
496
497
498
499
500 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
503
504
505 Please notice this Network Control Treatment Aggregate is meant to be
506 used for the customer's network control traffic. The provider may
507 choose to treat its own network control traffic differently, perhaps
508 in its own service class that is not aggregated with the customer's
509 network control traffic.
510
511 <span class="h4"><a class="selflink" id="section-4.1.2" href="#section-4.1.2">4.1.2</a>. Real-Time Treatment Aggregate</span>
512
513 The Real-Time Treatment Aggregate aggregates all real-time
514 (inelastic) service classes. The theory is that real-time traffic is
515 admitted under some model and controlled by an SLA managed at the
516 edge of the network prior to aggregation. As such, there is a
517 predictable and enforceable upper bound on the traffic that can enter
518 such a queue, and to provide predictable variation in delay it must
519 be protected from bursts of elastic traffic. The predictability of
520 traffic level may be based upon admission control for a well-known
521 community of interest (e.g., a point-point service) and/or based upon
522 historical measurements.
523
524 This treatment aggregate may include the following service classes
525 from the Diffserv service classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>], in addition to other locally
526 defined classes: Telephony, Signaling, Multimedia Conferencing, Real-
527 time Interactive, and Broadcast Video.
528
529 Traffic in each service class that is going to be aggregated into the
530 treatment aggregate should be conditioned prior to aggregation. It
531 is recommended that per-service-class admission control procedures be
532 used, followed by per-service-class policing so that any individual
533 service class does not generate more than what it is allowed.
534 Furthermore, additional admission control and policing may be used on
535 the sum of all traffic aggregated into this treatment aggregate.
536
537 Traffic in the Real-Time Treatment Aggregate should be carried in a
538 common queue or class with a PHB (Per Hop Behavior) as described in
539 <a href="./rfc3246">RFC 3246</a> [<a href="#ref-9" title="&quot;An Expedited Forwarding PHB (Per-Hop Behavior)&quot;">9</a>] and <a href="./rfc3247">RFC 3247</a> [<a href="#ref-10" title="&quot;Supplemental Information for the New Definition of the EF PHB (Expedited Forwarding Per-Hop Behavior)&quot;">10</a>].
540
541 <span class="h4"><a class="selflink" id="section-4.1.3" href="#section-4.1.3">4.1.3</a>. Assured Elastic Treatment Aggregate</span>
542
543 The Assured Elastic Treatment Aggregate aggregates all elastic
544 traffic that uses the Assured Forwarding model as described in <a href="./rfc2597">RFC</a>
545 <a href="./rfc2597">2597</a> [<a href="#ref-8" title="&quot;Assured Forwarding PHB Group&quot;">8</a>]. The premise of such a service is that an SLA that is
546 negotiated includes a "committed rate" and the ability to exceed that
547 rate (and perhaps a second "excess rate") in exchange for a higher
548 probability of loss using Active Queue Management (AQM) [<a href="#ref-7" title="&quot;Recommendations on Queue Management and Congestion Avoidance in the Internet&quot;">7</a>] or
549 Explicit Congestion Notification (ECN) marking [<a href="#ref-11" title="&quot;The Addition of Explicit Congestion Notification (ECN) to IP&quot;">11</a>] for the portion
550 of traffic deemed to be in excess.
551
552
553
554
555
556 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
559
560
561 This treatment aggregate may include the following service classes
562 from the Diffserv service classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>], in addition to other locally
563 defined classes: Multimedia Streaming, Low Latency Data, OAM, and
564 High-Throughput Data.
565
566 The DSCP values belonging to the Assured Forwarding (AF) PHB group
567 and class selector of the original service classes remain an
568 important consideration and should be preserved during aggregation.
569 This treatment aggregate should maintain the AF PHB group marking of
570 the original packet. For example, AF3x marked packets should remain
571 AF3x marked within this treatment aggregate. In addition, the class
572 selector DSCP value should not be changed. Traffic bearing these
573 DSCPs is carried in a common queue or class with a PHB as described
574 in <a href="./rfc2597">RFC 2597</a> [<a href="#ref-8" title="&quot;Assured Forwarding PHB Group&quot;">8</a>]. In effect, appropriate target rate thresholds have
575 been applied at the edge, dividing traffic into AFn1 (committed, for
576 any value of n), AFn2, and AFn3 (excess). The service should be
577 engineered so that AFn1 and CS2 marked packet flows have sufficient
578 bandwidth in the network to provide high assurance of delivery.
