Debian Codegraph @

rfc6661.html

raw
1 <pre>Internet Engineering Task Force (IETF) A. Charny
2 Request for Comments: 6661
3 Category: Experimental F. Huang
4 ISSN: 2070-1721 Huawei Technologies
5 G. Karagiannis
6 University of Twente
7 M. Menth
8 University of Tuebingen
9 T. Taylor, Ed.
10 Huawei Technologies
11 July 2012
12
13
14 <span class="h1">Pre-Congestion Notification (PCN) Boundary-Node Behavior for the</span>
15 <span class="h1">Controlled Load (CL) Mode of Operation</span>
16
17 Abstract
18
19 Pre-Congestion Notification (PCN) is a means for protecting the
20 quality of service for inelastic traffic admitted to a Diffserv
21 domain. The overall PCN architecture is described in <a href="./rfc5559">RFC 5559</a>. This
22 memo is one of a series describing possible boundary-node behaviors
23 for a PCN-domain. The behavior described here is that for a form of
24 measurement-based load control using three PCN marking states: not-
25 marked, threshold-marked, and excess-traffic-marked. This behavior
26 is known informally as the Controlled Load (CL) PCN-boundary-node
27 behavior.
28
29 Status of This Memo
30
31 This document is not an Internet Standards Track specification; it is
32 published for examination, experimental implementation, and
33 evaluation.
34
35 This document defines an Experimental Protocol for the Internet
36 community. This document is a product of the Internet Engineering
37 Task Force (IETF). It represents the consensus of the IETF
38 community. It has received public review and has been approved for
39 publication by the Internet Engineering Steering Group (IESG). Not
40 all documents approved by the IESG are a candidate for any level of
41 Internet Standard; see <a href="./rfc5741#section-2">Section&nbsp;2 of RFC 5741</a>.
42
43 Information about the current status of this document, any errata,
44 and how to provide feedback on it may be obtained at
45 <a href="http://www.rfc-editor.org/info/rfc6661">http://www.rfc-editor.org/info/rfc6661</a>.
46
47
48
49
50
51
52 <span class="grey">Charny, et al. Experimental [Page 1]</span></pre>
53 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-2" ></span>
54 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
55
56
57 Copyright Notice
58
59 Copyright (c) 2012 IETF Trust and the persons identified as the
60 document authors. All rights reserved.
61
62 This document is subject to <a href="https://www.rfc-editor.org/bcp/bcp78">BCP 78</a> and the IETF Trust's Legal
63 Provisions Relating to IETF Documents
64 (<a href="http://trustee.ietf.org/license-info">http://trustee.ietf.org/license-info</a>) in effect on the date of
65 publication of this document. Please review these documents
66 carefully, as they describe your rights and restrictions with respect
67 to this document. Code Components extracted from this document must
68 include Simplified BSD License text as described in Section 4.e of
69 the Trust Legal Provisions and are provided without warranty as
70 described in the Simplified BSD License.
71
72
73
74
75
76
77
78
79
80
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">Charny, et al. Experimental [Page 2]</span></pre>
109 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-3" ></span>
110 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
111
112
113 Table of Contents
114
115 <a href="#section-1">1</a>. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-4">4</a>
116 <a href="#section-1.1">1.1</a>. Terminology . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-5">5</a>
117 <a href="#section-2">2</a>. [CL-Specific] Assumed Core Network Behavior for CL . . . . . . <a href="#page-8">8</a>
118 <a href="#section-3">3</a>. Node Behaviors . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-9">9</a>
119 <a href="#section-3.1">3.1</a>. Overview . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-9">9</a>
120 <a href="#section-3.2">3.2</a>. Behavior of the PCN-Egress-Node . . . . . . . . . . . . . <a href="#page-10">10</a>
121 <a href="#section-3.2.1">3.2.1</a>. Data Collection . . . . . . . . . . . . . . . . . . . <a href="#page-10">10</a>
122 <a href="#section-3.2.2">3.2.2</a>. Reporting the PCN Data . . . . . . . . . . . . . . . . <a href="#page-11">11</a>
123 <a href="#section-3.2.3">3.2.3</a>. Optional Report Suppression . . . . . . . . . . . . . <a href="#page-11">11</a>
124 <a href="#section-3.3">3.3</a>. Behavior at the Decision Point . . . . . . . . . . . . . . <a href="#page-12">12</a>
125 <a href="#section-3.3.1">3.3.1</a>. Flow Admission . . . . . . . . . . . . . . . . . . . . <a href="#page-12">12</a>
126 <a href="#section-3.3.2">3.3.2</a>. Flow Termination . . . . . . . . . . . . . . . . . . . <a href="#page-13">13</a>
127 3.3.3. Decision Point Action for Missing
128 PCN-Boundary-Node Reports . . . . . . . . . . . . . . <a href="#page-15">15</a>
129 <a href="#section-3.4">3.4</a>. Behavior of the Ingress Node . . . . . . . . . . . . . . . <a href="#page-16">16</a>
130 3.5. Summary of Timers and Associated Configurable Durations . 16
131 <a href="#section-3.5.1">3.5.1</a>. Recommended Values for the Configurable Durations . . <a href="#page-18">18</a>
132 <a href="#section-4">4</a>. Specification of Diffserv Per-Domain Behavior . . . . . . . . <a href="#page-18">18</a>
133 <a href="#section-4.1">4.1</a>. Applicability . . . . . . . . . . . . . . . . . . . . . . <a href="#page-18">18</a>
134 <a href="#section-4.2">4.2</a>. Technical Specification . . . . . . . . . . . . . . . . . <a href="#page-19">19</a>
135 <a href="#section-4.2.1">4.2.1</a>. Classification and Traffic Conditioning . . . . . . . <a href="#page-19">19</a>
136 <a href="#section-4.2.2">4.2.2</a>. PHB Configuration . . . . . . . . . . . . . . . . . . <a href="#page-19">19</a>
137 <a href="#section-4.3">4.3</a>. Attributes . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-19">19</a>
138 <a href="#section-4.4">4.4</a>. Parameters . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-19">19</a>
139 <a href="#section-4.5">4.5</a>. Assumptions . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-20">20</a>
140 <a href="#section-4.6">4.6</a>. Example Uses . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-20">20</a>
141 <a href="#section-4.7">4.7</a>. Environmental Concerns . . . . . . . . . . . . . . . . . . <a href="#page-20">20</a>
142 <a href="#section-4.8">4.8</a>. Security Considerations . . . . . . . . . . . . . . . . . <a href="#page-20">20</a>
143 <a href="#section-5">5</a>. Operational and Management Considerations . . . . . . . . . . <a href="#page-20">20</a>
144 <a href="#section-5.1">5.1</a>. Deployment of the CL Edge Behavior . . . . . . . . . . . . <a href="#page-20">20</a>
145 5.1.1. Selection of Deployment Options and Global
146 Parameters . . . . . . . . . . . . . . . . . . . . . . <a href="#page-20">20</a>
147 5.1.2. Specification of Node- and Link-Specific Parameters . 22
148 <a href="#section-5.1.3">5.1.3</a>. Installation of Parameters and Policies . . . . . . . <a href="#page-23">23</a>
149 <a href="#section-5.1.4">5.1.4</a>. Activation and Verification of All Behaviors . . . . . <a href="#page-24">24</a>
150 <a href="#section-5.2">5.2</a>. Management Considerations . . . . . . . . . . . . . . . . <a href="#page-25">25</a>
151 <a href="#section-5.2.1">5.2.1</a>. Event Logging in the PCN-Domain . . . . . . . . . . . <a href="#page-25">25</a>
152 <a href="#section-5.2.1.1">5.2.1.1</a>. Logging Loss and Restoration of Contact . . . . . <a href="#page-25">25</a>
153 <a href="#section-5.2.1.2">5.2.1.2</a>. Logging Flow Termination Events . . . . . . . . . <a href="#page-27">27</a>
154 <a href="#section-5.2.2">5.2.2</a>. Provision and Use of Counters . . . . . . . . . . . . <a href="#page-28">28</a>
155 <a href="#section-6">6</a>. Security Considerations . . . . . . . . . . . . . . . . . . . <a href="#page-29">29</a>
156 <a href="#section-7">7</a>. IANA Considerations . . . . . . . . . . . . . . . . . . . . . <a href="#page-30">30</a>
157 <a href="#section-8">8</a>. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-30">30</a>
158 <a href="#section-9">9</a>. References . . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-31">31</a>
159 <a href="#section-9.1">9.1</a>. Normative References . . . . . . . . . . . . . . . . . . . <a href="#page-31">31</a>
160 <a href="#section-9.2">9.2</a>. Informative References . . . . . . . . . . . . . . . . . . <a href="#page-32">32</a>
161
162
163
164 <span class="grey">Charny, et al. Experimental [Page 3]</span></pre>
165 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-4" ></span>
166 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
167
168
169 <span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Introduction</span>
170
171 The objective of Pre-Congestion Notification (PCN) is to protect the
172 quality of service (QoS) of inelastic flows within a Diffserv domain,
173 in a simple, scalable, and robust fashion. Two mechanisms are used:
174 admission control to decide whether to admit or block a new flow
175 request and, in abnormal circumstances, flow termination to decide
176 whether to terminate some of the existing flows. To achieve this,
177 the overall rate of PCN-traffic is metered on every link in the PCN-
178 domain, and PCN-packets are appropriately marked when certain
179 configured rates are exceeded. These configured rates are below the
180 rate of the link, thus providing notification to PCN-boundary-nodes
181 about incipient overloads before any congestion occurs (hence the
182 "pre" part of "pre-congestion notification"). The level of marking
183 allows decisions to be made about whether to admit or terminate PCN-
184 flows. For more details, see [<a href="./rfc5559" title="&quot;Pre-Congestion Notification (PCN) Architecture&quot;">RFC5559</a>].
185
186 This document describes an experimental edge-node behavior to
187 implement PCN in a network. The experiment may be run in a network
188 in which a substantial proportion of the traffic carried is in the
189 form of inelastic flows and where admission control of micro-flows is
190 applied at the edge. For the effects of PCN to be observable, the
191 committed bandwidth (i.e., level of non-best-effort traffic) on at
192 least some links of the network should be near or at link capacity.
193 The amount of effort required to prepare the network for the
194 experiment (see <a href="#section-5.1">Section 5.1</a>) may constrain the size of network to
195 which it is applied. The purposes of the experiment are:
196
197 o to validate the specification of the CL edge behavior;
198
199 o to evaluate the effectiveness of the CL edge behavior in
200 preserving quality of service for admitted flows; and
201
202 o to evaluate PCN's potential for reducing the amount of capital and
203 operational costs in comparison to alternative methods of assuring
204 quality of service.
