idnits 2.17.1 draft-ietf-pcn-cl-edge-behaviour-08.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (December 16, 2010) is 4880 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Missing Reference: 'CL-specific' is mentioned on line 776, but not defined == Missing Reference: 'CL-Specific' is mentioned on line 298, but not defined == Missing Reference: 'CLE-specific' is mentioned on line 417, but not defined ** Obsolete normative reference: RFC 5696 (Obsoleted by RFC 6660) Summary: 1 error (**), 0 flaws (~~), 4 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Engineering Task Force A. Charny 3 Internet-Draft Cisco Systems 4 Intended status: Informational F. Huang 5 Expires: June 19, 2011 Huawei Technologies 6 G. Karagiannis 7 U. Twente 8 M. Menth 9 University of Tuebingen 10 T. Taylor, Ed. 11 Huawei Technologies 12 December 16, 2010 14 PCN Boundary Node Behaviour for the Controlled Load (CL) Mode of 15 Operation 16 draft-ietf-pcn-cl-edge-behaviour-08 18 Abstract 20 Pre-congestion notification (PCN) is a means for protecting the 21 quality of service for inelastic traffic admitted to a Diffserv 22 domain. The overall PCN architecture is described in RFC 5559. This 23 memo is one of a series describing possible boundary node behaviours 24 for a PCN-domain. The behaviour described here is that for a form of 25 measurement-based load control using three PCN marking states, not- 26 marked, threshold-marked, and excess-traffic-marked. This behaviour 27 is known informally as the Controlled Load (CL) PCN-boundary-node 28 behaviour. 30 Status of this Memo 32 This Internet-Draft is submitted in full conformance with the 33 provisions of BCP 78 and BCP 79. 35 Internet-Drafts are working documents of the Internet Engineering 36 Task Force (IETF). Note that other groups may also distribute 37 working documents as Internet-Drafts. The list of current Internet- 38 Drafts is at http://datatracker.ietf.org/drafts/current/. 40 Internet-Drafts are draft documents valid for a maximum of six months 41 and may be updated, replaced, or obsoleted by other documents at any 42 time. It is inappropriate to use Internet-Drafts as reference 43 material or to cite them other than as "work in progress." 45 This Internet-Draft will expire on June 19, 2011. 47 Copyright Notice 48 Copyright (c) 2010 IETF Trust and the persons identified as the 49 document authors. All rights reserved. 51 This document is subject to BCP 78 and the IETF Trust's Legal 52 Provisions Relating to IETF Documents 53 (http://trustee.ietf.org/license-info) in effect on the date of 54 publication of this document. Please review these documents 55 carefully, as they describe your rights and restrictions with respect 56 to this document. Code Components extracted from this document must 57 include Simplified BSD License text as described in Section 4.e of 58 the Trust Legal Provisions and are provided without warranty as 59 described in the Simplified BSD License. 61 Table of Contents 63 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 64 1.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 4 65 2. [CL-Specific] Assumed Core Network Behaviour for CL . . . . . 8 66 3. Node Behaviours . . . . . . . . . . . . . . . . . . . . . . . 8 67 3.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . 8 68 3.2. Behaviour of the PCN-Egress-Node . . . . . . . . . . . . . 9 69 3.2.1. Data Collection . . . . . . . . . . . . . . . . . . . 9 70 3.2.2. Reporting the PCN Data . . . . . . . . . . . . . . . . 10 71 3.2.3. Optional Report Suppression . . . . . . . . . . . . . 10 72 3.3. Behaviour at the Decision Point . . . . . . . . . . . . . 11 73 3.3.1. Flow Admission . . . . . . . . . . . . . . . . . . . . 11 74 3.3.2. Flow Termination . . . . . . . . . . . . . . . . . . . 12 75 3.3.3. Decision Point Action For Missing 76 PCN-Boundary-Node Reports . . . . . . . . . . . . . . 13 77 3.4. Behaviour of the Ingress Node . . . . . . . . . . . . . . 14 78 3.5. Summary of Timers . . . . . . . . . . . . . . . . . . . . 14 79 4. Identifying Ingress and Egress Nodes For PCN Traffic . . . . . 15 80 5. Specification of Diffserv Per-Domain Behaviour . . . . . . . . 15 81 5.1. Applicability . . . . . . . . . . . . . . . . . . . . . . 16 82 5.2. Technical Specification . . . . . . . . . . . . . . . . . 16 83 5.2.1. Classification and Traffic Conditioning . . . . . . . 16 84 5.2.2. PHB Configuration . . . . . . . . . . . . . . . . . . 16 85 5.3. Attributes . . . . . . . . . . . . . . . . . . . . . . . . 17 86 5.4. Parameters . . . . . . . . . . . . . . . . . . . . . . . . 17 87 5.5. Assumptions . . . . . . . . . . . . . . . . . . . . . . . 18 88 5.6. Example Uses . . . . . . . . . . . . . . . . . . . . . . . 18 89 5.7. Environmental Concerns . . . . . . . . . . . . . . . . . . 19 90 5.8. Security Considerations . . . . . . . . . . . . . . . . . 19 91 6. Security Considerations . . . . . . . . . . . . . . . . . . . 19 92 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 19 93 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 19 94 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 19 95 9.1. Normative References . . . . . . . . . . . . . . . . . . . 