579 Since the traffic is elastic and responds dynamically to packet loss,
580 Active Queue Management [<a href="#ref-7" title="&quot;Recommendations on Queue Management and Congestion Avoidance in the Internet&quot;">7</a>] should be used primarily to reduce the
581 forwarding rate to the minimum assured rate at congestion points.
582 The probability of loss of AFn1 and CS2 traffic must not exceed the
583 probability of loss of AFn2 traffic, which in turn must not exceed
584 the probability of loss of AFn3 traffic.
585
586 If RED [<a href="#ref-7" title="&quot;Recommendations on Queue Management and Congestion Avoidance in the Internet&quot;">7</a>] is used as an AQM algorithm, the min-threshold specifies a
587 target queue depth for each of AFn1+CS2, AFn2, and AFn3, and the max-
588 threshold specifies the queue depth above which all traffic with such
589 a DSCP is dropped or ECN marked. Thus, in this treatment aggregate,
590 the following inequalities SHOULD hold in queue configurations:
591
592 o min-threshold AFn3 &lt; max-threshold AFn3
593
594 o max-threshold AFn3 &lt;= min-threshold AFn2
595
596 o min-threshold AFn2 &lt; max-threshold AFn2
597
598 o max-threshold AFn2 &lt;= min-threshold AFn1+CS2
599
600 o min-threshold AFn1+CS2 &lt; max-threshold AFn1+CS2
601
602 o max-threshold AFn1+CS2 &lt;= memory assigned to the queue
603
604 Note: This configuration tends to drop AFn3 traffic before AFn2, and
605 AFn2 before AFn1 and CS2. Many other AQM algorithms exist and are
606 used; they should be configured to achieve a similar result.
607
608
609
610
611
612 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
615
616
617 <span class="h4"><a class="selflink" id="section-4.1.4" href="#section-4.1.4">4.1.4</a>. Elastic Treatment Aggregate</span>
618
619 The Elastic Treatment Aggregate aggregates all remaining elastic
620 traffic. The premise of such a service is that there is no intrinsic
621 SLA differentiation of traffic, but that AQM [<a href="#ref-7" title="&quot;Recommendations on Queue Management and Congestion Avoidance in the Internet&quot;">7</a>] or ECN flagging [<a href="#ref-11" title="&quot;The Addition of Explicit Congestion Notification (ECN) to IP&quot;">11</a>]
622 is appropriate for such traffic.
623
624 This treatment aggregate may include the following service classes
625 from the Diffserv service classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>], in addition to other locally
626 defined classes: Standard and Low-Priority Data.
627
628 Treatment aggregates should be well specified, each indicating the
629 service classes it will handle. But in cases where unspecified or
630 unknown service classes are encountered, they may be dropped or be
631 treated using the Elastic Treatment Aggregate. The choice of how to
632 treat unspecified service classes should be well defined, based on
633 some agreements.
634
635 Traffic in the Elastic Treatment Aggregate should be carried in a
636 common queue or class with a PHB as described in <a href="./rfc2474">RFC 2474</a> [<a href="#ref-2" title="&quot;Definition of the Differentiated Services Field (DS Field) in the IPv4 and IPv6 Headers&quot;">2</a>],
637 section 4.1, "A Default PHB". The AQM thresholds for Elastic traffic
638 MAY be separately set, so that Low Priority Data traffic is dropped
639 before Standard traffic, but this is not a requirement.
640
641 <span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a>. Treatment Aggregates and Inter-Provider Relationships</span>
642
643 When treatment aggregates are used at provider boundaries, we
644 recommend that the inter-provider relationship be based on Diffserv
645 service classes [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>]. This allows the admission control into each
646 treatment aggregate of a provider domain to be based on the admission
647 control of traffic into the supported service classes, as indicated
648 by the discussion in <a href="#section-4">section 4</a> of this document.
649
650 If the inter-provider relationship needs to be based on treatment
651 aggregates specified by this document, then the exact treatment
652 aggregate content and representation must be agreed to by the peering
653 providers.