205
206 For the first two objectives, the experiment should run long enough
207 for the network to experience sharp peaks of traffic in at least some
208 directions. It would also be desirable to observe PCN performance in
209 the face of failures in the network. A period on the order of a
210 month or two in busy season may be enough. The third objective is
211 more difficult and could require observation over a period long
212 enough for traffic demand to grow to the point where additional
213 capacity must be provisioned at some points in the network.
214
215
216
217
218
219
220 <span class="grey">Charny, et al. Experimental [Page 4]</span></pre>
221 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-5" ></span>
222 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
223
224
225 <a href="#section-3">Section 3</a> of this document specifies a detailed set of algorithms and
226 procedures used to implement the PCN mechanisms for the CL mode of
227 operation. Since the algorithms depend on specific metering and
228 marking behavior at the interior nodes, it is also necessary to
229 specify the assumptions made about PCN-interior-node behavior
230 (<a href="#section-2">Section 2</a>). Finally, because PCN uses Diffserv codepoint (DSCP)
231 values to carry its markings, a specification of PCN-boundary-node
232 behavior must include the per-domain behavior (PDB) template
233 specified in [<a href="./rfc3086" title="&quot;Definition of Differentiated Services Per Domain Behaviors and Rules for their Specification&quot;">RFC3086</a>], filled out with the appropriate content
234 (<a href="#section-4">Section 4</a>).
235
236 Note that the terms "block" or "terminate" actually translate to one
237 or more of several possible courses of action, as discussed in
238 <a href="./rfc5559#section-3.6">Section&nbsp;3.6 of [RFC5559]</a>. The choice of which action to take for
239 blocked or terminated flows is a matter of local policy.
240
241 A companion document [<a href="./rfc6662" title="&quot;Pre-Congestion Notification (PCN) Boundary Node Behavior for the Single Marking (SM) Mode of Operation&quot;">RFC6662</a>] specifies the Single Marking (SM) PCN-
242 boundary-node behavior. This document and [<a href="./rfc6662" title="&quot;Pre-Congestion Notification (PCN) Boundary Node Behavior for the Single Marking (SM) Mode of Operation&quot;">RFC6662</a>] have a great
243 deal of text in common. To simplify the task of the reader, the text
244 in the present document that is specific to the CL PCN-boundary-node
245 behavior is preceded by the phrase "[<a href="#ref-CL-specific">CL-specific</a>]". A similar
246 distinction for SM-specific text is made in [<a href="./rfc6662" title="&quot;Pre-Congestion Notification (PCN) Boundary Node Behavior for the Single Marking (SM) Mode of Operation&quot;">RFC6662</a>].
247
248 <span class="h3"><a class="selflink" id="section-1.1" href="#section-1.1">1.1</a>. Terminology</span>
249
250 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
251 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
252 document are to be interpreted as described in [<a href="./rfc2119" title="&quot;Key words for use in RFCs to Indicate Requirement Levels&quot;">RFC2119</a>].
253
254 This document uses the following terms defined in <a href="./rfc5559#section-2">Section&nbsp;2 of
255 [RFC5559]</a>:
256
257 o PCN-domain
258
259 o PCN-ingress-node
260
261 o PCN-egress-node
262
263 o PCN-interior-node
264
265 o PCN-boundary-node
266
267 o PCN-flow
268
269 o ingress-egress-aggregate
270
271 o [<a href="#ref-CL-specific">CL-specific</a>] PCN-threshold-rate
272
273
274
275
276 <span class="grey">Charny, et al. Experimental [Page 5]</span></pre>
277 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-6" ></span>
278 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
279
280
281 o PCN-excess-rate
282
283 o PCN-admissible-rate
284
285 o PCN-supportable-rate
286
287 o PCN-marked
288
289 o [<a href="#ref-CL-specific">CL-specific</a>] threshold-marked
290
291 o excess-traffic-marked
292
293 It also uses the terms PCN-traffic and PCN-packet, for which the
294 definition is repeated from [<a href="./rfc5559" title="&quot;Pre-Congestion Notification (PCN) Architecture&quot;">RFC5559</a>] because of their importance to
295 the understanding of the text that follows:
296
297 PCN-traffic, PCN-packets, PCN-BA
298 A PCN-domain carries traffic of different Diffserv behavior
299 aggregates (BAs) [<a href="./rfc2474" title="&quot;Definition of the Differentiated Services Field (DS Field) in the IPv4 and IPv6 Headers&quot;">RFC2474</a>]. The PCN-BA uses the PCN mechanisms to
300 carry PCN-traffic, and the corresponding packets are PCN-packets.
301 The same network will carry traffic of other Diffserv BAs. The
302 PCN-BA is distinguished by a combination of the Diffserv codepoint
303 and the ECN field.
304
305 This document uses the following terms from [<a href="./rfc5670" title="&quot;Metering and Marking Behaviour of PCN- Nodes&quot;">RFC5670</a>]:
306
307 o [<a href="#ref-CL-specific">CL-specific</a>] threshold-meter;
308
309 o excess-traffic-meter.
310
311 To complete the list of borrowed terms, this document reuses the
312 following terms and abbreviations defined in <a href="./rfc6660#section-2">Section&nbsp;2 of [RFC6660]</a>:
313
314 o not-PCN codepoint;
315
316 o not-marked (NM) codepoint;
317
318 o [<a href="#ref-CL-specific">CL-specific</a>] threshold-marked (ThM) codepoint;
319
320 o excess-traffic-marked (ETM) codepoint.
321
322 This document defines the following additional terms:
323
324 Decision Point
325 The node that makes the decision about which flows to admit and to
326 terminate. In a given network deployment, this can be the PCN-
327 ingress-node or a centralized control node. In either case, the
328 PCN-ingress-node is the point where the decisions are enforced.
329
330
331
332 <span class="grey">Charny, et al. Experimental [Page 6]</span></pre>
333 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-7" ></span>
334 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
335
336
337 NM-rate
338 The rate of not-marked PCN-traffic received at a PCN-egress-node
339 for a given ingress-egress-aggregate in octets per second. For
340 further details, see <a href="#section-3.2.1">Section 3.2.1</a>.
341
342 [<a id="ref-CL-specific">CL-specific</a>] ThM-rate
343 The rate of threshold-marked PCN-traffic received at a PCN-egress-
344 node for a given ingress-egress-aggregate in octets per second.
345 For further details, see <a href="#section-3.2.1">Section 3.2.1</a>.
346
347 ETM-rate
348 The rate of excess-traffic-marked PCN-traffic received at a PCN-
349 egress-node for a given ingress-egress-aggregate in octets per
350 second. For further details, see <a href="#section-3.2.1">Section 3.2.1</a>.
351
352 PCN-sent-rate
353 The rate of PCN-traffic received at a PCN-ingress-node and
354 destined for a given ingress-egress-aggregate in octets per
355 second. For further details, see <a href="#section-3.4">Section 3.4</a>.
356
357 Congestion level estimate (CLE)
358 The ratio of PCN-marked to total PCN-traffic (measured in octets)
359 received for a given ingress-egress-aggregate during a given
360 measurement period. The CLE is used to derive the PCN-admission-
361 state (<a href="#section-3.3.1">Section 3.3.1</a>) and is also used by the report suppression
362 procedure (<a href="#section-3.2.3">Section 3.2.3</a>) if report suppression is activated.
363
364 PCN-admission-state
365 The state ("admit" or "block") derived by the Decision Point for a
366 given ingress-egress-aggregate based on statistics about PCN-
367 packet marking. The Decision Point decides to admit or block new
368 flows offered to the aggregate based on the current value of the
369 PCN-admission-state. For further details, see <a href="#section-3.3.1">Section 3.3.1</a>.
370
371 Sustainable aggregate rate (SAR)
372 The estimated maximum rate of PCN-traffic that can be carried in a
373 given ingress-egress-aggregate at a given moment without risking
374 degradation of quality of service for the admitted flows. The
375 intention is that if the PCN-sent-rate of every ingress-egress-
376 aggregate passing through a given link is limited to its
377 sustainable aggregate rate, the total rate of PCN-traffic flowing
378 through the link will be limited to the PCN-supportable-rate for
379 that link. An estimate of the sustainable aggregate rate for a
380 given ingress-egress-aggregate is derived as part of the flow
381 termination procedure and is used to determine how much PCN-
382 traffic needs to be terminated. For further details, see
383 <a href="#section-3.3.2">Section 3.3.2</a>.
384
385
386
387
388 <span class="grey">Charny, et al. Experimental [Page 7]</span></pre>
389 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-8" ></span>
390 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
391
392
393 CLE-reporting-threshold
394 A configurable value against which the CLE is compared as part of
395 the report suppression procedure. For further details, see
396 <a href="#section-3.2.3">Section 3.2.3</a>.
397
398 CLE-limit
399 A configurable value against which the CLE is compared to
400 determine the PCN-admission-state for a given ingress-egress-
401 aggregate. For further details, see <a href="#section-3.3.1">Section 3.3.1</a>.
402
403 T_meas
404 A configurable time interval that defines the measurement period
405 over which the PCN-egress-node collects statistics relating to
406 PCN-traffic marking. At the end of the interval, the PCN-egress-
407 node calculates the values NM-rate, [<a href="#ref-CL-specific">CL-specific</a>] ThM-rate, and
408 ETM-rate as defined above and sends a report to the Decision
409 Point, subject to the operation of the report suppression feature.
410 For further details, see <a href="#section-3.2">Section 3.2</a>.
411
412 T_maxsuppress
413 A configurable time interval after which the PCN-egress-node MUST
414 send a report to the Decision Point for a given ingress-egress-
415 aggregate regardless of the most recent values of the CLE. This
416 mechanism provides the Decision Point with a periodic confirmation
417 of liveness when report suppression is activated. For further
418 details, see <a href="#section-3.2.3">Section 3.2.3</a>.
419
420 T_fail
421 An interval after which the Decision Point concludes that
422 communication from a given PCN-egress-node has failed if it has
423 received no reports from the PCN-egress-node during that interval.
424 For further details, see <a href="#section-3.3.3">Section 3.3.3</a>.
425
426 T_crit
427 A configurable interval used in the calculation of T_fail. For
428 further details, see <a href="#section-3.3.3">Section 3.3.3</a>.