19 96 9.2. Informative References . . . . . . . . . . . . . . . . . . 20 97 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 21 99 1. Introduction 101 The objective of Pre-Congestion Notification (PCN) is to protect the 102 quality of service (QoS) of inelastic flows within a Diffserv domain, 103 in a simple, scalable, and robust fashion. Two mechanisms are used: 104 admission control, to decide whether to admit or block a new flow 105 request, and (in abnormal circumstances) flow termination to decide 106 whether to terminate some of the existing flows. To achieve this, 107 the overall rate of PCN-traffic is metered on every link in the PCN- 108 domain, and PCN-packets are appropriately marked when certain 109 configured rates are exceeded. These configured rates are below the 110 rate of the link thus providing notification to PCN-boundary-nodes 111 about incipient overloads before any congestion occurs (hence the 112 "pre" part of "pre-congestion notification"). The level of marking 113 allows decisions to be made about whether to admit or terminate PCN- 114 flows. For more details see [RFC5559]. 116 PCN-boundary-node behaviours specify a detailed set of algorithms and 117 procedures used to implement the PCN mechanisms. Since the 118 algorithms depend on specific metering and marking behaviour at the 119 interior nodes, it is also necessary to specify the assumptions made 120 about PCN-interior-node behaviour. Finally, because PCN uses DSCP 121 values to carry its markings, a specification of PCN-boundary-node 122 behaviour MUST include the per domain behaviour (PDB) template 123 specified in [RFC3086], filled out with the appropriate content. The 124 present document accomplishes these tasks for the Controlled Load 125 (CL) mode of operation. 127 [RFC EDITOR'S NOTE: you may choose to delete the following paragraph 128 and the "[CL-specific]" tags throughout this document when publishing 129 it, since they are present primarily to aid reviewers. RFCyyyy is 130 the published version of draft-ietf-pcn-sm-edge-behaviour.] 132 A companion document [RFCyyyy] specifies the Single Marking (SM) PCN- 133 boundary-node behaviour. This document and [RFCyyyy] have a great 134 deal of text in common. To simplify the task of the reader, the text 135 in the present document that is specific to the CL PCN-boundary-node 136 behaviour is preceded by the phrase: "[CL-specific]". A similar 137 distinction for SM-specific text is made in [RFCyyyy]. 139 1.1. Terminology 141 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 142 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 143 document are to be interpreted as described in [RFC2119]. 145 This document uses the following terms defined in Section 2 of 146 [RFC5559]: 148 o PCN-domain; 150 o PCN-ingress-node; 152 o PCN-egress-node; 154 o PCN-interior-node; 156 o PCN-boundary-node; 158 o PCN-flow; 160 o ingress-egress-aggregate (IEA); 162 o [CL-specific] PCN-threshold-rate; 164 o PCN-excess-rate; 166 o PCN-admissible-rate; 168 o PCN-supportable-rate; 170 o PCN-marked; 172 o [CL-specific] threshold-marked; 174 o excess-traffic-marked. 176 It also uses the following terms, for which the definition is 177 repeated from [RFC5559] because of their importance to the 178 understanding of the text that follows: 180 PCN-traffic, PCN-packets, PCN-BA 181 A PCN-domain carries traffic of different Diffserv behaviour 182 aggregates (BAs) [RFC2474]. The PCN-BA uses the PCN mechanisms to 183 carry PCN-traffic, and the corresponding packets are PCN-packets. 184 The same network will carry traffic of other Diffserv BAs. The 185 PCN-BA is distinguished by a combination of the Diffserv codepoint 186 and ECN fields. 188 This document uses the following terms from [RFC5670]: 190 o [CL-specific] threshold-meter; 192 o excess-traffic-meter. 194 To complete the list of borrowed terms, this document reuses the 195 following terms and abbreviations defined in Section 3 of [RFC5696]: 197 o not-PCN codepoint; 199 o Not-marked (NM) codepoint; 201 o PCN-marked (PM) codepoint; 203 o EXP (experimental) [codepoint]. 205 This document defines the following additional terms: 207 Decision Point 208 The node that makes the decision about which flows to admit and to 209 terminate. In a given network deployment, this can be the PCN- 210 ingress-node or a centralized control node. Regardless of the 211 location of the Decision Point, the PCN-ingress-node is the point 212 where the decisions are enforced. 214 NM-rate 215 The rate of not-marked PCN-traffic received at a PCN-egress-node 216 for a given ingress-egress-aggregate in octets per second. For 217 further details see Section 3.2.1. 219 [CL-specific] ThM-rate 220 The rate of threshold-marked PCN-traffic received at a PCN-egress- 221 node for a given ingress-egress-aggregate in octets per second. 222 For further details see Section 3.2.1. 