654
655 Some additional work on inter-provider relationships is provided by
656 inter-provider QoS [<a href="#ref-15" title="&quot;Inter-provider Quality of Service&quot;">15</a>], where details on supporting real-time
657 services between service providers are discussed. Some related work
658 in ITU-T provided by <a href="#appendix-V">Appendix V</a>I of Y.1541 [<a href="#ref-16" title="&quot;Network Performance Objectives for IP-Based Services&quot;">16</a>] may also help with
659 inter-provider relationships, especially with international
660 providers.
661
662
663
664
665
666
667
668 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
671
672
673 <span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a>. Security Considerations</span>
674
675 This document discusses the policy of using Differentiated Services
676 and its service classes. If implemented as described, it should
677 require that the network do nothing that the network has not already
678 allowed. If that is the case, no new security issues should arise
679 from the use of such a policy.
680
681 As this document is based on <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>], the Security Consideration
682 discussion of no new security issues indicated by <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>] also
683 applies to treatment aggregates of this document.
684
685 <span class="h2"><a class="selflink" id="section-7" href="#section-7">7</a>. Acknowledgements</span>
686
687 This document has benefited from discussions with numerous people,
688 especially Shane Amante, Brian Carpenter, and Dave McDysan. It has
689 also benefited from detailed reviews by David Black, Marvin Krym,
690 Bruce Davie, Fil Dickinson, and Julie Ann Connary.
691
692 <span class="h2"><a class="selflink" id="section-8" href="#section-8">8</a>. References</span>
693
694 <span class="h3"><a class="selflink" id="section-8.1" href="#section-8.1">8.1</a>. Normative References</span>
695
696 [<a id="ref-1">1</a>] Bradner, S., "Key words for use in RFCs to Indicate Requirement
697 Levels", <a href="https://www.rfc-editor.org/bcp/bcp14">BCP 14</a>, <a href="./rfc2119">RFC 2119</a>, March 1997.
698
699 [<a id="ref-2">2</a>] Nichols, K., Blake, S., Baker, F., and D. Black, "Definition of
700 the Differentiated Services Field (DS Field) in the IPv4 and
701 IPv6 Headers", <a href="./rfc2474">RFC 2474</a>, December 1998.
702
703 [<a id="ref-3">3</a>] Babiarz, J., Chan, K., and F. Baker, "Configuration Guidelines
704 for DiffServ Service Classes", <a href="./rfc4594">RFC 4594</a>, August 2006.
705
706 [<a id="ref-4">4</a>] Braden, B., Clark, D., and S. Shenker, "Integrated Services in
707 the Internet Architecture: an Overview", <a href="./rfc1633">RFC 1633</a>, June 1994.
708
709 [<a id="ref-5">5</a>] Black, D., "Differentiated Services and Tunnels", <a href="./rfc2983">RFC 2983</a>,
710 October 2000.
711
712 [<a id="ref-6">6</a>] Le Faucheur, F., Wu, L., Davie, B., Davari, S., Vaananen, P.,
713 Krishnan, R., Cheval, P., and J. Heinanen, "Multi-Protocol
714 Label Switching (MPLS) Support of Differentiated Services",
715 <a href="./rfc3270">RFC 3270</a>, May 2002.
716
717
718
719
720
721
722
723
724 <span class="grey">Chan, 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="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
727
728
729 [<a id="ref-7">7</a>] Braden, B., Clark, D., Crowcroft, J., Davie, B., Deering, S.,
730 Estrin, D., Floyd, S., Jacobson, V., Minshall, G., Partridge,
731 C., Peterson, L., Ramakrishnan, K., Shenker, S., Wroclawski,
732 J., and L. Zhang, "Recommendations on Queue Management and
733 Congestion Avoidance in the Internet", <a href="./rfc2309">RFC 2309</a>, April 1998.
734
735 [<a id="ref-8">8</a>] Heinanen, J., Baker, F., Weiss, W., and J. Wroclawski, "Assured
736 Forwarding PHB Group", <a href="./rfc2597">RFC 2597</a>, June 1999.
737
738 [<a id="ref-9">9</a>] Davie, B., Charny, A., Bennet, J., Benson, K., Le Boudec, J.,
739 Courtney, W., Davari, S., Firoiu, V., and D. Stiliadis, "An
740 Expedited Forwarding PHB (Per-Hop Behavior)", <a href="./rfc3246">RFC 3246</a>,
741 March 2002.
742
743 [<a id="ref-10">10</a>] Charny, A., Bennet, J., Benson, K., Boudec, J., Chiu, A.,
744 Courtney, W., Davari, S., Firoiu, V., Kalmanek, C., and K.
745 Ramakrishnan, "Supplemental Information for the New Definition
746 of the EF PHB (Expedited Forwarding Per-Hop Behavior)",
747 <a href="./rfc3247">RFC 3247</a>, March 2002.