429
430 <span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. [CL-Specific] Assumed Core Network Behavior for CL</span>
431
432 This section describes the assumed behavior for PCN-interior-nodes in
433 the PCN-domain. The CL mode of operation assumes that:
434
435 o PCN-interior-nodes perform both threshold-marking and excess-
436 traffic-marking of PCN-packets, according to the rules specified
437 in [<a href="./rfc5670" title="&quot;Metering and Marking Behaviour of PCN- Nodes&quot;">RFC5670</a>];
438
439
440
441
442
443
444 <span class="grey">Charny, et al. Experimental [Page 8]</span></pre>
445 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-9" ></span>
446 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
447
448
449 o For IP transport, threshold-marking of PCN-packets uses the ThM
450 codepoint defined in [<a href="./rfc6660" title="&quot;Encoding Three Pre-Congestion Notification (PCN) States in the IP Header Using a Single Diffserv Codepoint (DSCP)&quot;">RFC6660</a>]; for MPLS transport, an equivalent
451 marking is used as discussed in <a href="./rfc6660#appendix-C">Appendix&nbsp;C of [RFC6660]</a>;
452
453 o For IP transport, excess-traffic-marking of PCN-packets uses the
454 ETM codepoint defined in [<a href="./rfc6660" title="&quot;Encoding Three Pre-Congestion Notification (PCN) States in the IP Header Using a Single Diffserv Codepoint (DSCP)&quot;">RFC6660</a>]; for MPLS transport, an
455 equivalent marking is used as discussed in <a href="./rfc6660#appendix-C">Appendix&nbsp;C of
456 [RFC6660]</a>;
457
458 o On each link, the reference rate for the threshold-meter is
459 configured to be equal to the PCN-admissible-rate for the link;
460
461 o On each link, the reference rate for the excess-traffic-meter is
462 configured to be equal to the PCN-supportable-rate for the link;
463
464 o The set of valid codepoint transitions is as shown in Sections
465 5.2.1 and 5.2.2 of [<a href="./rfc6660" title="&quot;Encoding Three Pre-Congestion Notification (PCN) States in the IP Header Using a Single Diffserv Codepoint (DSCP)&quot;">RFC6660</a>].
466
467 <span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. Node Behaviors</span>
468
469 <span class="h3"><a class="selflink" id="section-3.1" href="#section-3.1">3.1</a>. Overview</span>
470
471 This section describes the behavior of the PCN-ingress-node, PCN-
472 egress-node, and the Decision Point (which MAY be collocated with the
473 PCN-ingress-node).
474
475 The PCN-egress-node collects the rates of not-marked, [<a href="#ref-CL-specific">CL-specific</a>]
476 threshold-marked, and excess-traffic-marked PCN-traffic for each
477 ingress-egress-aggregate and reports them to the Decision Point.
478 [<a href="#ref-CL-specific">CL-specific</a>] It MAY also identify and report PCN-flows that have
479 experienced excess-traffic-marking. For a detailed description, see
480 <a href="#section-3.2">Section 3.2</a>.
481
482 The PCN-ingress-node enforces flow admission and termination
483 decisions. It also reports the rate of PCN-traffic sent to a given
484 ingress-egress-aggregate when requested by the Decision Point. For
485 details, see <a href="#section-3.4">Section 3.4</a>.
486
487 Finally, the Decision Point makes flow admission decisions and
488 selects flows to terminate based on the information provided by the
489 PCN-ingress-node and PCN-egress-node for a given ingress-egress-
490 aggregate. For details, see <a href="#section-3.3">Section 3.3</a>.
491
492 Specification of a signaling protocol to report rates to the Decision
493 Point is out of scope of this document. If the PCN-ingress-node is
494 chosen as the Decision Point, [<a href="#ref-RSVP-PCN" title="&quot;Generic Aggregation of Resource ReSerVation Protocol (RSVP) for IPv4 And IPv6 Reservations over PCN domains&quot;">RSVP-PCN</a>] specifies an appropriate
495 signaling protocol.
496
497
498
499
500 <span class="grey">Charny, et al. Experimental [Page 9]</span></pre>
501 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-10" ></span>
502 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
503
504
505 <a href="#section-5.1.2">Section 5.1.2</a> describes how to derive the filters by means of which
506 PCN-ingress-nodes and PCN-egress-nodes are able to classify incoming
507 packets into ingress-egress-aggregates.
508
509 <span class="h3"><a class="selflink" id="section-3.2" href="#section-3.2">3.2</a>. Behavior of the PCN-Egress-Node</span>
510
511 <span class="h4"><a class="selflink" id="section-3.2.1" href="#section-3.2.1">3.2.1</a>. Data Collection</span>
512
513 The PCN-egress-node needs to meter the PCN-traffic it receives in
514 order to calculate the following rates for each ingress-egress-
515 aggregate passing through it. These rates SHOULD be calculated at
516 the end of each measurement period based on the PCN-traffic observed
517 during that measurement period. The duration of a measurement period
518 is equal to the configurable value T_meas. For further information,
519 see <a href="#section-3.5">Section 3.5</a>.
520
521 o NM-rate: octets per second of PCN-traffic in PCN-packets that are
522 not-marked (i.e., marked with the NM codepoint);
523
524 o [<a href="#ref-CL-specific">CL-specific</a>] ThM-rate: octets per second of PCN-traffic in PCN-
525 packets that are threshold-marked (i.e., marked with the ThM
526 codepoint);
527
528 o ETM-rate: octets per second of PCN-traffic in PCN-packets that are
529 excess-traffic-marked (i.e., marked with the ETM codepoint).
530
531 Note: metering the PCN-traffic continuously and using equal-length
532 measurement intervals minimizes the statistical variance introduced
533 by the measurement process itself. On the other hand, the operation
534 of PCN is not affected if the starting and ending times of the
535 measurement intervals for different ingress-egress-aggregates are
536 different.
537
538 [<a id="ref-CL-specific">CL-specific</a>] As a configurable option, the PCN-egress-node MAY
539 record flow identifiers of the PCN-flows for which excess-traffic-
540 marked packets have been observed during this measurement interval.
541 If this set is large (e.g., more than 20 flows), the PCN-egress-node
542 MAY record only the most recently excess-traffic-marked PCN-flow
543 identifiers rather than the complete set.
544
545 These can be used by the Decision Point when it selects flows for
546 termination. In networks using multipath routing, it is possible
547 that congestion is not occurring on all paths carrying a given
548 ingress-egress-aggregate. Assuming that specific PCN-flows are
549 routed via specific paths, identifying the PCN-flows that are
550 experiencing excess-traffic-marking helps to avoid termination of
551 PCN-flows not contributing to congestion.
552
553
554
555
556 <span class="grey">Charny, et al. Experimental [Page 10]</span></pre>
557 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-11" ></span>
558 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
559
560
561 <span class="h4"><a class="selflink" id="section-3.2.2" href="#section-3.2.2">3.2.2</a>. Reporting the PCN Data</span>
562
563 Unless the report suppression option described in <a href="#section-3.2.3">Section 3.2.3</a> is
564 activated, the PCN-egress-node MUST report the latest values of NM-
565 rate, [<a href="#ref-CL-specific">CL-specific</a>] ThM-rate, and ETM-rate to the Decision Point each
566 time that it calculates them.
567
568 [<a id="ref-CL-specific">CL-specific</a>] If the PCN-egress-node recorded a set of flow
569 identifiers of PCN-flows for which excess-traffic-marking was
570 observed in the most recent measurement interval, then it MUST also
571 include these identifiers in the report.
572
573 <span class="h4"><a class="selflink" id="section-3.2.3" href="#section-3.2.3">3.2.3</a>. Optional Report Suppression</span>
574
575 Report suppression MUST be provided as a configurable option, along
576 with two configurable parameters, the CLE-reporting-threshold and the
577 maximum report suppression interval T_maxsuppress. The default value
578 of the CLE-reporting-threshold is zero. The CLE-reporting-threshold
579 MUST NOT exceed the CLE-limit configured at the Decision Point. For
580 further information on T_maxsuppress, see <a href="#section-3.5">Section 3.5</a>.
581
582 If the report suppression option is enabled, the PCN-egress-node MUST
583 apply the following procedure to decide whether to send a report to
584 the Decision Point, rather than sending a report automatically at the
585 end of each measurement interval.
586
587 1. As well as the quantities NM-rate, [<a href="#ref-CL-specific">CL-specific</a>] ThM-rate, and
588 ETM-rate, the PCN-egress-node MUST calculate the congestion level
589 estimate (CLE) for each measurement interval. The CLE is
590 computed as:
591
592 [<a id="ref-CL-specific">CL-specific</a>]
593 CLE = (ThM-rate + ETM-rate) / (NM-rate + ThM-rate + ETM-rate)
594
595 if any PCN-traffic was observed, or CLE = 0 if all the rates are
596 zero.
597
598 2. If the CLE calculated for the latest measurement interval is
599 greater than the CLE-reporting-threshold and/or the CLE
600 calculated for the immediately previous interval was greater than
601 the CLE-reporting-threshold, then the PCN-egress-node MUST send a
602 report to the Decision Point. The contents of the report are
603 described below.
604
605
606
607
608
609
610
611
612 <span class="grey">Charny, et al. Experimental [Page 11]</span></pre>
613 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-12" ></span>
614 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
615
616
617 The reason for taking into account the CLE of the previous
618 interval is to ensure that the Decision Point gets immediate
619 feedback if the CLE has dropped below CLE-reporting-threshold.
620 This is essential if the Decision Point is running the flow
621 termination procedure and observing whether (further) flow
622 termination is needed. See <a href="#section-3.3.2">Section 3.3.2</a>.
623
624 3. If an interval T_maxsuppress has elapsed since the last report
625 was sent to the Decision Point, then the PCN-egress-node MUST
626 send a report to the Decision Point regardless of the CLE value.
627
628 4. If neither of the preceding conditions holds, the PCN-egress-node
629 MUST NOT send a report for the latest measurement interval.
630
631 Each report sent to the Decision Point when report suppression has
632 been activated MUST contain the values of NM-rate, [<a href="#ref-CL-specific">CL-specific</a>] ThM-
633 rate, ETM-rate, and CLE that were calculated for the most recent
634 measurement interval. [<a href="#ref-CL-specific">CL-specific</a>] If the PCN-egress-node recorded
635 a set of flow identifiers of PCN-flows for which excess-traffic-
636 marking was observed in the most recent measurement interval, then it
637 MUST also include these identifiers in the report.
638
639 The above procedure ensures that at least one report is sent per
640 interval (T_maxsuppress + T_meas). This demonstrates to the Decision
641 Point that both the PCN-egress-node and the communication path
642 between that node and the Decision Point are in operation.
643
644 <span class="h3"><a class="selflink" id="section-3.3" href="#section-3.3">3.3</a>. Behavior at the Decision Point</span>
645
646 Operators can choose to use PCN procedures just for flow admission,
647 or just for flow termination, or for both. Decision Points MUST
648 implement both mechanisms, but configurable options MUST be provided
649 to activate or deactivate PCN-based flow admission and flow
650 termination independently of each other at a given Decision Point.
651
652 If PCN-based flow termination is enabled but PCN-based flow admission
653 is not, flow termination operates as specified in this document.
654
655 Logically, some other system of flow admission control is in
656 operation, but the description of such a system is out of scope of
657 this document and depends on local arrangements.