224 ETM-rate 225 The rate of excess-traffic-marked PCN-traffic received at a PCN- 226 egress-node for a given ingress-egress-aggregate in octets per 227 second. For further details see Section 3.2.1. 229 PCN-sent-rate 230 The rate of PCN-traffic received at a PCN-ingress-node and 231 destined for a given ingress-egress-aggregate in octets per 232 second. For further details see Section 3.4. 234 Congestion level estimate (CLE) 235 A value derived from the measurement of PCN-packets received at a 236 PCN-egress-node for a given ingress-egress-aggregate, representing 237 the ratio of marked to total PCN-traffic (measured in octets) 238 received over a short period. The CLE is used to derive the PCN- 239 admission-state (Section 3.3.1) and also by the report suppression 240 procedure (Section 3.2.3) if report suppression is activated. 242 PCN-admission-state 243 The state ("admit" or "block") derived by the Decision Point for a 244 given ingress-egress-aggregate based on PCN packet marking 245 statistics. The Decision Point decides to admit or block new 246 flows offered to the aggregate based on the current value of the 247 PCN-admission-state. For further details see Section 3.3.1. 249 Sustainable aggregate rate (SAR) 250 The estimated maximum rate of PCN-traffic that can be admitted to 251 a given ingress-egress-aggregate at a given moment without risking 252 degradation of quality of service for the admitted flows. The 253 intention is that if the PCN-sent-rate of every ingress-egress- 254 aggregate passing through a given link is limited to its 255 sustainable aggregate rate, the total rate of PCN-traffic flowing 256 through the link will be limited to the PCN-supportable-rate for 257 that link. An estimate of the sustainable aggregate rate for a 258 given ingress-egress-aggregate is derived as part of the flow 259 termination procedure, and is used to determine how much PCN- 260 traffic needs to be terminated. For further details see 261 Section 3.3.2. 263 CLE-reporting-threshold 264 A configurable value against which the CLE is compared as part of 265 the report suppression procedure. For further details, see 266 Section 3.2.3. 268 CLE-limit 269 A configurable value against which the CLE is compared in order to 270 derive the PCN-admission-state for a given ingress-egress- 271 aggregate. For further details, see Section 3.3.1. 273 T-meas 274 An interval, the value of which is configurable, defining the 275 measurement period at the PCN-egress-node during which statistics 276 relating to PCN-traffic marking are collected. At the end of the 277 interval the values NM-rate, [CL-specific] ThM-rate, and ETM-rate 278 as defined above are calculated and a report is sent to the 279 Decision Point, subject to the operation of the report suppression 280 feature. For further details see Section 3.2. 282 T-maxsuppress 283 An interval, the value of which is configurable, after which the 284 PCN-egress-node MUST send a report to the Decision Point for a 285 given ingress-egress-aggregate regardless of the most recent 286 values of the CLE. This is used as a keep-alive mechanism for 287 signalling between the PCN-egress-node and the Decision Point when 288 report suppression is activated. For further details, see 289 Section 3.2.3. 291 T-fail 292 An interval, the value of which is configurable, after which the 293 Decision Point concludes that communication from a given PCN- 294 egress-node has failed if it has received no reports from the PCN- 295 egress-node during that interval. For further details see 296 Section 3.3.3. 298 2. [CL-Specific] Assumed Core Network Behaviour for CL 300 This section describes the assumed behaviour for nodes of the PCN- 301 domain when acting in their role as PCN-interior-nodes. The CL mode 302 of operation assumes that: 304 o PCN-interior-nodes perform both threshold-marking and excess- 305 traffic-marking of packets, according to the rules specified in 306 [RFC5670]; 308 o excess-traffic-marking of packets uses the PCN-Marked (PM) 309 codepoint defined in [RFC5696]; 311 o threshold-marking of packets uses the EXP codepoint defined in 312 [RFC5696]; 314 o the PCN-domain satisfies the conditions specified in [RFC5696]; 316 o on each link the reference rate for the threshold-meter is 317 configured to be equal to the PCN-admissible-rate for the link; 319 o on each link the reference rate for the excess-traffic-meter is 320 configured to be equal to the PCN-supportable-rate for the link; 322 o the set of valid codepoint transitions is as shown in Section 4.2 323 of [RFC5696]. 325 3. Node Behaviours 327 3.1. Overview 329 This section describes the behaviour of the PCN-ingress-node, PCN- 330 egress-node, and the Decision Point (which MAY be collocated with the 331 PCN-ingress-node). 333 The PCN-egress-node collects the rates of not-marked, [CL-specific] 334 threshold-marked, and excess-traffic-marked PCN-traffic for each 335 ingress-egress-aggregate and reports them to the Decision Point. 336 [CL-specific] It MAY also identify and report PCN-flows that have 337 experienced excess-traffic-marking. For a detailed description, see 338 Section 3.2. 