748
749 [<a id="ref-11">11</a>] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition of
750 Explicit Congestion Notification (ECN) to IP", <a href="./rfc3168">RFC 3168</a>,
751 September 2001.
752
753 <span class="h3"><a class="selflink" id="section-8.2" href="#section-8.2">8.2</a>. Informative References</span>
754
755 [<a id="ref-12">12</a>] Choi, B., Moon, S., Zhang, Z., Papagiannaki, K., and C. Diot,
756 "Analysis of Point-To-Point Packet Delay in an Operational
757 Network", INFOCOMM 2004, March 2004,
758 &lt;<a href="http://www.ieee-infocom.org/2004/Papers/37_4.PDF">http://www.ieee-infocom.org/2004/Papers/37_4.PDF</a>&gt;.
759
760 [<a id="ref-13">13</a>] Ogielski, A. and J. Cowie, "Internet Routing Behavior on 9/11",
761 March 2002, &lt;<a href="http://www.renesys.com/tech/presentations/pdf/renesys-030502-NRC-911.pdf">http://www.renesys.com/tech/presentations/pdf/</a>
762 <a href="http://www.renesys.com/tech/presentations/pdf/renesys-030502-NRC-911.pdf">renesys-030502-NRC-911.pdf</a>&gt;.
763
764 [<a id="ref-14">14</a>] Nichols, K. and B. Carpenter, "Definition of Differentiated
765 Services Per Domain Behaviors and Rules for their
766 Specification", <a href="./rfc3086">RFC 3086</a>, April 2001.
767
768 [<a id="ref-15">15</a>] MIT Communications Futures Program, "Inter-provider Quality of
769 Service", November 2006, &lt;
770 <a href="http://cfp.mit.edu/resources/papers/Interprovider">http://cfp.mit.edu/resources/papers/Interprovider</a> QoS
771 MIT_CFP_WP_9_14_06.pdf&gt;.
772
773 [<a id="ref-16">16</a>] International Telecommunications Union, "Network Performance
774 Objectives for IP-Based Services", Recommendation Y.1541,
775 February 2006.
776
777
778
779
780 <span class="grey">Chan, et al. Informational [Page 14]</span></pre>
781 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-15" ></span>
782 <span class="grey"><a href="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
783
784
785 <span class="h2"><a class="selflink" id="appendix-A" href="#appendix-A">Appendix A</a>. Using MPLS for Treatment Aggregates</span>
786
787 <a href="./rfc2983">RFC 2983</a> on Diffserv and Tunnels [<a href="#ref-5" title="&quot;Differentiated Services and Tunnels&quot;">5</a>] and <a href="./rfc3270">RFC 3270</a> on MPLS Support of
788 Diffserv [<a href="#ref-6" title="&quot;Multi-Protocol Label Switching (MPLS) Support of Differentiated Services&quot;">6</a>] provide a very good background on this topic. This
789 document provides an example of using the E-LSP, EXP Inferred PHB
790 Scheduled Class (PSC) Label Switched Path (LSP), defined by MPLS
791 Support of Diffserv [<a href="#ref-6" title="&quot;Multi-Protocol Label Switching (MPLS) Support of Differentiated Services&quot;">6</a>] for realizing the Treatment Aggregates.
792
793 When treatment aggregates are represented in MPLS using EXP Inferred
794 PSC LSP, we recommend the following usage of the MPLS EXP field for
795 treatment aggregates.
