idnits 2.17.1 draft-bryant-mpls-rfc6374-sfl-03.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 (October 27, 2016) is 2735 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-09) exists of draft-bryant-mpls-sfl-control-00 == Outdated reference: A later version (-05) exists of draft-bryant-mpls-sfl-framework-02 == Outdated reference: A later version (-14) exists of draft-ietf-ippm-alt-mark-01 == Outdated reference: A later version (-07) exists of draft-ietf-mpls-flow-ident-02 Summary: 0 errors (**), 0 flaws (~~), 5 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 MPLS Working Group S. Bryant 3 Internet-Draft M. Chen 4 Intended status: Informational Z. Li 5 Expires: April 30, 2017 Huawei 6 G. Swallow 7 S. Sivabalan 8 Cisco Systems 9 G. Mirsky 10 Ericsson 11 G. Fioccola 12 Telecom Italia 13 October 27, 2016 15 RFC6374 Synonymous Flow Labels 16 draft-bryant-mpls-rfc6374-sfl-03 18 Abstract 20 This document describes a method of making RFC6374 performance 21 measurements on flows carried over an MPLS Label Switched path. This 22 allows loss and delay measurements to be made on multi-point to point 23 LSPs and allows the measurement of flows within an MPLS construct 24 using RFC6374. 26 Status of This Memo 28 This Internet-Draft is submitted in full conformance with the 29 provisions of BCP 78 and BCP 79. 31 Internet-Drafts are working documents of the Internet Engineering 32 Task Force (IETF). Note that other groups may also distribute 33 working documents as Internet-Drafts. The list of current Internet- 34 Drafts is at http://datatracker.ietf.org/drafts/current/. 36 Internet-Drafts are draft documents valid for a maximum of six months 37 and may be updated, replaced, or obsoleted by other documents at any 38 time. It is inappropriate to use Internet-Drafts as reference 39 material or to cite them other than as "work in progress." 41 This Internet-Draft will expire on April 30, 2017. 43 Copyright Notice 45 Copyright (c) 2016 IETF Trust and the persons identified as the 46 document authors. All rights reserved. 48 This document is subject to BCP 78 and the IETF Trust's Legal 49 Provisions Relating to IETF Documents 50 (http://trustee.ietf.org/license-info) in effect on the date of 51 publication of this document. Please review these documents 52 carefully, as they describe your rights and restrictions with respect 53 to this document. Code Components extracted from this document must 54 include Simplified BSD License text as described in Section 4.e of 55 the Trust Legal Provisions and are provided without warranty as 56 described in the Simplified BSD License. 58 Table of Contents 60 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 61 2. Requirements Language . . . . . . . . . . . . . . . . . . . . 3 62 3. RFC6374 Packet Loss Measurement with SFL . . . . . . . . . . 4 63 4. RFC6374 Single Packet Delay Measurement . . . . . . . . . . . 4 64 5. Data Service Packet Delay Measurement . . . . . . . . . . . . 4 65 6. Some Simplifying Rules . . . . . . . . . . . . . . . . . . . 6 66 7. Multiple Packet Delay Characteristics . . . . . . . . . . . . 6 67 7.1. Method 1: Time Buckets . . . . . . . . . . . . . . . . . 7 68 7.2. Method 2 Classic Standard Deviation . . . . . . . . . . . 9 69 7.3. Per Packet Delay Measurement . . . . . . . . . . . . . . 9 70 7.3.1. RFC6374 Multi-Packet Delay Measurement Message Format 10 71 7.4. Average Delay . . . . . . . . . . . . . . . . . . . . . . 11 72 8. Sampled Measurement . . . . . . . . . . . . . . . . . . . . . 13 73 9. Carrying RFC6374 Packets over an LSP using an SFL . . . . . . 13 74 9.1. RFC6374 SFL TLV . . . . . . . . . . . . . . . . . . . . . 15 75 10. RFC6374 Combined Loss-Delay Measurement . . . . . . . . . . . 16 76 11. Privacy Considerations . . . . . . . . . . . . . . . . . . . 16 77 12. Security Considerations . . . . . . . . . . . . . . . . . . . 17 78 13. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 17 79 13.1. Allocation of PW Associated Channel Type . . . . . . . . 17 80 13.2. MPLS Loss/Delay TLV Object . . . . . . . . . . . . . . . 17 81 14. References . . . . . . . . . . . . . . . . . . . . . . . . . 17 82 14.1. Normative References . . . . . . . . . . . . . . . . . . 17 83 14.2. Informative References . . . . . . . . . . . . . . . . . 18 84 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 19 86 1. Introduction 88 [I-D.ietf-mpls-flow-ident] describes the requirement for introducing 89 flow identities when using RFC6374 [RFC6374] packet Loss Measurements 90 (LM). In summary RFC6374 uses the loss-measurement (LM) packet as 91 the packet accounting demarcation point. Unfortunately this gives 92 rise to a number of problems that may lead to significant packet 93 accounting errors in certain situations. For example: 95 1. Where a flow is subjected to Equal Cost Multi-Path (ECMP) 96 treatment packets can arrive out of order with respect to the LM 97 packet. 