658
659 <span class="h4"><a class="selflink" id="section-3.3.1" href="#section-3.3.1">3.3.1</a>. Flow Admission</span>
660
661 The Decision Point determines the PCN-admission-state for a given
662 ingress-egress-aggregate each time it receives a report from the
663 egress node. It makes this determination on the basis of the
664 congestion level estimate (CLE). If the CLE is provided in the
665
666
667
668 <span class="grey">Charny, et al. Experimental [Page 12]</span></pre>
669 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-13" ></span>
670 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
671
672
673 egress-node report, the Decision Point SHOULD use the reported value.
674 If the CLE was not provided in the report, the Decision Point MUST
675 calculate it based on the other values provided in the report, using
676 the formula:
677
678 [<a id="ref-CL-specific">CL-specific</a>]
679 CLE = (ThM-rate + ETM-rate) / (NM-rate + ThM-rate + ETM-rate)
680
681 if any PCN-traffic was observed, or CLE = 0 if all the rates are
682 zero.
683
684 The Decision Point MUST compare the reported or calculated CLE to a
685 configurable value, the CLE-limit. If the CLE is less than the CLE-
686 limit, the PCN-admission-state for that aggregate MUST be set to
687 "admit"; otherwise, it MUST be set to "block".
688
689 If the PCN-admission-state for a given ingress-egress-aggregate is
690 "admit", the Decision Point SHOULD allow new flows to be admitted to
691 that aggregate. If the PCN-admission-state for a given ingress-
692 egress-aggregate is "block", the Decision Point SHOULD NOT allow new
693 flows to be admitted to that aggregate. These actions MAY be
694 modified by policy in specific cases, but such policy intervention
695 risks defeating the purpose of using PCN.
696
697 A performance study of this admission control method is presented in
698 [<a href="#ref-MeLe12" title="&quot;Performance of PCN-Based Admission Control under Challenging Conditions&quot;">MeLe12</a>].
699
700 <span class="h4"><a class="selflink" id="section-3.3.2" href="#section-3.3.2">3.3.2</a>. Flow Termination</span>
701
702 [<a id="ref-CL-specific">CL-specific</a>] When the report from the PCN-egress-node includes a
703 non-zero value of the ETM-rate for some ingress-egress-aggregate, the
704 Decision Point MUST request the PCN-ingress-node to provide an
705 estimate of the rate (PCN-sent-rate) at which the PCN-ingress-node is
706 receiving PCN-traffic that is destined for the given ingress-egress-
707 aggregate.
708
709 If the Decision Point is collocated with the PCN-ingress-node, the
710 request and response are internal operations.
711
712 The Decision Point MUST then wait, for both the requested rate from
713 the PCN-ingress-node and the next report from the PCN-egress-node for
714 the ingress-egress-aggregate concerned. If this next egress-node
715 report also includes a non-zero value for the ETM-rate, the Decision
716 Point MUST determine the amount of PCN-traffic to terminate using the
717 following steps:
718
719
720
721
722
723
724 <span class="grey">Charny, et al. Experimental [Page 13]</span></pre>
725 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-14" ></span>
726 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
727
728
729 1. [<a href="#ref-CL-specific">CL-specific</a>] The sustainable aggregate rate (SAR) for the given
730 ingress-egress-aggregate is estimated by the sum:
731
732 SAR = NM-rate + ThM-rate
733
734 for the latest reported interval.
735
736 2. The amount of traffic to be terminated is the difference:
737
738 PCN-sent-rate - SAR,
739
740 where PCN-sent-rate is the value provided by the PCN-ingress-
741 node.
742
743 See <a href="#section-3.3.3">Section 3.3.3</a> for a discussion of appropriate actions if the
744 Decision Point fails to receive a timely response to its request for
745 the PCN-sent-rate.
746
747 If the difference calculated in the second step is positive (traffic
748 rate to be terminated), the Decision Point SHOULD select PCN-flows
749 for termination. To that end, the Decision Point MAY use upper rate
750 limits for individual PCN-flows (known, e.g., from resource signaling
751 used to establish the PCN-flows) and select a set of PCN-flows whose
752 sum of upper rate limits is up to the traffic rate to be terminated.
753 Then, these PCN-flows are terminated. The use of upper limits on
754 PCN-flow rates avoids over-termination.
755
756 Termination may be continuously needed after consecutive measurement
757 intervals for various reasons, e.g., if the used upper rate limits
758 overestimate the actual flow rates. For such cases it is RECOMMENDED
759 that enough time elapses between successive termination events to
760 allow the effects of previous termination events to be reflected in
761 the measurements upon which the termination decisions are based;
762 otherwise, over-termination may occur. See [<a href="#ref-Satoh10" title="&quot;Cause and Countermeasure of Overtermination for PCN-Based Flow Termination&quot;">Satoh10</a>] and Sections
763 4.2 and 4.3 of [<a href="#ref-MeLe10" title="&quot;PCN-Based Measured Rate Termination&quot;">MeLe10</a>].
764
765 In general, the selection of flows for termination MAY be guided by
766 policy. [<a href="#ref-CL-specific">CL-specific</a>] If the egress node has supplied a list of
767 identifiers of PCN-flows that experienced excess-traffic-marking
768 (<a href="#section-3.2">Section 3.2</a>), the Decision Point SHOULD first consider terminating
769 PCN-flows in that list.
770
771 The Decision Point SHOULD log each round of termination as described
772 in <a href="#section-5.2.1.2">Section 5.2.1.2</a>.
773
774
775
776
777
778
779
780 <span class="grey">Charny, et al. Experimental [Page 14]</span></pre>
781 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-15" ></span>
782 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
783
784
785 <span class="h4"><a class="selflink" id="section-3.3.3" href="#section-3.3.3">3.3.3</a>. Decision Point Action for Missing PCN-Boundary-Node Reports</span>
786
787 The Decision Point SHOULD start a timer t_recvFail when it receives a
788 report from the PCN-egress-node. t_recvFail is reset each time a new
789 report is received from the PCN-egress-node. t_recvFail expires if it
790 reaches the value T_fail. T_fail is calculated according to the
791 following logic:
792
793 a. T_fail = the configurable duration T_crit, if report suppression
794 is not deployed;
795
796 b. T_fail = T_crit also if report suppression is deployed and the
797 last report received from the PCN-egress-node contained a CLE
798 value greater than CLE-reporting-threshold (<a href="#section-3.2.3">Section 3.2.3</a>);
799
800 c. T_fail = 3 * T_maxsuppress (<a href="#section-3.2.3">Section 3.2.3</a>) if report suppression
801 is deployed and the last report received from the PCN-egress-node
802 contained a CLE value less than or equal to CLE-reporting-
803 threshold.
804
805 If timer t_recvFail expires for a given PCN-egress-node, the Decision
806 Point SHOULD notify management. A log format is defined for that
807 purpose in <a href="#section-5.2.1.1">Section 5.2.1.1</a>. Other actions depend on local policy,
808 but MAY include blocking of new flows destined for the PCN-egress-
809 node concerned until another report is received from it. Termination
810 of already admitted flows is also possible, but could be triggered by
811 "Destination unreachable" messages received at the PCN-ingress-node.
812
813 If a centralized Decision Point sends a request for the estimated
814 value of PCN-sent-rate to a given PCN-ingress-node and fails to
815 receive a response in a reasonable amount of time, the Decision Point
816 SHOULD repeat the request once. [<a href="#ref-CL-specific">CL-specific</a>] While waiting after
817 sending this second request, the Decision Point MAY begin selecting
818 flows to terminate, using ETM-rate as an estimate of the amount of
819 traffic to be terminated in place of the quantity specified in
820 <a href="#section-3.3.2">Section 3.3.2</a>:
821
822 PCN-sent-rate - SAR
823
824 Because ETM-rate will over-estimate the amount of traffic to be
825 terminated due to dropping of PCN-packets by interior nodes, the
826 Decision Point SHOULD terminate less than the full amount ETM-rate in
827 the first pass and recalculate the additional amount to terminate in
828 additional passes based on subsequent reports from the PCN-egress-
829 node. If the second request to the PCN-ingress-node also fails, the
830 Decision Point MUST select flows to terminate based on the ETM-rate
831
832
833
834
835
836 <span class="grey">Charny, et al. Experimental [Page 15]</span></pre>
837 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-16" ></span>
838 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
839
840
841 approximation as just described and SHOULD notify management. The
842 log format described in <a href="#section-5.2.1.1">Section 5.2.1.1</a> is also suitable for this
843 purpose.
844
845 The response timer t_sndFail with upper bound T_crit is specified
846 in <a href="#section-3.5">Section 3.5</a>. The use of T_crit is an approximation. A more
847 precise limit would be on the order of two round-trip times, plus
848 an allowance for processing at each end, plus an allowance for
849 variance in these values.
850
851 See <a href="#section-3.5">Section 3.5</a> for suggested values of the configurable durations
852 T_crit and T_maxsuppress.
853
854 <span class="h3"><a class="selflink" id="section-3.4" href="#section-3.4">3.4</a>. Behavior of the Ingress Node</span>
855
856 The PCN-ingress-node MUST provide the estimated current rate of PCN-
857 traffic received at that node and destined for a given ingress-
858 egress-aggregate in octets per second (the PCN-sent-rate) when the
859 Decision Point requests it. The way this rate estimate is derived is
860 a matter of implementation.
861
862 For example, the rate that the PCN-ingress-node supplies can be
863 based on a quick sample taken at the time the information is
864 required.
865
866 <span class="h3"><a class="selflink" id="section-3.5" href="#section-3.5">3.5</a>. Summary of Timers and Associated Configurable Durations</span>
867
868 Here is a summary of the timers used in the procedures just
869 described:
870
871 t_meas
872
873 Where used: PCN-egress-node.
874
875 Used in procedure: data collection (<a href="#section-3.2.1">Section 3.2.1</a>).
876
877 Incidence: one per ingress-egress-aggregate.
878
879 Reset: immediately on expiry.
880
881 Expiry: when it reaches the configurable duration T_meas.
882
883 Action on expiry: calculate NM-rate, [<a href="#ref-CL-specific">CL-specific</a>] ThM-rate,
884 and ETM-rate and proceed to the applicable reporting procedure
885 (<a href="#section-3.2.2">Section 3.2.2</a> or <a href="#section-3.2.3">Section 3.2.3</a>).
886
887
888
889
890
891
892 <span class="grey">Charny, et al. Experimental [Page 16]</span></pre>
893 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-17" ></span>
894 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
895
896
897 t_maxsuppress
898
899 Where used: PCN-egress-node.
900
901 Used in procedure: report suppression (<a href="#section-3.2.3">Section 3.2.3</a>).
902
903 Incidence: one per ingress-egress-aggregate.
904
905 Reset: when the next report is sent, either after expiry or
906 because the CLE has exceeded the reporting threshold.