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836 <span class="grey">Chan, et al. Informational [Page 15]</span></pre>
837 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-16" ></span>
838 <span class="grey"><a href="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
839
840
841 -------------------------------------------
842 |Treatment || MPLS || DSCP | DSCP |
843 |Aggregate || EXP || name | value |
844 |==========++======++=========|=============|
845 | Network || 110 || CS6 | 110000 |
846 | Control || || | |
847 |==========++======++=========|=============|
848 | Real- || 100 || EF | 101110 |
849 | Time || ||---------|-------------|
850 | || || CS5 | 101000 |
851 | || ||---------|-------------|
852 | || ||AF41,AF42|100010,100100|
853 | || || AF43 | 100110 |
854 | || ||---------|-------------|
855 | || || CS4 | 100000 |
856 | || ||---------|-------------|
857 | || || CS3 | 011000 |
858 |==========++======++=========|=============|
859 | Assured || 010* || CS2 | 010000 |
860 | Elastic || || AF31 | 011010 |
861 | || || AF21 | 010010 |
862 | || || AF11 | 001010 |
863 | ||------||---------|-------------|
864 | || 011* || AF32 | 011100 |
865 | || || AF22 | 010100 |
866 | || || AF12 | 001100 |
867 | || || AF33 | 011110 |
868 | || || AF23 | 010110 |
869 | || || AF13 | 001110 |
870 |==========++======++=========|=============|
871 | Elastic || 000* || Default | 000000 |
872 | || || (CS0) | |
873 | ||------||---------|-------------|
874 | || 001* || CS1 | 001000 |
875 -------------------------------------------
876
877 Figure 3: Treatment Aggregate and MPLS EXP Field Usage
878
879 * Note: For Assured Elastic (and Elastic) Treatment Aggregate, the
880 usage of 010 or 011 (000 or 001) as EXP field value depends on the
881 drop probability. Packets in the LSP with EXP field of 011 (001)
882 have a higher probability of being dropped than packets with an
883 EXP field of 010 (000).
884
885
886
887
888
889
890
891
892 <span class="grey">Chan, et al. Informational [Page 16]</span></pre>
893 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-17" ></span>
894 <span class="grey"><a href="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
895
896
897 The above table indicates the recommended usage of EXP fields for
898 treatment aggregates. Because many deployments of MPLS are on a per-
899 domain basis, each domain has total control of its EXP usage and each
900 domain may use a different EXP field allocation for the domain's
901 supported treatment aggregates.
902
903 <span class="h3"><a class="selflink" id="appendix-A.1" href="#appendix-A.1">A.1</a>. Network Control Treatment Aggregate with E-LSP</span>
904
905 The usage of E-LSP for Network Control Treatment Aggregate needs to
906 adhere to the recommendations indicated in <a href="#section-4.1.1">section 4.1.1</a> of this
907 document and <a href="./rfc4594#section-3.2">section&nbsp;3.2 of RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>]. Reinforcing these
908 recommendations, there should be no drop precedence associated with
909 the MPLS PSC used for Network Control Treatment Aggregate because
910 dropping of Network Control Treatment Aggregate traffic should be
911 prevented.
912
913 <span class="h3"><a class="selflink" id="appendix-A.2" href="#appendix-A.2">A.2</a>. Real-Time Treatment Aggregate with E-LSP</span>
914
915 In addition to the recommendations provided in <a href="#section-4.1.2">section 4.1.2</a> of this
916 document and in member service classes' sections of <a href="./rfc4594">RFC 4594</a> [<a href="#ref-3" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">3</a>], we
917 want to indicate that Real-Time Treatment Aggregate traffic should
918 not be dropped, as some of the applications whose traffic is carried
919 in the Real-Time Treatment Aggregate do not react well to dropped
920 packets. As indicated in <a href="#section-4.1.2">section 4.1.2</a> of this document, admission
921 control should be performed on each service class contributing to the
922 Real-Time Treatment Aggregate to prevent packet loss due to
923 insufficient resources allocated to Real-Time Treatment Aggregate.
924 Further, admission control and policing may also be applied on the
925 sum of all traffic aggregated into this treatment aggregate.
926
927 <span class="h3"><a class="selflink" id="appendix-A.3" href="#appendix-A.3">A.3</a>. Assured Elastic Treatment Aggregate with E-LSP</span>
928
929 EXP field markings of 010 and 011 are used for the Assured Elastic
930 Treatment Aggregate. The two encodings are used to provide two
931 levels of drop precedence indications, with 010 encoded traffic
932 having a lower probability of being dropped than 011 encoded traffic.
933 This provides for the mapping of CS2, AF31, AF21, and AF11 into EXP
934 010; and AF32, AF22, AF12 and AF33, AF23, AF13 into EXP 011. If the
935 domain chooses to support only one drop precedence for this treatment
936 aggregate, we recommend the use of 010 for EXP field marking.