99 2. Where a flow is subjected to ECMP treatment, packets can arrive 100 at different hardware interfaces, thus requiring reception of an 101 LM packet on one interface to trigger a packet accounting action 102 on a different interface which may not be co-located with it. 103 This is a difficult technical problem to address with the 104 required degree of accuracy. 106 3. Even where there is no ECMP (for example on RSVP-TE, MPLS-TP LSPs 107 and PWs) local processing may be distributed over a number of 108 processor cores, leading to synchronization problems. 110 4. Link aggregation techniques may also lead to synchronization 111 issues. 113 5. Some forwarder implementations have a long pipeline between 114 processing a packet and incrementing the associated counter again 115 leading to synchronization difficulties. 117 An approach to mitigating these synchronization issue is described in 118 [I-D.tempia-ippm-p3m] and 119 [I-D.chen-ippm-coloring-based-ipfpm-framework] in which packets are 120 batched by the sender and each batch is marked in some way such that 121 adjacent batches can be easily recognized by the receiver. 123 An additional problem arises where the LSP is a multi-point to point 124 LSP, since MPLS does not include a source address in the packet. 125 Network management operations require the measurement of packet loss 126 between a source and destination. It is thus necessary to introduce 127 some source specific information into the packet to identify packet 128 batches from a specific source. 130 [I-D.bryant-mpls-sfl-framework] describes a method of encoding per 131 flow instructions in an MPLS label stack using a technique called 132 Synonymous Flow Labels (SFL) in which labels which mimic the 133 behaviour of other labels provide the packet batch identifiers and 134 enable the per batch packet accounting. This memo specifies how SFLs 135 are used to perform RFC6374 packet loss and RFC6374 delay 136 measurements. 138 2. Requirements Language 140 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 141 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 142 "OPTIONAL" in this document are to be interpreted as described in 143 [RFC2119]. 145 3. RFC6374 Packet Loss Measurement with SFL 147 The data service packets of the flow being instrumented are grouped 148 into batches, and all the packets within a batch are marked with the 149 SFL [I-D.ietf-mpls-flow-ident] corresponding to that batch. The 150 sender counts the number of packets in the batch. When the batch has 151 completed and the sender is confident that all of the packets in that 152 batch will have been received, the sender issues an RFC6374 Query 153 message to determine the number actually received and hence the 154 number of packets lost. The RFC6374 Query message is sent using the 155 same SFL as the co-responding batch of data service packets. The 156 format of the Query and Response packet is described in Section 9. 158 4. RFC6374 Single Packet Delay Measurement 160 RFC6374 describes how to measure the packet delay by measuring the 161 transit time of an RFC6374 packet over an LSP. The format of an 162 RFC6374 packet carried over the LSP using an SFL is shown in 163 Section 9. 165 Such a packet may not need to be carried over an SFL since the delay 166 over a particular LSP should be a function of the TC bits. However 167 if label inferred scheduling is used [RFC3270] then the SFL would be 168 REQUIRED to ensure that the RFC6374 packet which was being used as a 169 proxy for a data service packet experienced a representative delay. 171 5. Data Service Packet Delay Measurement 173 Where it is desired to more thoroughly instrument a packet flow and 174 to determine the delay of a number of packets it is undesirable to 175 send a large number of RFC6374 packets acting as proxy data service 176 packets Section 4. A method of directly measuring the delay 177 characteristics of a batch of packets is therefore needed. 179 Given the long intervals over which it is necessary to measure packet 180 loss, it is not necessarily the case that the batch times for the two 181 measurement types would be identical. This it is proposed that the 182 two measurements are relatively independent. The notion that they 183 are relatively independent arises for the potential for the two 184 batches to overlap in time, in which case either the delay batch time 185 will need to be cut short or the loss time will need to be extended 186 to allow correct reconciliation of the various counters. 