907
908 Expiry: when it reaches the configurable duration
909 T_maxsuppress.
910
911 Action on expiry: send a report to the Decision Point the next
912 time the reporting procedure (<a href="#section-3.2.3">Section 3.2.3</a>) is invoked,
913 regardless of the value of CLE.
914
915 t_recvFail
916
917 Where used: Decision Point.
918
919 Used in procedure: failure detection (<a href="#section-3.3.3">Section 3.3.3</a>).
920
921 Incidence: one per ingress-egress-aggregate.
922
923 Reset: when a report is received for the ingress-egress-
924 aggregate.
925
926 Expiry: when it reaches the calculated duration T_fail. As
927 described in <a href="#section-3.3.3">Section 3.3.3</a>, T_fail is equal either to the
928 configured duration T_crit or to the calculated value 3 *
929 T_maxsuppress, where T_maxsuppress is a configured duration.
930
931 Action on expiry: notify management, and possibly other
932 actions.
933
934 t_sndFail
935
936 Where used: centralized Decision Point.
937
938 Used in procedure: failure detection (<a href="#section-3.3.3">Section 3.3.3</a>).
939
940 Incidence: only as required, one per outstanding request to a
941 PCN-ingress-node.
942
943 Started: when a request for the value of PCN-sent-traffic for a
944 given ingress-egress-aggregate is sent to the PCN-ingress-node.
945
946
947
948 <span class="grey">Charny, et al. Experimental [Page 17]</span></pre>
949 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-18" ></span>
950 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
951
952
953 Terminated without action: when a response is received before
954 expiry.
955
956 Expiry: when it reaches the configured duration T_crit.
957
958 Action on expiry: as described in <a href="#section-3.3.3">Section 3.3.3</a>.
959
960 <span class="h4"><a class="selflink" id="section-3.5.1" href="#section-3.5.1">3.5.1</a>. Recommended Values for the Configurable Durations</span>
961
962 The timers just described depend on three configurable durations,
963 T_meas, T_maxsuppress, and T_crit. The recommendations given below
964 for the values of these durations are all related to the intended PCN
965 reaction time of 1 to 3 seconds. However, they are based on
966 judgement rather than operational experience or mathematical
967 derivation.
968
969 The value of T_meas is RECOMMENDED to be on the order of 100 to 500
970 ms to provide a reasonable trade-off between demands on network
971 resources (PCN-egress-node and Decision Point processing, network
972 bandwidth) and the time taken to react to impending congestion.
973
974 The value of T_maxsuppress is RECOMMENDED to be on the order of 3 to
975 6 seconds, for similar reasons to those for the choice of T_meas.
976
977 The value of T_crit SHOULD NOT be less than 3 * T_meas. Otherwise,
978 it could cause too many management notifications due to transient
979 conditions in the PCN-egress-node or along the signaling path. A
980 reasonable upper bound on T_crit is on the order of 3 seconds.
981
982 <span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a>. Specification of Diffserv Per-Domain Behavior</span>
983
984 This section provides the specification required by [<a href="./rfc3086" title="&quot;Definition of Differentiated Services Per Domain Behaviors and Rules for their Specification&quot;">RFC3086</a>] for a
985 per-domain behavior.
986
987 <span class="h3"><a class="selflink" id="section-4.1" href="#section-4.1">4.1</a>. Applicability</span>
988
989 This section quotes [<a href="./rfc5559" title="&quot;Pre-Congestion Notification (PCN) Architecture&quot;">RFC5559</a>].
990
991 The PCN CL boundary node behavior specified in this document is
992 applicable to inelastic traffic (particularly video and voice) where
993 quality of service for admitted flows is protected primarily by
994 admission control at the ingress to the domain.
995
996 In exceptional circumstances (e.g., due to rerouting as a result of
997 network failures) already admitted flows may be terminated to protect
998 the quality of service of the remaining flows. [<a href="#ref-CL-specific">CL-specific</a>] The
999 performance results in, e.g., [<a href="#ref-MeLe10" title="&quot;PCN-Based Measured Rate Termination&quot;">MeLe10</a>], indicate that the CL boundary
1000 node behavior provides better service outcomes under such
1001
1002
1003
1004 <span class="grey">Charny, et al. Experimental [Page 18]</span></pre>
1005 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-19" ></span>
1006 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1007
1008
1009 circumstances than the SM boundary node behavior described in
1010 [<a href="./rfc6662" title="&quot;Pre-Congestion Notification (PCN) Boundary Node Behavior for the Single Marking (SM) Mode of Operation&quot;">RFC6662</a>], because CL is less likely to terminate PCN-flows
1011 unnecessarily.
1012
1013 <span class="h3"><a class="selflink" id="section-4.2" href="#section-4.2">4.2</a>. Technical Specification</span>
1014
1015 <span class="h4"><a class="selflink" id="section-4.2.1" href="#section-4.2.1">4.2.1</a>. Classification and Traffic Conditioning</span>
1016
1017 Packet classification and treatment at the PCN-ingress-node is
1018 described in <a href="./rfc6660#section-5.1">Section&nbsp;5.1 of [RFC6660]</a>.
1019
1020 PCN packets are further classified as belonging or not belonging to
1021 an admitted flow. PCN packets not belonging to an admitted flow are
1022 "blocked". (See <a href="#section-1">Section 1</a> for an understanding of how this term is
1023 interpreted.) Packets belonging to an admitted flow are policed to
1024 ensure that they adhere to the rate or flowspec that was negotiated
1025 during flow admission.
1026
1027 <span class="h4"><a class="selflink" id="section-4.2.2" href="#section-4.2.2">4.2.2</a>. PHB Configuration</span>
1028
1029 The PCN CL boundary node behavior is a metering and marking behavior
1030 rather than a scheduling behavior. As a result, while the encoding
1031 uses a single DSCP value, that value can vary from one deployment to
1032 another. The PCN working group suggests using admission control for
1033 the following service classes (defined in [<a href="./rfc4594" title="&quot;Configuration Guidelines for DiffServ Service Classes&quot;">RFC4594</a>]):
1034
1035 o Telephony (EF)
1036
1037 o Real-time interactive (CS4)
1038
1039 o Broadcast Video (CS3)
1040
1041 o Multimedia Conferencing (AF4)
1042
1043 For a fuller discussion, see <a href="./rfc6660#appendix-A">Appendix&nbsp;A of [RFC6660]</a>.
1044
1045 <span class="h3"><a class="selflink" id="section-4.3" href="#section-4.3">4.3</a>. Attributes</span>
1046
1047 The purpose of this per-domain behavior is to achieve low loss and
1048 jitter for the target class of traffic. The design requirement for
1049 PCN was that recovery from overloads through the use of flow
1050 termination should happen within 1-3 seconds. PCN probably performs
1051 better than that.
1052
1053 <span class="h3"><a class="selflink" id="section-4.4" href="#section-4.4">4.4</a>. Parameters</span>
1054
1055 The set of parameters that needs to be configured at each PCN-node
1056 and at the Decision Point is described in <a href="#section-5.1">Section 5.1</a>.
1057
1058
1059
1060 <span class="grey">Charny, et al. Experimental [Page 19]</span></pre>
1061 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-20" ></span>
1062 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1063
1064
1065 <span class="h3"><a class="selflink" id="section-4.5" href="#section-4.5">4.5</a>. Assumptions</span>
1066
1067 It is assumed that a specific portion of link capacity has been
1068 reserved for PCN-traffic.
1069
1070 <span class="h3"><a class="selflink" id="section-4.6" href="#section-4.6">4.6</a>. Example Uses</span>
1071
1072 The PCN CL behavior may be used to carry real-time traffic,
1073 particularly voice and video.
1074
1075 <span class="h3"><a class="selflink" id="section-4.7" href="#section-4.7">4.7</a>. Environmental Concerns</span>
1076
1077 The PCN CL per-domain behavior could theoretically interfere with the
1078 use of end-to-end ECN due to reuse of ECN bits for PCN marking.
1079 <a href="./rfc6660#section-5.1">Section&nbsp;5.1 of [RFC6660]</a> describes the actions that can be taken to
1080 protect ECN signaling. <a href="#appendix-B">Appendix B</a> of that document provides further
1081 discussion of how ECN and PCN can coexist.
1082
1083 <span class="h3"><a class="selflink" id="section-4.8" href="#section-4.8">4.8</a>. Security Considerations</span>
1084
1085 Please see the security considerations in [<a href="./rfc5559" title="&quot;Pre-Congestion Notification (PCN) Architecture&quot;">RFC5559</a>] as well as those
1086 in [<a href="./rfc2474" title="&quot;Definition of the Differentiated Services Field (DS Field) in the IPv4 and IPv6 Headers&quot;">RFC2474</a>] and [<a href="./rfc2475" title="&quot;An Architecture for Differentiated Services&quot;">RFC2475</a>].
1087
1088 <span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a>. Operational and Management Considerations</span>
1089
1090 <span class="h3"><a class="selflink" id="section-5.1" href="#section-5.1">5.1</a>. Deployment of the CL Edge Behavior</span>
1091
1092 Deployment of the PCN Controlled Load edge behavior requires the
1093 following steps:
1094
1095 o selection of deployment options and global parameter values;
1096
1097 o derivation of per-node and per-link information;
1098
1099 o installation, but not activation, of parameters and policies at
1100 all of the nodes in the PCN-domain;
1101
1102 o activation and verification of all behaviors.
1103
1104 <span class="h4"><a class="selflink" id="section-5.1.1" href="#section-5.1.1">5.1.1</a>. Selection of Deployment Options and Global Parameters</span>
1105
1106 The first set of decisions affects the operation of the network as a
1107 whole. To begin with, the operator needs to make basic design
1108 decisions such as whether the Decision Point is centralized or
1109 collocated with the PCN-ingress-nodes, and whether per-flow and
1110 aggregate resource signaling as described in [<a href="#ref-RSVP-PCN" title="&quot;Generic Aggregation of Resource ReSerVation Protocol (RSVP) for IPv4 And IPv6 Reservations over PCN domains&quot;">RSVP-PCN</a>] is deployed
1111 in the network. After that, the operator needs to decide:
1112
1113
1114
1115
1116 <span class="grey">Charny, et al. Experimental [Page 20]</span></pre>
1117 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-21" ></span>
1118 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1119
1120
1121 o whether PCN packets will be forwarded unencapsulated or in tunnels
1122 between the PCN-ingress-node and the PCN-egress-node.
1123 Encapsulation preserves incoming ECN settings and simplifies the
1124 PCN-egress-node's job when it comes to relating incoming packets
1125 to specific ingress-egress-aggregates, but lowers the path MTU and
1126 imposes the extra labor of encapsulation/decapsulation on the PCN-
1127 edge-nodes.