937
938 <span class="h3"><a class="selflink" id="appendix-A.4" href="#appendix-A.4">A.4</a>. Elastic Treatment Aggregate with E-LSP</span>
939
940 EXP field markings of 000 and 001 are used for the Elastic Treatment
941 Aggregate. The two encodings are used to provide two levels of drop
942 precedence indications, with 000 encoded traffic having a lower
943 probability of being dropped than 001 encoded traffic. This provides
944 for the mapping of Default/CS0 into 000; and CS1 into 001. Notice
945
946
947
948 <span class="grey">Chan, et al. Informational [Page 17]</span></pre>
949 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-18" ></span>
950 <span class="grey"><a href="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
951
952
953 that with this mapping, during congestion, CS1-marked traffic may be
954 starved. If the domain chooses to support only one drop precedence
955 for this treatment aggregate, we recommend the use of 000 for EXP
956 field marking.
957
958 <span class="h3"><a class="selflink" id="appendix-A.5" href="#appendix-A.5">A.5</a>. Treatment Aggregates and L-LSP</span>
959
960 Because L-LSP (Label Only Inferred PSC LSP) supports a single PSC per
961 LSP, the support of each treatment aggregate is on a per-LSP basis.
962 This document does not further specify any additional recommendation
963 (beyond what has been indicated in <a href="#section-4">section 4</a> of this document) for
964 treatment aggregate to L-LSP mapping, leaving this to each individual
965 MPLS domain administration.
966
967 Authors' Addresses
968
969 Kwok Ho Chan
970 Nortel
971 600 Technology Park Drive
972 Billerica, MA 01821
973 US
974
975 Phone: +1-978-288-8175
976 Fax: +1-978-288-8700
977 EMail: [email protected]
978
979
980 Jozef Z. Babiarz
981 Nortel
982 3500 Carling Avenue
983 Ottawa, Ont. K2H 8E9
984 Canada
985
986 Phone: +1-613-763-6098
987 Fax: +1-613-768-2231
988 EMail: [email protected]
989
990
991 Fred Baker
992 Cisco Systems
993 1121 Via Del Rey
994 Santa Barbara, CA 93117
995 US
996
997 Phone: +1-408-526-4257
998 Fax: +1-413-473-2403
999 EMail: [email protected]
1000
1001
1002
1003
1004 <span class="grey">Chan, et al. Informational [Page 18]</span></pre>
1005 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-19" ></span>
1006 <span class="grey"><a href="./rfc5127">RFC 5127</a> Aggregation of Diffserv Service Classes February 2008</span>
1007
1008
1009 Full Copyright Statement
1010
1011 Copyright (C) The IETF Trust (2008).
1012
1013 This document is subject to the rights, licenses and restrictions
1014 contained in <a href="https://www.rfc-editor.org/bcp/bcp78">BCP 78</a>, and except as set forth therein, the authors
1015 retain all their rights.
1016
1017 This document and the information contained herein are provided on an
1018 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
1019 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND
1020 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS
1021 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF
1022 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
1023 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
1024
1025 Intellectual Property
1026
1027 The IETF takes no position regarding the validity or scope of any
1028 Intellectual Property Rights or other rights that might be claimed to
1029 pertain to the implementation or use of the technology described in
1030 this document or the extent to which any license under such rights
1031 might or might not be available; nor does it represent that it has
1032 made any independent effort to identify any such rights. Information
1033 on the procedures with respect to rights in RFC documents can be
1034 found in <a href="https://www.rfc-editor.org/bcp/bcp78">BCP 78</a> and <a href="https://www.rfc-editor.org/bcp/bcp79">BCP 79</a>.
1035
1036 Copies of IPR disclosures made to the IETF Secretariat and any
1037 assurances of licenses to be made available, or the result of an
1038 attempt made to obtain a general license or permission for the use of
1039 such proprietary rights by implementers or users of this
1040 specification can be obtained from the IETF on-line IPR repository at
1041 <a href="http://www.ietf.org/ipr">http://www.ietf.org/ipr</a>.
1042
1043 The IETF invites any interested party to bring to its attention any
1044 copyrights, patents or patent applications, or other proprietary
1045 rights that may cover technology that may be required to implement
1046 this standard. Please address the information to the IETF at
1047 [email protected].
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060 Chan, et al. Informational [Page 19]
1061 </pre>