188 The problem is illustrated in Figure 1 below: 190 (1) AAAAAAAAAABBBBBBBBBBAAAAAAAAAABBBBBBBBBB 192 SFL Marking of a packet batch for loss measurement 194 (2) AADDDDAAAABBBBBBBBBBAAAAAAAAAABBBBBBBBBB 196 SFL Marking of a subset if the packets for delay 198 (3) AAAAAAAADDDDBBBBBBBBAAAAAAAAAABBBBBBBBBB 200 SFL Marking of a subset of the packets across a 201 packet loss measurement boundary 203 (4) AACDCDCDAABBBBBBBBBBAAAAAAAAAABBBBBBBBBB 205 The case of multiple delay measurements within 206 a packet loss measurement 208 Figure 1: RFC6734 Query Packet with SFL 210 In case 1 of Figure 1 we show the case were loss measurement alone is 211 being carried out on the flow under analysis. For illustrative 212 purposes consider that in the time interval being analyzed, 10 213 packets always flow. 215 Now consider case 2 of Figure 1 where a small batch of packets need 216 to analyzed for delay. These are marked with a different SFL type 217 indicating that they are to be monitored for both loss and delay. 218 The SFL=A indicates loss batch A, SFL=D indicates a batch of packets 219 that are to be instrumented for delay, but SFL D is synonymous with 220 SFL A, which in turn is synonymous with the underlying FEC. Thus a 221 packet marked D will be accumulated into the A loss batch, into the 222 delay statistics and will be forwarded as normal. Whether the packet 223 is actually counted twice (for loss and delay) or whether the two 224 counters are reconciled during reporting is a local matter. 226 Now consider case 3 of Figure 1 where a small batch of packets are 227 marked for delay across a loss batch boundary. These packets need to 228 considered as part of batch A or a part of batch B, and any RFC6374 229 Query needs to take place after all the packets A or D (which ever 230 option is chosen) have arrived at the receiving LSR. 232 Now consider case 4 of Figure 1. Here we have a case where it is 233 required to take a number of delay measurements within a batch of 234 packets that we are measuring for loss. To do this we need two SFLs 235 for delay (C and D) and alternate between them (on a delay batch by 236 delay batch basis) for the purposes of measuring the delay 237 characteristics of the different batches of packets. 239 6. Some Simplifying Rules 241 It is possible to construct a large set of overlapping measurement 242 type, in terms of loss, delay, loss and delay and batch overlap. If 243 we allow all combination of cases, this leads to configuration, 244 testing and implementation complexity and hence increased operation 245 and capital cost. The following simplifying rules represent the 246 default case: 248 1. Any system that needs to measure delay MUST be able to measure 249 loss. 251 2. Any system that is to measure delay MUST be configured to measure 252 loss. Whether the loss statistics are collected or not is a 253 local matter. 255 3. A delay measurement MAY start at any point during a loss 256 measurement batch, subject to rule 4. 258 4. A delay measurement interval MUST be short enough that it will 259 complete before the enclosing loss batch completes. 261 5. The duration of a second delay (D in Figure 1 batch must be such 262 that all packets from the packets belonging to a first delay 263 batch (C in Figure 1)will have been received before the second 264 delay batch completes. 266 Given that the sender controls both the start and duration of a loss 267 and a delay packet batch, these rules are readily implemented in the 268 control plane. 270 7. Multiple Packet Delay Characteristics 272 A number of methods are described. The expectation is that the MPLS 273 WG possibly with the assistance of the IPPM WG will select one or 274 maybe more than one of these methods for standardization. 276 Three Methods are discussed: 278 1. Time Buckets 280 2. Classic Standard Deviation 282 3. Average Delay 284 7.1. Method 1: Time Buckets 286 In this method the receiving LSR measures the inter-packet gap, 287 classifies the delay into a number of delay buckets and records the 288 number of packets in each bucket. As an example, if the operator 289 were concerned about packets with a delay of up to 1us, 2us, 4us, 290 8us, and over 8us then there would be five buckets and packets that 291 arrived up to 1us would cause the 1us bucket counter to increase, 292 between 1us and 2us the 2us bucket counter would increase etc. In 293 practice it might be better in terms of processing and potential 294 parallelism if, when a packet had a delay relative to its predecessor 295 of 2us both the up to 1us and the 2us counter were incremented and 296 any more detailed information was calculated in the analytics system. 