1128
1129 o which service classes will be subject to PCN control and what DSCP
1130 will be used for each. (See <a href="./rfc6660#appendix-A">[RFC6660] Appendix&nbsp;A</a> for advice on
1131 this topic.)
1132
1133 o the markings to be used at all nodes in the PCN-domain to indicate
1134 not-marked (NM), [<a href="#ref-CL-specific">CL-specific</a>] threshold-marked (ThM), and excess-
1135 traffic-marked (ETM) PCN packets;
1136
1137 o the marking rules for re-marking PCN-traffic leaving the PCN-
1138 domain;
1139
1140 o whether PCN-based flow admission is enabled;
1141
1142 o whether PCN-based flow termination is enabled.
1143
1144 The following parameters affect the operation of PCN itself. The
1145 operator needs to choose:
1146
1147 o the value of CLE-limit if PCN-based flow admission is enabled.
1148 [<a href="#ref-CL-specific">CL-specific</a>] In practice, the operation of flow admission is not
1149 very sensitive to the value of the CLE-limit, because when
1150 threshold-marking occurs it tends to persist long enough that
1151 threshold-marked traffic becomes a large proportion of the
1152 received traffic in a given interval.
1153
1154 o the value of the collection interval T_meas. For a recommended
1155 range of values, see <a href="#section-3.5.1">Section 3.5.1</a> above.
1156
1157 o whether report suppression is to be enabled at the PCN-egress-
1158 nodes and if so, the values of CLE-reporting-threshold and
1159 T_maxsuppress. It is reasonable to leave CLE-reporting-threshold
1160 at its default value (zero, as specified in <a href="#section-3.2.3">Section 3.2.3</a>). For a
1161 recommended range of values of T_maxsuppress, see <a href="#section-3.5.1">Section 3.5.1</a>
1162 above.
1163
1164 o the value of the duration T_crit, which the Decision Point uses in
1165 deciding whether communications with a given PCN-edge-node have
1166 failed. For a recommended range of values of T_crit, see
1167 <a href="#section-3.5.1">Section 3.5.1</a> above.
1168
1169
1170
1171
1172 <span class="grey">Charny, et al. Experimental [Page 21]</span></pre>
1173 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-22" ></span>
1174 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1175
1176
1177 o [<a href="#ref-CL-specific">CL-specific</a>] Activation/deactivation of recording of individual
1178 flow identifiers when excess-traffic-marked PCN-traffic is
1179 observed. Reporting these identifiers has value only if PCN-based
1180 flow termination is activated and Equal Cost Multi-Path (ECMP)
1181 routing is enabled in the PCN-domain.
1182
1183 <span class="h4"><a class="selflink" id="section-5.1.2" href="#section-5.1.2">5.1.2</a>. Specification of Node- and Link-Specific Parameters</span>
1184
1185 Filters are required at both the PCN-ingress-node and the PCN-egress-
1186 node to classify incoming PCN packets by ingress-egress-aggregate.
1187 Because of the potential use of multipath routing in domains upstream
1188 of the PCN-domain, it is impossible to do such classification
1189 reliably at the PCN-egress-node based on the packet header contents
1190 as originally received at the PCN-ingress-node. (Packets with the
1191 same header contents could enter the PCN-domain at multiple PCN-
1192 ingress-nodes.) As a result, the only way to construct such filters
1193 reliably is to tunnel the packets from the PCN-ingress-node to the
1194 PCN-egress-node.
1195
1196 The PCN-ingress-node needs filters in order to place PCN packets into
1197 the right tunnel in the first instance, and also to satisfy requests
1198 from the Decision Point for admission rates into specific ingress-
1199 egress-aggregates. These filters select the PCN-egress-node, but not
1200 necessarily a specific path through the network to that node. As a
1201 result, they are likely to be stable even in the face of failures in
1202 the network, except when the PCN-egress-node itself becomes
1203 unreachable. If all PCN packets will be tunneled, the PCN-ingress-
1204 node also needs to know the address of the peer PCN-egress-node
1205 associated with each filter.
1206
1207 Operators may wish to give some thought to the provisioning of
1208 alternate egress points for some or all ingress-egress-aggregates in
1209 case of failure of the PCN-egress-node. This could require the
1210 setting up of standby tunnels to these alternate egress points.
1211
1212 Each PCN-egress-node needs filters to classify incoming PCN packets
1213 by ingress-egress-aggregate, in order to gather measurements on a
1214 per-aggregate basis. If tunneling is used, these filters are
1215 constructed on the basis of the identifier of the tunnel from which
1216 the incoming packet has emerged (e.g., the source address in the
1217 outer header if IP encapsulation is used). The PCN-egress-node also
1218 needs to know the address of the Decision Point to which it sends
1219 reports for each ingress-egress-aggregate.
1220
1221
1222
1223
1224
1225
1226
1227
1228 <span class="grey">Charny, et al. Experimental [Page 22]</span></pre>
1229 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-23" ></span>
1230 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1231
1232
1233 A centralized Decision Point needs to have the address of the PCN-
1234 ingress-node corresponding to each ingress-egress-aggregate.
1235 Security considerations require that information also be prepared for
1236 a centralized Decision Point and each PCN-edge-node to allow them to
1237 authenticate each other.
1238
1239 Turning to link-specific parameters, the operator needs to derive
1240 values for the PCN-admissible-rate and [<a href="#ref-CL-specific">CL-specific</a>] PCN-supportable-
1241 rate on each link in the network. The first two paragraphs of
1242 <a href="./rfc5559#section-5.2.2">Section&nbsp;5.2.2 of [RFC5559]</a> discuss how these values may be derived.
1243
1244 <span class="h4"><a class="selflink" id="section-5.1.3" href="#section-5.1.3">5.1.3</a>. Installation of Parameters and Policies</span>
1245
1246 As discussed in the previous two sections, every PCN node needs to be
1247 provisioned with a number of parameters and policies relating to its
1248 behavior in processing incoming packets. The Diffserv MIB [<a href="./rfc3289" title="&quot;Management Information Base for the Differentiated Services Architecture&quot;">RFC3289</a>]
1249 can be useful for this purpose, although it needs to be extended in
1250 some cases. This MIB covers packet classification, metering,
1251 counting, policing, dropping, and marking. The required extensions
1252 specifically include an encapsulation action following
1253 reclassification by ingress-egress-aggregate. In addition, the MIB
1254 has to be extended to include objects for marking the ECN field in
1255 the outer header at the PCN-ingress-node and an extension to the
1256 classifiers to include the ECN field at PCN-interior and PCN-egress-
1257 nodes. Finally, new objects may need to be defined at the PCN-
1258 interior-nodes to represent the metering algorithms for threshold-
1259 marking and packet-size-independent excess-traffic-marking.
1260
1261 Values for the PCN-admissible-rate and [<a href="#ref-CL-specific">CL-specific</a>] PCN-supportable-
1262 rate on each link on a node appear as metering parameters. Operators
1263 should take note of the need to deploy meters of a given type
1264 (threshold or excess-traffic) either on the ingress or the egress
1265 side of each interior link, but not both (Appendix B.2 of [<a href="./rfc5670" title="&quot;Metering and Marking Behaviour of PCN- Nodes&quot;">RFC5670</a>].
1266
1267 The following additional information has to be configured by other
1268 means (e.g., additional MIBs, NETCONF models).
1269
1270 At the PCN-egress-node:
1271
1272 o the measurement interval T_meas (units of ms, range 50 to 1000);
1273
1274 o [<a href="#ref-CL-specific">CL-specific</a>] whether specific flow identifiers must be captured
1275 when excess-traffic-marked packets are observed;
1276
1277 o whether report suppression is to be applied;
1278
1279
1280
1281
1282
1283
1284 <span class="grey">Charny, et al. Experimental [Page 23]</span></pre>
1285 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-24" ></span>
1286 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1287
1288
1289 o if so, the interval T_maxsuppress (units of 100 ms, range 1 to
1290 100) and the CLE-reporting-threshold (units of tenths of one
1291 percent, range 0 to 1000, default value 0);
1292
1293 o the address of the PCN-ingress-node for each ingress-egress-
1294 aggregate, if the Decision Point is collocated with the PCN-
1295 ingress-node and [<a href="#ref-RSVP-PCN" title="&quot;Generic Aggregation of Resource ReSerVation Protocol (RSVP) for IPv4 And IPv6 Reservations over PCN domains&quot;">RSVP-PCN</a>] is not deployed;
1296
1297 o the address of the centralized Decision Point to which it sends
1298 its reports, if there is one.
1299
1300 At the Decision Point:
1301
1302 o whether PCN-based flow admission is enabled;
1303
1304 o whether PCN-based flow termination is enabled;
1305
1306 o the value of CLE-limit (units of tenths of one percent, range 0 to
1307 1000);
1308
1309 o the value of the interval T_crit (units of 100 ms, range 1 to
1310 100);
1311
1312 o whether report suppression is to be applied;
1313
1314 o if so, the interval T_maxsuppress (units of 100 ms, range 1 to
1315 100) and the CLE-reporting-threshold (units of tenths of one
1316 percent, range 0 to 1000, default value 0). These MUST be the
1317 same values that are provisioned in the PCN-egress-nodes;
1318
1319 o if the Decision Point is centralized, the address of the PCN-
1320 ingress-node (and any other information needed to establish a
1321 security association) for each ingress-egress-aggregate.
1322
1323 Depending on the testing strategy, it may be necessary to install the
1324 new configuration data in stages. This is discussed further below.
1325
1326 <span class="h4"><a class="selflink" id="section-5.1.4" href="#section-5.1.4">5.1.4</a>. Activation and Verification of All Behaviors</span>
1327
1328 It is certainly not within the scope of this document to advise on
1329 testing strategy, which operators undoubtedly have well in hand.
1330 Quite possibly an operator will prefer an incremental approach to
1331 activation and testing. Implementing the PCN marking scheme at PCN-
1332 ingress-nodes, corresponding scheduling behavior in downstream nodes,
1333 and re-marking at the PCN-egress-nodes is a large enough step in
1334 itself to require thorough testing before going further.
1335
1336
1337
1338
1339
1340 <span class="grey">Charny, et al. Experimental [Page 24]</span></pre>
1341 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-25" ></span>
1342 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1343
1344
1345 Testing will probably involve the injection of packets at individual
1346 nodes and tracking of how the node processes them. This work can
1347 make use of the counter capabilities included in the Diffserv MIB.
1348 The application of these capabilities to the management of PCN is
1349 discussed in the next section.
1350
1351 <span class="h3"><a class="selflink" id="section-5.2" href="#section-5.2">5.2</a>. Management Considerations</span>
1352
1353 This section focuses on the use of event logging and the use of
1354 counters supported by the Diffserv MIB [<a href="./rfc3289" title="&quot;Management Information Base for the Differentiated Services Architecture&quot;">RFC3289</a>] for the various
1355 monitoring tasks involved in management of a PCN network.