340 The PCN-ingress-node enforces flow admission and termination 341 decisions. It also reports the rate of PCN-traffic sent to a given 342 ingress-egress-aggregate when requested by the Decision Point. For 343 details, see Section 3.4. 345 Finally, the Decision Point makes flow admission decisions and 346 selects flows to terminate based on the information provided by the 347 PCN-ingress-node and PCN-egress-node for a given ingress-egress- 348 aggregate. For details, see Section 3.3. 350 3.2. Behaviour of the PCN-Egress-Node 352 3.2.1. Data Collection 354 The PCN-egress-node MUST meter received PCN-traffic in order to 355 derive periodically the following rates for each ingress-egress- 356 aggregate passing through it: 358 o NM-rate: octets per second of PCN-traffic in PCN-packets that are 359 not-marked (i.e., marked with the NM codepoint); 361 o [CL-specific] ThM-rate: octets per second of PCN-traffic in PCN- 362 packets that are threshold-marked (i.e., marked with the PM 363 codepoint); 365 o [CL-specific] ETM-rate: octets per second of PCN-traffic in PCN- 366 packets that are excess-traffic-marked (i.e., marked with the EXP 367 codepoint). 369 The PCN-traffic SHOULD be metered continuously and the measurement 370 intervals themselves SHOULD be of equal length, to minimize the 371 statistical variance introduced by the measurement process itself. 372 The starting and ending times of the measurement intervals for 373 different ingress-egress-aggregates MAY be the same or MAY be 374 different. 376 [CL-specific] As a configurable option, the PCN-egress-node MAY 377 record flow identifiers of the PCN-flows for which excess-traffic- 378 marked packets have been observed. These can be used by the Decision 379 Point when it selects flows for termination. 381 In networks using multipath routing it is possible that congestion 382 is not occurring on all paths carrying a given ingress-egress- 383 aggregate. Assuming that specific PCN-flows are routed via 384 specific paths, identifying the PCN-flows that are experiencing 385 excess-traffic-marking helps to avoid termination of PCN-flows not 386 contributing to congestion. 388 3.2.2. Reporting the PCN Data 390 If the report suppression option described in the next sub-section is 391 not activated, the PCN-egress-node MUST report the latest values of 392 NM-rate, [CL-specific] ThM-rate, and ETM-rate to the Decision Point 393 each time that it calculates them. 395 [CL-specific] If so configured (e.g., because multipath routing is 396 being used, as explained in the previous section), the PCN-egress- 397 node MUST also report the set of flow identifiers of PCN-flows for 398 which excess-traffic-marking was observed in the most recent 399 measurement interval. If this set is large, the PCN-egress-node MAY 400 report only the most recently excess-traffic-marked PCN-flows rather 401 than the complete set. 403 3.2.3. Optional Report Suppression 405 Report suppression MUST be provided as a configurable option, along 406 with two configurable parameters, the CLE-reporting-threshold and the 407 maximum report suppression interval T-maxsuppress. The default value 408 of the CLE-reporting-threshold is zero. T-maxsuppress functions as a 409 keep-alive mechanism for signalling between the PCN-egress-node and 410 the Decision Point. 412 If the report suppression option is enabled, the PCN-egress-node MUST 413 apply the following procedure to decide whether to send a report to 414 the Decision Point, rather than sending a report automatically at the 415 end of each measurement interval. 417 1. As well as the quantities NM-rate, [CLE-specific] ThM-rate, and 418 ETM-rate, the PCN-egress-node MUST calculate the congestion level 419 estimate (CLE) for each measurement interval. The CLE is 420 computed as: 422 [CL-specific] 423 CLE = (ThM-rate + ETM-rate) / (NM-rate + ThM-rate + ETM-rate) 425 if any PCN-traffic was observed, or CLE = 0 if all the rates are 426 zero. 428 2. If the calculated CLE for the latest measurement interval is 429 greater than the CLE-reporting-threshold and/or the calculated 430 CLE for the immediately previous interval was greater than the 431 CLE-reporting-threshold, then the PCN-egress-node MUST send a 432 report to the Decision Point. The contents of the report are 433 described below. 435 3. If an interval T-maxsuppress has elapsed since the last report 436 was sent to the Decision Point, then the PCN-egress-node MUST 437 send a report to the Decision Point regardless of the CLE value. 439 4. If neither of the preceding conditions holds, the PCN-egress-node 440 MUST NOT send a report for the latest measurement interval. 442 Each report sent to the Decision Point when report suppression has 443 been activated MUST contain the values of NM-rate, [CL-specific] ThM- 444 rate, ETM-rate, and CLE that were calculated for the most recent 445 measurement interval. [CL-specific] If so configured, the PCN- 446 egress-node MUST also report the set of flow identifiers of PCN-flows 447 for which excess-traffic-marking was observed in the most recent 448 measurement interval. 