298 This method allows the operator to see more structure in the jitter 299 characteristics than simply measuring the average jitter, and avoids 300 the complication of needing to perform a per packet multiply, but 301 will probably need to time intervals between buckets to be 302 programmable by the operator. 304 The packet format of an RFC6374 Bucket Jitter Measurement Message is 305 shown below: 307 0 1 2 3 308 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 309 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 310 |Version| Flags | Control Code | Message Length | 311 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 312 | QTF | RTF | RPTF | Reserved | 313 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 314 | Session Identifier | DS | 315 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 316 | Number of | Reserved | 317 | Buckets | | 318 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 319 | Interval in 10ns units | 320 | | 321 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 322 | Number pkts in Bucket | 323 | | 324 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 325 ~ ~ 326 ~ ~ 327 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 328 ~ ~ 329 ~ TLV Block ~ 330 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 332 Figure 2: Bucket Jitter Measurement Message Format 334 The Version, Flags, Control Code, Message Length, QTF, RTF, RPTF, 335 Session Identifier, and DS Fields are as defined in section 3.7 of 336 RFC6374. The remaining fields are as follows: 338 o Number of Buckets in the measurement 340 o Reserved must be sent as zero and ignored on receipt 342 o Interval in 10ns units is the inter-packet interval for 343 this bucket 345 o Number Pkts in Bucket is the number of packets found in 346 this bucket. 348 There will be a number of Interval/Number pairs depending on the 349 number of buckets being specified by the Querier. If an RFC6374 350 message is being used to configure the buckets, the Responder MUST 351 respond with 0 packets in each bucket until it has been configured 352 for a full measurement period (i.e. it was configured at the time of 353 the last response message). Out of band configuration is permitted 354 by this mode of operation. 356 Note this is a departure from the normal fixed format used in 357 RFC6374. We need to establish if this is problematic or not. 359 This RFC6374 message is carried over an LSP in the way described in 360 [RFC6374] and over an LSP with an SFL as described in Section 9. 362 7.2. Method 2 Classic Standard Deviation 364 In this method, provision is made for reporting the following delay 365 characteristics: 367 1. Number of packets in the batch (n). 369 2. Sum of delays in a batch (S) 371 3. Maximum Delay. 373 4. Minimum Delay. 375 5. Sum of squares of Inter-packet delay (SS). 377 Characteristic's 1 and 2 give the mean delay. Measuring the delay of 378 each pair in the batch is discussed in Section 7.3. 380 Characteristics 3 and 4 give the outliers. 382 Characteristics 1, 2 and 5 can be used to calculate the variance of 383 the inter-packet gap and hence the standard deviation giving a view 384 of the distribution of packet delays and hence the jitter. The 385 equation for the variance (var) is given by: 387 var = (SS - S*S/n)/(n-1) 389 There is some concern over the use of this algorithm for measuring 390 variance, because SS and S*S/n can be similar numbers, particularly 391 where variance is low. However the method commends it self by not 392 requiring a division in the hardware. A future version of this 393 document will look at using improved statistical methods such as the 394 assumed mean. 396 7.3. Per Packet Delay Measurement 398 If detailed packet delay measurement is required then it might be 399 possible to record the inter-packet gap for each packet pair. In 400 other that exception cases of slow flows or small batch sizes, this 401 would create a large demand on storage in the instrumentation system, 402 bandwidth to such a storage system and bandwidth to the analytics 403 system. Such a measurement technique is outside the scope of this 404 document. 