1356
1357 <span class="h4"><a class="selflink" id="section-5.2.1" href="#section-5.2.1">5.2.1</a>. Event Logging in the PCN-Domain</span>
1358
1359 It is anticipated that event logging using SYSLOG [<a href="./rfc5424" title="&quot;The Syslog Protocol&quot;">RFC5424</a>] will be
1360 needed for fault management and potentially for capacity management.
1361 Implementations MUST be capable of generating logs for the following
1362 events:
1363
1364 o detection of loss of contact between a Decision Point and a PCN-
1365 edge-node, as described in <a href="#section-3.3.3">Section 3.3.3</a>;
1366
1367 o successful receipt of a report from a PCN-egress-node, following
1368 detection of loss of contact with that node;
1369
1370 o flow termination events.
1371
1372 All of these logs are generated by the Decision Point. There is a
1373 strong likelihood in the first and third cases that the events are
1374 correlated with network failures at a lower level. This has
1375 implications for how often specific event types should be reported,
1376 so as not to contribute unnecessarily to log buffer overflow.
1377 Recommendations on this topic follow for each event report type.
1378
1379 The field names (e.g., HOSTNAME, STRUCTURED-DATA) used in the
1380 following subsections are defined in [<a href="./rfc5424" title="&quot;The Syslog Protocol&quot;">RFC5424</a>].
1381
1382 <span class="h5"><a class="selflink" id="section-5.2.1.1" href="#section-5.2.1.1">5.2.1.1</a>. Logging Loss and Restoration of Contact</span>
1383
1384 <a href="#section-3.3.3">Section 3.3.3</a> describes the circumstances under which the Decision
1385 Point may determine that it has lost contact, either with a PCN-
1386 ingress-node or a PCN-egress-node, due to failure to receive an
1387 expected report. Loss of contact with a PCN-ingress-node is a case
1388 primarily applicable when the Decision Point is in a separate node.
1389 However, implementations MAY implement logging in the collocated case
1390 if the implementation is such that non-response to a request from the
1391 Decision Point function can occasionally occur due to processor load
1392 or other reasons.
1393
1394
1395
1396 <span class="grey">Charny, et al. Experimental [Page 25]</span></pre>
1397 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-26" ></span>
1398 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1399
1400
1401 The log reporting the loss of contact with a PCN-ingress-node or PCN-
1402 egress-node MUST include the following content:
1403
1404 o The HOSTNAME field MUST identify the Decision Point issuing the
1405 log.
1406
1407 o A STRUCTURED-DATA element MUST be present, containing parameters
1408 identifying the node for which an expected report has not been
1409 received and the type of report lost (ingress or egress). It is
1410 RECOMMENDED that the SD-ID for the STRUCTURED-DATA element have
1411 the form "PCNNode" (without the quotes), which has been registered
1412 with IANA. The node identifier PARAM-NAME is RECOMMENDED to be
1413 "ID" (without the quotes). The identifier itself is subject to
1414 the preferences expressed in <a href="./rfc5424#section-6.2.4">Section&nbsp;6.2.4 of [RFC5424]</a> for the
1415 HOSTNAME field. The report type PARAM-NAME is RECOMMENDED to be
1416 "RTyp" (without the quotes). The PARAM-VALUE for the RTyp field
1417 MUST be either "ingr" or "egr".
1418
1419 The following values are also RECOMMENDED for the indicated fields in
1420 this log, subject to local practice:
1421
1422 o PRI initially set to 115, representing a Facility value of (14)
1423 "log alert" and a Severity level of (3) "Error Condition". Note
1424 that loss of contact with a PCN-egress-node implies that no new
1425 flows will be admitted to one or more ingress-egress-aggregates
1426 until contact is restored. The reason a higher severity level
1427 (lower value) is not proposed for the initial log is because any
1428 corrective action would probably be based on alerts at a lower
1429 subsystem level.
1430
1431 o APPNAME set to "PCN" (without the quotes).
1432
1433 o MSGID set to "LOST" (without the quotes).
1434
1435 If contact is not regained with a PCN-egress-node in a reasonable
1436 period of time (say, one minute), the log SHOULD be repeated, this
1437 time with a PRI value of 113, implying a Facility value of (14) "log
1438 alert" and a Severity value of (1) "Alert: action must be taken
1439 immediately". The reasoning is that by this time, any more general
1440 conditions should have been cleared, and the problem lies
1441 specifically with the PCN-egress-node concerned and the PCN
1442 application in particular.
1443
1444 Whenever a loss-of-contact log is generated for a PCN-egress-node, a
1445 log indicating recovery SHOULD be generated when the Decision Point
1446 next receives a report from the node concerned. The log SHOULD have
1447 the same content as just described for the loss-of-contact log, with
1448 the following differences:
1449
1450
1451
1452 <span class="grey">Charny, et al. Experimental [Page 26]</span></pre>
1453 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-27" ></span>
1454 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1455
1456
1457 o PRI changes to 117, indicating a Facility value of (14) "log
1458 alert" and a Severity of (5) "Notice: normal but significant
1459 condition".
1460
1461 o MSGID changes to "RECVD" (without the quotes).
1462
1463 <span class="h5"><a class="selflink" id="section-5.2.1.2" href="#section-5.2.1.2">5.2.1.2</a>. Logging Flow Termination Events</span>
1464
1465 <a href="#section-3.3.2">Section 3.3.2</a> describes the process whereby the Decision Point
1466 decides that flow termination is required for a given ingress-egress-
1467 aggregate, calculates how much flow to terminate, and selects flows
1468 for termination. This section describes a log that SHOULD be
1469 generated each time such an event occurs. (In the case where
1470 termination occurs in multiple rounds, one log SHOULD be generated
1471 per round.) The log may be useful in fault management, to indicate
1472 the service impact of a fault occurring in a lower-level subsystem.
1473 In the absence of network failures, it may also be used as an
1474 indication of an urgent need to review capacity utilization along the
1475 path of the ingress-egress-aggregate concerned.
1476
1477 The log reporting a flow termination event MUST include the following
1478 content:
1479
1480 o The HOSTNAME field MUST identify the Decision Point issuing the
1481 log.
1482
1483 o A STRUCTURED-DATA element MUST be present, containing parameters
1484 identifying the ingress and egress nodes for the ingress-egress-
1485 aggregate concerned, indicating the total amount of flow being
1486 terminated, and giving the number of flows terminated to achieve
1487 that objective.
1488
1489 It is RECOMMENDED that the SD-ID for the STRUCTURED-DATA element
1490 have the form: "PCNTerm" (without the quotes), which has been
1491 registered with IANA. The parameter identifying the ingress node
1492 for the ingress-egress-aggregate is RECOMMENDED to have PARAM-NAME
1493 "IngrID" (without the quotes). The parameter identifying the
1494 egress node for the ingress-egress-aggregate is RECOMMENDED to
1495 have PARAM-NAME "EgrID" (without the quotes). Both identifiers
1496 are subject to the preferences expressed in <a href="./rfc5424#section-6.2.4">Section&nbsp;6.2.4 of
1497 [RFC5424]</a> for the HOSTNAME field.
1498
1499 The parameter giving the total amount of flow being terminated is
1500 RECOMMENDED to have PARAM-NAME "TermRate" (without the quotes).
1501 The PARAM-VALUE MUST be the target rate as calculated according to
1502 the procedures of <a href="#section-3.3.2">Section 3.3.2</a>, as an integer value in thousands
1503 of octets per second. The parameter giving the number of flows
1504
1505
1506
1507
1508 <span class="grey">Charny, et al. Experimental [Page 27]</span></pre>
1509 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-28" ></span>
1510 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1511
1512
1513 selected for termination is RECOMMENDED to have PARAM-NAME "FCnt"
1514 (without the quotes). The PARAM-VALUE for this parameter MUST be
1515 an integer, the number of flows selected.
1516
1517 The following values are also RECOMMENDED for the indicated fields in
1518 this log, subject to local practice:
1519
1520 o PRI initially set to 116, representing a Facility value of (14)
1521 "log alert" and a Severity level of (4) "Warning: warning
1522 conditions".
1523
1524 o APPNAME set to "PCN" (without the quotes).
1525
1526 o MSGID set to "TERM" (without the quotes).
1527
1528 <span class="h4"><a class="selflink" id="section-5.2.2" href="#section-5.2.2">5.2.2</a>. Provision and Use of Counters</span>
1529
1530 The Diffserv MIB [<a href="./rfc3289" title="&quot;Management Information Base for the Differentiated Services Architecture&quot;">RFC3289</a>] allows for the provision of counters along
1531 the various possible processing paths associated with an interface
1532 and flow direction. It is RECOMMENDED that the PCN-nodes be
1533 instrumented as described below. It is assumed that the cumulative
1534 counts so obtained will be collected periodically for use in
1535 debugging, fault management, and capacity management.
1536
1537 PCN-ingress-nodes SHOULD provide the following counts for each
1538 ingress-egress-aggregate. Since the Diffserv MIB installs counters
1539 by interface and direction, aggregation of counts over multiple
1540 interfaces may be necessary to obtain total counts by ingress-egress-
1541 aggregate. It is expected that such aggregation will be performed by
1542 a central system rather than at the PCN-ingress-node.
1543
1544 o total PCN packets and octets that were received for that ingress-
1545 egress-aggregate but were dropped;
1546
1547 o total PCN packets and octets admitted to that aggregate.
1548
1549 PCN-interior-nodes SHOULD provide the following counts for each
1550 interface, noting that a given packet MUST NOT be counted more than
1551 once as it passes through the node:
1552
1553 o total PCN packets and octets dropped;
1554
1555 o total PCN packets and octets forwarded without re-marking;
1556
1557 o [<a href="#ref-CL-specific">CL-specific</a>] total PCN packets and octets re-marked to threshold-
1558 marked;
1559
1560 o total PCN packets and octets re-marked to excess-traffic-marked.
1561
1562
1563
1564 <span class="grey">Charny, et al. Experimental [Page 28]</span></pre>
1565 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-29" ></span>
1566 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1567
1568
1569 PCN-egress-nodes SHOULD provide the following counts for each
1570 ingress-egress-aggregate. As with the PCN-ingress-node, so with the
1571 PCN-egress-node it is expected that any necessary aggregation over
1572 multiple interfaces will be done by a central system.
1573
1574 o total not-marked PCN packets and octets received;
1575
1576 o [<a href="#ref-CL-specific">CL-specific</a>] total threshold-marked PCN packets and octets
1577 received;
1578
1579 o total excess-traffic-marked PCN packets and octets received.