450 The above procedure ensures that at least one report is sent per 451 interval (T-maxsuppress + T-meas). This provides some protection 452 against loss of egress reports and also demonstrates to the Decision 453 Point that both the PCN-egress-node and the communication path 454 between that node and the Decision Point are in operation. 456 3.3. Behaviour at the Decision Point 458 Operators can choose to use PCN procedures just for flow admission, 459 or just for flow termination, or for both. A compliant Decision 460 Point MUST implement both mechanisms, but configurable options MUST 461 be provided to activate or deactivate PCN-based flow admission and 462 flow termination independently of each other at a given Decision 463 Point. 465 If PCN-based flow termination is enabled but PCN-based flow admission 466 is not, flow termination operates as specified in this document. 467 Logically, some other system of flow admission control is in 468 operation, but the description of such a system is out of scope of 469 this document and depends on local arrangements. 471 3.3.1. Flow Admission 473 The Decision Point determines the PCN-admission-state for a given 474 ingress-egress-aggregate each time it receives a report from the 475 egress node. It makes this determination on the basis of the 476 congestion level estimate (CLE). If the CLE is provided in the 477 egress node report, the Decision Point SHOULD use the reported value. 478 If the CLE was not provided in the report, the Decision Point MUST 479 calculate it based on the other values provided in the report, using 480 the formula: 482 [CL-specific] 483 CLE = (ThM-rate + ETM-rate) / (NM-rate + ThM-rate + ETM-rate) 485 if any PCN-traffic was observed, or CLE = 0 if all the rates are 486 zero. 488 The Decision Point MUST compare the reported or calculated CLE to a 489 configurable value, the CLE-limit. If the CLE is less than the CLE- 490 limit, the PCN-admission-state for that aggregate MUST be set to 491 "admit"; otherwise it MUST be set to "block". 493 [CL-specific] The outcome of the comparison is not very sensitive 494 to the value of the CLE-limit in practice, because when threshold- 495 marking occurs it tends to persist long enough that threshold- 496 marked traffic becomes a large proportion of the received traffic 497 in a given interval. 499 If the PCN-admission-state for a given ingress-egress-aggregate is 500 "admit", the Decision Point SHOULD allow new flows to be admitted to 501 that aggregate. If the PCN-admission-state for a given ingress- 502 egress-aggregate is "block", the Decision Point SHOULD NOT allow new 503 flows to be admitted to that aggregate. These actions MAY be 504 modified by policy in specific cases, but such policy intervention 505 risks defeating the purpose of using PCN. 507 3.3.2. Flow Termination 509 [CL-specific] When the report from the PCN-egress-node includes a 510 non-zero value of the ETM-rate for some ingress-egress-aggregate, the 511 Decision Point MUST request the PCN-ingress-node to provide an 512 estimate of the rate (PCN-sent-rate) at which the PCN-ingress-node is 513 receiving PCN-traffic that is destined for the given ingress-egress- 514 aggregate. 516 If the Decision Point is collocated with the PCN-ingress-node, the 517 request and response are internal operations. 519 The Decision Point MUST then wait, for both the requested rate from 520 the PCN-ingress-node and the next report from the PCN-egress-node for 521 the ingress-egress-aggregate concerned. If this next egress node 522 report also includes a non-zero value for the ETM-rate, the Decision 523 Point MUST determine an amount of flow to terminate using the 524 following steps: 526 1. [CL-specific] The sustainable aggregate rate (SAR) for the given 527 ingress-egress-aggregate is estimated by the sum: 529 SAR = NM-rate + ThM-rate 531 for the latest reported interval. 533 2. The amount of traffic to be terminated is the difference: 535 PCN-sent-rate - SAR, 537 where PCN-sent-rate is the value provided by the PCN-ingress- 538 node. 540 If the difference calculated in the second step is positive, the 541 Decision Point SHOULD select PCN-flows to terminate, until it 542 determines that the PCN-traffic admission rate will no longer be 543 greater than the estimated sustainable aggregate rate. If the 544 Decision Point knows the bandwidth required by individual PCN-flows 545 (e.g., from resource signalling used to establish the flows), it MAY 546 choose to complete its selection of PCN-flows to terminate in a 547 single round of decisions. 549 Alternatively, the Decision Point MAY spread flow termination over 550 multiple rounds to avoid over-termination. If this is done, it is 551 RECOMMENDED that enough time elapse between successive rounds of 552 termination to allow the effects of previous rounds to be reflected 553 in the measurements upon which the termination decisions are based 554 (see [IEEE-Satoh] and sections 4.2 and 4.3 of [MeLe10]). 556 In general, the selection of flows for termination MAY be guided by 557 policy. [CL-specific] If the egress node has supplied a list of 558 identifiers of PCN-flows that experienced excess-traffic-marking 559 (Section 3.2), the Decision Point SHOULD first consider terminating 560 PCN-flows in that list. 562 3.3.3. Decision Point Action For Missing PCN-Boundary-Node Reports 564 If the Decision Point fails to receive any report from a given PCN- 565 egress-node for a configurable interval T-fail, it SHOULD raise an 566 alarm to management. A Decision Point collocated with a PCN-ingress- 567 node SHOULD cease to admit PCN-flows to the ingress-egress-aggregate 568 passing from the PCN-ingress-node to the given PCN-egress-node, until 569 it again receives a report from that node. A centralized Decision 570 Point MAY cease to admit PCN-flows to all ingress-egress-aggregates 571 destined to the PCN-egress-node concerned, until it again receives a 572 report from that node. 574 If a centralized Decision Point fails to receive a reply within a 575 reasonable period of time to a request for a PCN-sent-rate value sent 576 to a given PCN-ingress-node, it SHOULD raise an alarm to management. 578 3.4. Behaviour of the Ingress Node 580 The PCN-ingress-node MUST provide the estimated current rate of PCN- 581 traffic received at that node and destined for a given ingress- 582 egress-aggregate in octets per second (the PCN-sent-rate) when the 583 Decision Point requests it. The way this rate estimate is derived is 584 a matter of implementation. 586 For example, the rate that the PCN-ingress-node supplies MAY be 587 based on a quick sample taken at the time the information is 588 required. It is RECOMMENDED that such a sample be based on 589 observation of at least thirty PCN-packets to achieve reasonable 590 statistical reliability. 592 3.5. Summary of Timers 594 Table 1 summarizes the timers implied by the preceding procedures. 595 The three configurable limits T-meas, T-maxsuppress, and T-fail apply 596 to the three timers t-meas, t-maxsuppress, and t-fail respectively. 597 t-meas and t-maxsuppress are reset upon expiry. t-fail is reset by 598 management action or by receipt of a report from the PCN-egress-node 599 concerned. 601 IEA = ingress-egress-aggregate. 603 +---------------+----------+---------------+------------------------+ 604 | Limit | Where | Incidence | Action on Expiry | 605 +---------------+----------+---------------+------------------------+ 606 | T-meas | Egress | One per node | Calculate and possibly | 607 | | node | | report NM-rate, | 608 | | | | ThM-rate*, ETM-rate | 609 | | | | and CLE for each IEA. | 610 | - | - | - | - | 611 | T-maxsuppress | Egress | One per IEA | Send a report for that | 612 | | node | if report | IEA at the next expiry | 613 | | | suppression | of t-meas. | 614 | | | is enabled. | | 615 | - | - | - | - | 616 | T-fail | Decision | One per | Assume failure and | 617 | | Point | egress node | cease to admit flows | 618 | | | | passing through that | 619 | | | | egress node. | 620 +---------------+----------+---------------+------------------------+ 622 * ThM-rate is [CL-specific]. 624 Table 1: Timers Used For the CL Boundary Node Behaviour 626 The value of T-meas SHOULD be configurable, and is RECOMMENDED to be 627 of the order of 100 to 500 ms to provide a reasonable tradeoff 628 between signalling demands on the network and the time taken to react 629 to impending congestion. 631 t-maxsuppress is active only when report suppression is enabled. The 632 value of T-maxsuppress SHOULD be configurable. The appropriate value 633 for T-maxsuppress depends on whether the transport protocol between 634 the PCN-egress-node and the Decision Point is reliable, and whether 635 it implements its own keep-alive procedures. At the time of writing, 636 that transport protocol has not yet been specified. This 637 specification therefore requires that any transport protocol 638 specification for carrying PCN reports MUST specify an appropriate 639 default value for T-maxsuppress. 641 The value of T-fail MUST be configurable. As for T-maxsuppress, the 642 appropriate value of T-fail depends on the transport protocol between 643 the PCN-boundary-nodes and the Decision Point. It is RECOMMENDED 644 that the default value for T-fail be three times the default value 645 for T-maxsuppress as proposed by the transport protocol 646 specification. The transport protocol specification MAY propose a 647 different default value for T-fail in view of the particular 648 characteristics of that protocol. 650 4. Identifying Ingress and Egress Nodes For PCN Traffic 652 The operation of PCN depends on the ability of the PCN-ingress-node 653 to identify the ingress-egress-aggregate to which each new PCN-flow 654 belongs and the ability of the egress node to identify the ingress- 655 egress-aggregate to which each received PCN-packet belongs. If the 656 Decision Point is collocated with the PCN-ingress-node, the PCN- 657 egress-node also needs to associate each ingress-egress-aggregate 658 with the address of the PCN-ingress-node to which it MUST send its 659 reports. 