1580
1581 The following continuously cumulative counters SHOULD be provided as
1582 indicated, but require new MIBs to be defined. If the Decision Point
1583 is not collocated with the PCN-ingress-node, the latter SHOULD
1584 provide a count of the number of requests for PCN-sent-rate received
1585 from the Decision Point and the number of responses returned to the
1586 Decision Point. The PCN-egress-node SHOULD provide a count of the
1587 number of reports sent to each Decision Point. Each Decision Point
1588 SHOULD provide the following:
1589
1590 o total number of requests for PCN-sent-rate sent to each PCN-
1591 ingress-node with which it is not collocated;
1592
1593 o total number of reports received from each PCN-egress-node;
1594
1595 o total number of loss-of-contact events detected for each PCN-
1596 boundary-node;
1597
1598 o total cumulative duration of "block" state in hundreds of
1599 milliseconds for each ingress-egress-aggregate;
1600
1601 o total number of rounds of flow termination exercised for each
1602 ingress-egress-aggregate.
1603
1604 <span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a>. Security Considerations</span>
1605
1606 [<a id="ref-RFC5559">RFC5559</a>] provides a general description of the security
1607 considerations for PCN. This memo introduces one new consideration,
1608 related to the use of a centralized Decision Point. The Decision
1609 Point itself is a trusted entity. However, its use implies the
1610 existence of an interface on the PCN-ingress-node through which
1611 communication of policy decisions takes place. That interface is a
1612 point of vulnerability that must be protected from denial-of-service
1613 attacks.
1614
1615
1616
1617
1618
1619
1620 <span class="grey">Charny, et al. Experimental [Page 29]</span></pre>
1621 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-30" ></span>
1622 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1623
1624
1625 <span class="h2"><a class="selflink" id="section-7" href="#section-7">7</a>. IANA Considerations</span>
1626
1627 IANA has added the following entries to the "syslog Structured Data
1628 ID Values" registry.
1629
1630 Structured Data ID: PCNNode OPTIONAL
1631
1632 Structured Data Parameter: ID MANDATORY
1633
1634 Structured Data Parameter: Rtyp MANDATORY
1635
1636 Reference: <a href="./rfc6661">RFC 6661</a>
1637
1638 Structured Data ID: PCNTerm OPTIONAL
1639
1640 Structured Data Parameter: IngrID MANDATORY
1641
1642 Structured Data Parameter: EgrID MANDATORY
1643
1644 Structured Data Parameter: TermRate MANDATORY
1645
1646 Structured Data Parameter: FCnt MANDATORY
1647
1648 Reference: <a href="./rfc6661">RFC 6661</a>
1649
1650
1651 <span class="h2"><a class="selflink" id="section-8" href="#section-8">8</a>. Acknowledgements</span>
1652
1653 The content of this memo bears a family resemblance to [<a href="#ref-Briscoe-CL" title="&quot;An edge- to-edge Deployment Model for Pre-Congestion Notification: Admission Control over a DiffServ Region&quot;">Briscoe-CL</a>].
1654 The authors of that document were Bob Briscoe, Philip Eardley, and
1655 Dave Songhurst of BT, Anna Charny and Francois Le Faucheur of Cisco,
1656 Jozef Babiarz, Kwok Ho Chan, and Stephen Dudley of Nortel, Giorgios
1657 Karagiannis of U. Twente and Ericsson, and Attila Bader and Lars
1658 Westberg of Ericsson.
1659
1660 Ruediger Geib, Philip Eardley, and Bob Briscoe have helped to shape
1661 the present document with their comments. Toby Moncaster gave a
1662 careful review to get it into shape for Working Group Last Call.
1663
1664 Amongst the authors, Michael Menth deserves special mention for his
1665 constant and careful attention to both the technical content of this
1666 document and the manner in which it was expressed.
1667
1668 David Harrington's careful AD review resulted not only in necessary
1669 changes throughout the document, but also the addition of the
1670 operations and management considerations (<a href="#section-5">Section 5</a>).
1671
1672
1673
1674
1675
1676 <span class="grey">Charny, et al. Experimental [Page 30]</span></pre>
1677 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-31" ></span>
1678 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1679
1680
1681 As part of the broader review process, the document saw further
1682 improvements as a result of comments by Joel Halpern, Brian
1683 Carpenter, Stephen Farrell, Sean Turner, and Pete Resnick.
1684
1685 <span class="h2"><a class="selflink" id="section-9" href="#section-9">9</a>. References</span>
1686
1687 <span class="h3"><a class="selflink" id="section-9.1" href="#section-9.1">9.1</a>. Normative References</span>
1688
1689 [<a id="ref-RFC2119">RFC2119</a>] Bradner, S., "Key words for use in RFCs to Indicate
1690 Requirement Levels", <a href="https://www.rfc-editor.org/bcp/bcp14">BCP 14</a>, <a href="./rfc2119">RFC 2119</a>, March 1997.
1691
1692 [<a id="ref-RFC2474">RFC2474</a>] Nichols, K., Blake, S., Baker, F., and D. Black,
1693 "Definition of the Differentiated Services Field (DS
1694 Field) in the IPv4 and IPv6 Headers", <a href="./rfc2474">RFC 2474</a>,
1695 December 1998.
1696
1697 [<a id="ref-RFC2475">RFC2475</a>] Blake, S., Black, D., Carlson, M., Davies, E., Wang,
1698 Z., and W. Weiss, "An Architecture for Differentiated
1699 Services", <a href="./rfc2475">RFC 2475</a>, December 1998.
1700
1701 [<a id="ref-RFC3086">RFC3086</a>] Nichols, K. and B. Carpenter, "Definition of
1702 Differentiated Services Per Domain Behaviors and Rules
1703 for their Specification", <a href="./rfc3086">RFC 3086</a>, April 2001.
1704
1705 [<a id="ref-RFC3289">RFC3289</a>] Baker, F., Chan, K., and A. Smith, "Management
1706 Information Base for the Differentiated Services
1707 Architecture", <a href="./rfc3289">RFC 3289</a>, May 2002.
1708
1709 [<a id="ref-RFC5424">RFC5424</a>] Gerhards, R., "The Syslog Protocol", <a href="./rfc5424">RFC 5424</a>,
1710 March 2009.
1711
1712 [<a id="ref-RFC5559">RFC5559</a>] Eardley, P., "Pre-Congestion Notification (PCN)
1713 Architecture", <a href="./rfc5559">RFC 5559</a>, June 2009.
1714
1715 [<a id="ref-RFC5670">RFC5670</a>] Eardley, P., "Metering and Marking Behaviour of PCN-
1716 Nodes", <a href="./rfc5670">RFC 5670</a>, November 2009.
1717
1718 [<a id="ref-RFC6660">RFC6660</a>] Briscoe, B., Moncaster, T., and M. Menth, "Encoding
1719 Three Pre-Congestion Notification (PCN) States in the
1720 IP Header Using a Single Diffserv Codepoint (DSCP)",
1721 <a href="./rfc6660">RFC 6660</a>, July 2012.
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732 <span class="grey">Charny, et al. Experimental [Page 31]</span></pre>
1733 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-32" ></span>
1734 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1735
1736
1737 <span class="h3"><a class="selflink" id="section-9.2" href="#section-9.2">9.2</a>. Informative References</span>
1738
1739 [<a id="ref-Briscoe-CL">Briscoe-CL</a>] Briscoe, B., Eardley, P., Songhurst, D., Le Faucheur,
1740 F., Charny, A., Babiarz, J., Chan, K., Dudley, S.,
1741 Karagiannis, G., Bader, A., and L. Westberg, "An edge-
1742 to-edge Deployment Model for Pre-Congestion
1743 Notification: Admission Control over a DiffServ
1744 Region", Work in Progress, October 2006.
1745
1746 [<a id="ref-MeLe10">MeLe10</a>] Menth, M. and F. Lehrieder, "PCN-Based Measured Rate
1747 Termination", Computer Networks Journal (Elsevier) vol.
1748 54, no. 13, pp. 2099-2116, September 2010.
1749
1750 [<a id="ref-MeLe12">MeLe12</a>] Menth, M. and F. Lehrieder, "Performance of PCN-Based
1751 Admission Control under Challenging Conditions", IEEE/
1752 ACM Transactions on Networking, vol. 20, no. 2,
1753 April 2012.
1754
1755 [<a id="ref-RFC4594">RFC4594</a>] Babiarz, J., Chan, K., and F. Baker, "Configuration
1756 Guidelines for DiffServ Service Classes", <a href="./rfc4594">RFC 4594</a>,
1757 August 2006.
1758
1759 [<a id="ref-RFC6662">RFC6662</a>] Charny, A., Zhang, J., Karagiannis, G., Menth, M., and
1760 T. Taylor, Ed., "Pre-Congestion Notification (PCN)
1761 Boundary Node Behavior for the Single Marking (SM) Mode
1762 of Operation", <a href="./rfc6662">RFC 6662</a>, July 2012.
1763
1764 [<a id="ref-RSVP-PCN">RSVP-PCN</a>] Karagiannis, G. and A. Bhargava, "Generic Aggregation
1765 of Resource ReSerVation Protocol (RSVP) for IPv4 And
1766 IPv6 Reservations over PCN domains", Work in Progress,
1767 July 2012.
1768
1769 [<a id="ref-Satoh10">Satoh10</a>] Satoh, D. and H. Ueno, "Cause and Countermeasure of
1770 Overtermination for PCN-Based Flow Termination",
1771 Proceedings of IEEE Symposium on Computers and
1772 Communications (ISCC '10), pp. 155-161,
1773 Riccione, Italy, June 2010.
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788 <span class="grey">Charny, et al. Experimental [Page 32]</span></pre>
1789 <hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-33" ></span>
1790 <span class="grey"><a href="./rfc6661">RFC 6661</a> PCN CL Boundary-Node Behavior July 2012</span>
1791
1792
1793 Authors' Addresses
1794
1795 Anna Charny
1796 USA
1797
1798 EMail: [email protected]
1799
1800
1801 Fortune Huang
1802 Huawei Technologies
1803 Section F, Huawei Industrial Base,
1804 Bantian Longgang, Shenzhen 518129
1805 P.R. China
1806
1807 Phone: +86 15013838060
1808 EMail: [email protected]
1809
1810
1811 Georgios Karagiannis
1812 University of Twente
1813 P.O. Box 217
1814 7500 AE Enschede,
1815 The Netherlands
1816
1817 Phone: +31 53 4894099
1818 EMail: [email protected]
1819
1820
1821 Michael Menth
1822 University of Tuebingen
1823 Sand 13
1824 72076 Tuebingen
1825 Germany
1826
1827 Phone: +49-7071-2970505
1828 EMail: [email protected]
1829
1830
1831 Tom Taylor (editor)
1832 Huawei Technologies
1833 Ottawa
1834 Canada
1835
1836 EMail: [email protected]
1837
1838
1839
1840
1841
1842
1843
1844 Charny, et al. Experimental [Page 33]
1845 </pre>