661 The means by which this is done depends on the packet routing 662 technology in use in the network. The procedure to provide the 663 required information is out of the scope of this document. 665 5. Specification of Diffserv Per-Domain Behaviour 667 This section provides the specification required by [RFC3086] for a 668 per-domain behaviour. 670 5.1. Applicability 672 This section draws heavily upon points made in the PCN architecture 673 document, [RFC5559]. 675 The PCN CL boundary node behaviour specified in this document is 676 applicable to inelastic traffic (particularly video and voice) where 677 quality of service for admitted flows is protected primarily by 678 admission control at the ingress to the domain. In exceptional 679 circumstances (e.g., due to network failures) already-admitted flows 680 MAY be terminated to protect the quality of service of the remaining 681 flows. [CL-specific] The CL boundary node behaviour is less likely 682 to terminate too many flows under such circumstances than the SM 683 boundary node behaviour [RFCyyyy]. 685 [RFC EDITOR'S NOTE: please replace RFCyyyy above by the reference to 686 the published version of draft-ietf-pcn-sm-edge-behaviour.] 688 5.2. Technical Specification 690 5.2.1. Classification and Traffic Conditioning 692 This section paraphrases the applicable portions of Sections 3.6 and 693 4.2 of [RFC5559]. 695 Packets at the ingress to the domain are classified as either PCN or 696 non-PCN. Non-PCN packets MAY share the network with PCN packets 697 within the domain. Because the encoding specified in [RFC5696] and 698 used in this document requires the use of the ECN fields, PCN- 699 ingress-nodes MUST block ECN-capable traffic that uses the same DSCP 700 as PCN from entering the PCN-domain directly. "Blocking" means it is 701 dropped or downgraded to a lower-priority behaviour aggregate. 702 Alternatively such traffic MAY be tunnelled through the PCN-domain. 704 PCN packets are further classified as belonging or not belonging to 705 an admitted flow. PCN packets not belonging to an admitted flow are 706 dropped. (This assumes that requests for flow admission are 707 signalled in advance of the arrival of the flows themselves.) 708 Packets belonging to an admitted flow are policed to ensure that they 709 adhere to the agreed rate or flowspec. 711 5.2.2. PHB Configuration 713 The PCN SM and CL boundary node behaviours are metering and marking 714 behaviours rather than scheduling behaviours. As a result, they are 715 not tied to the selection of a specific DSCP value. The PCN working 716 group suggests using admission control for the following service 717 classes (defined in [RFC4594]): 719 o Telephony (EF) 721 o Real-time interactive (CS4) 723 o Broadcast Video (CS3) 725 o Multimedia Conferencing (AF4) 727 For a fuller discussion, see Section A.1 of Appendix A of [RFC5696]. 729 5.3. Attributes 731 The purpose of this per-domain behaviour is to achieve low loss and 732 jitter for the target class of traffic. The design requirement for 733 PCN was that recovery from overloads through the use of flow 734 termination SHOULD happen within 1-3 seconds. PCN probably performs 735 better than that. 737 5.4. Parameters 739 In the list that follows, note that most PCN-ingress-nodes are also 740 PCN-egress-nodes, and vice versa. Furthermore, the PCN-ingress-nodes 741 MAY be collocated with Decision Points. 743 Parameters at the PCN-ingress-node: 744 ----------------------------------- 746 o Filters for distinguishing PCN from non-PCN inbound traffic. 748 o The markings to be applied to PCN-traffic. 750 o Reference rates on each inward link for the [CL-specific] 751 threshold-meter and the excess-traffic-meter; see Section 2. 753 o The information needed to distinguish PCN-traffic belonging to a 754 given ingress-egress-aggregate. 756 Parameters at the PCN-egress-node: 757 ---------------------------------- 759 o The measurement interval T-meas. 761 o Whether report suppression is enabled and, if so, the values of 762 the CLE-reporting-threshold and T-maxsuppress. 764 o [CL-specific] Whether individual flow identifiers will be reported 765 for excess-traffic-marked PCN-traffic. 767 o The information needed to distinguish PCN-traffic belonging to a 768 given ingress-egress-aggregate. 770 o The marking rules for re-marking PCN-traffic leaving the PCN 771 domain. 773 Parameters at each interior node: 774 --------------------------------- 776 o Reference rates on each link for the [CL-specific] threshold-meter 777 and the excess-traffic-meter; see Section 2. 779 o The markings to be applied to PCN-traffic, including the 780 identification of PCN-packets and the encodings to indicate [CL- 781 specific] threshold-marking and excess-traffic-marking. 783 Parameters at the Decision Point: 784 --------------------------------- 786 o Activation/deactivation of PCN-based flow admission. 788 o Activation/deactivation of PCN-based flow termination. 790 o The value of CLE-limit. 792 o The maximum interval T-fail between reports from a given PCN- 793 egress-node, for detecting failure of communications with that 794 node. 796 o The information needed to map between each ingress-egress- 797 aggregate and the corresponding PCN-ingress-node and PCN-egress- 798 node. 800 5.5. Assumptions 802 Assumed that a specific portion of link capacity has been reserved 803 for PCN-traffic. Assumed that the Decision Point receives requests 804 for admission of PCN-flows before the packets in the PCN-flows 805 arrive. This is not a critical assumption, but in its absence, 806 packets will be dropped by the PCN-ingress-node until it obtains the 807 admission decision from the Decision Point. 809 5.6. Example Uses 811 The PCN CL behaviour MAY be used to carry real-time traffic, 812 particularly voice and video. 814 5.7. Environmental Concerns 816 The PCN CL per-domain behaviour can interfere with the use of end-to- 817 end ECN due to reuse of ECN bits for PCN marking. See Appendix B of 818 [RFC5696] for details. 820 5.8. Security Considerations 822 Please see the security considerations in Section 6 as well as those 823 in [RFC2474] and [RFC2475]. 825 6. Security Considerations 827 [RFC5559] provides a general description of the security 828 considerations for PCN. This memo introduces no new considerations. 830 7. IANA Considerations 832 This memo includes no request to IANA. 834 8. Acknowledgements 836 The content of this memo bears a family resemblance to 837 [ID.briscoe-CL]. The authors of that document were Bob Briscoe, 838 Philip Eardley, and Dave Songhurst of BT, Anna Charny and Francois Le 839 Faucheur of Cisco, Jozef Babiarz, Kwok Ho Chan, and Stephen Dudley of 840 Nortel, Giorgios Karagiannis of U. Twente and Ericsson, and Attila 841 Bader and Lars Westberg of Ericsson. 843 Ruediger Geib, Philip Eardley, and Bob Briscoe have helped to shape 844 the present document with their comments. Toby Moncaster gave a 845 careful review to get it into shape for Working Group Last Call. 847 Amongst the authors, Michael Menth deserves special mention for his 848 constant and careful attention to both the technical content of this 849 document and the manner in which it was expressed. 851 9. References 853 9.1. Normative References 855 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 856 Requirement Levels", BCP 14, RFC 2119, March 1997. 858 [RFC2474] Nichols, K., Blake, S., Baker, F., and D. Black, 859 "Definition of the Differentiated Services Field (DS 860 Field) in the IPv4 and IPv6 Headers", RFC 2474, 861 December 1998. 863 [RFC2475] Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z., 864 and W. Weiss, "An Architecture for Differentiated 865 Services", RFC 2475, December 1998. 867 [RFC3086] Nichols, K. and B. Carpenter, "Definition of 868 Differentiated Services Per Domain Behaviors and Rules for 869 their Specification", RFC 3086, April 2001. 871 [RFC5559] Eardley, P., "Pre-Congestion Notification (PCN) 872 Architecture", RFC 5559, June 2009. 874 [RFC5670] Eardley, P., "Metering and Marking Behaviour of PCN- 875 Nodes", RFC 5670, November 2009. 877 [RFC5696] Moncaster, T., Briscoe, B., and M. Menth, "Baseline 878 Encoding and Transport of Pre-Congestion Information", 879 RFC 5696, November 2009. 881 9.2. Informative References 883 [ID.briscoe-CL] 884 Briscoe, B., "An edge-to-edge Deployment Model for Pre- 885 Congestion Notification: Admission Control over a DiffServ 886 Region (expired Internet Draft)", 2006. 888 [IEEE-Satoh] 889 Satoh, D. and H. Ueno, ""Cause and Countermeasure of 890 Overtermination for PCN-Based Flow Termination", 891 Proceedings of IEEE Symposium on Computers and 892 Communications (ISCC '10), pp. 155-161, Riccione, Italy", 893 June 2010. 895 [MeLe10] Menth, M. and F. Lehrieder, "PCN-Based Measured Rate 896 Termination", Computer Networks Journal (Elsevier) vol. 897 54, no. 13, pages 2099 - 2116, September 2010. 899 [RFC4594] Babiarz, J., Chan, K., and F. Baker, "Configuration 900 Guidelines for DiffServ Service Classes", RFC 4594, 901 August 2006. 903 [RFCyyyy] Charny, A., Zhang, J., Karagiannis, G., Menth, M., and T. 904 Taylor, "PCN Boundary Node Behaviour for the Single 905 Marking (SM) Mode of Operation (Work in progress)", 906 December 2010. 908 Authors' Addresses 910 Anna Charny 911 Cisco Systems 912 300 Apollo Drive 913 Chelmsford, MA 01824 914 USA 916 Email: acharny@cisco.com 918 Fortune Huang 919 Huawei Technologies 920 Section F, Huawei Industrial Base, 921 Bantian Longgang, Shenzhen 518129 922 P.R. China 924 Phone: +86 15013838060 925 Email: fqhuang@huawei.com 927 Georgios Karagiannis 928 U. Twente 930 Phone: 931 Email: karagian@cs.utwente.nl 933 Michael Menth 934 University of Tuebingen 935 Sand 13 936 Tuebingen D-97074 937 Germany 939 Phone: +49-7071-2970505 940 Email: menth@informatik.uni-tuebingen.de 941 Tom Taylor (editor) 942 Huawei Technologies 943 1852 Lorraine Ave 944 Ottawa, Ontario K1H 6Z8 945 Canada 947 Phone: +1 613 680 2675 948 Email: tom111.taylor@bell.net