406 7.3.1. RFC6374 Multi-Packet Delay Measurement Message Format 408 The packet format of an RFC6374 Multi-Packet Delay Measurement 409 Message is shown below: 411 0 1 2 3 412 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 413 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 414 |Version| Flags | Control Code | Message Length | 415 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 416 | QTF | RTF | RPTF | Reserved | 417 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 418 | Session Identifier | DS | 419 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 420 | Number of Packets | 421 | | 422 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 423 | Sum of Delays for Batch | 424 | | 425 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 426 | Minimum Delay | 427 | | 428 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 429 | Maximum Delay | 430 | | 431 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 432 | Sum of squares of Inter-packet delay | 433 | | 434 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 435 ~ ~ 436 ~ TLV Block ~ 437 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 439 Figure 3: Multi-packet Delay Measurement Message Format 441 The Version, Flags, Control Code, Message Length, QTF, RTF, RPTF, 442 Session Identifier, and DS Fields are as defined in section 3.7 of 443 RFC6374. The remaining fields are as follows: 445 o Number of Packets is the number of packets in this batch 447 o Sum of Delays for Batch is the duration of the batch in the 448 time measurement format specified in the RTF field. 450 o Minimum Delay is the minimum inter-packet gap observed during 451 the batch in the time format specified in the RTF field. 453 o Maximum Delay is the maximum inter-packet gap observed during 454 the batch in the time format specified in the RTF field. 456 This RFC6374 message is carried over an LSP in the way described in 457 [RFC6374] and over an LSP with an SFL as described in Section 9. 459 7.4. Average Delay 461 Introduced in [I-D.ietf-ippm-alt-mark] is the concept of a one way 462 delay measurement in which the average time of arrival of a set of 463 packets is measured. In this approach the packet is time-stamped at 464 arrival and the Responder returns the sum of the time-stamps and the 465 number of times-tamps. From this the analytics engine can determine 466 the mean delay. An alternative model is that the Responder returns 467 the time stamp of the first and last packet and the number of 468 packets. This method has the advantage of allowing the average delay 469 to be determined at a number of points along the packet path and 470 allowing the components of the delay to be characterized. 472 The packet format of an RFC6374 Average Delay Measurement Message is 473 shown below: 475 0 1 2 3 476 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 477 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 478 |Version| Flags | Control Code | Message Length | 479 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 480 | QTF | RTF | RPTF | Reserved | 481 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 482 | Session Identifier | DS | 483 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 484 | Number of Packets | 485 | | 486 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 487 | Time of First Packet | 488 | | 489 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 490 | Time of Last Packet | 491 | | 492 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 493 | Sum of Timestamps of Batch | 494 | | 495 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 497 ~ ~ 498 ~ TLV Block ~ 499 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 501 Figure 4: Average Delay Measurement Message Format 503 The Version, Flags, Control Code, Message Length, QTF, RTF, RPTF, 504 Session Identifier, and DS Fields are as defined in section 3.7 of 505 RFC6374. The remaining fields are as follows: 507 o Number of Packets is the number of packets in this batch. 509 o Time of First Packet is the time of arrival of the first 510 packet in the batch. 512 o Time of Last Packet is the time of arrival of the last 513 packet in the batch. 515 o Sum of Timestamps of Batch. 517 This RFC6374 message is carried over an LSP in the way described in 518 [RFC6374] and over an LSP with an SFL as described in Section 9 520 8. Sampled Measurement 522 In the discussion so far it has been assumed that we would measure 523 the delay characteristics of every packet in a delay measurement 524 interval defined by an SL of constant colour. In 525 [I-D.ietf-ippm-alt-mark] the concept of a sampled measurement is 526 considered. That is the Responder only measures a packet at the 527 start of a group of packets being marked for delay measurement by a 528 particular colour, rather than every packet in the marked batch. A 529 measurement interval is not defined by the duration of a marked batch 530 of packets but the interval between a pair of RFC6374 packets taking 531 a readout of the delay characteristic. This approach has the 532 advantage that the measurement is not impacted by ECMP effects. 534 9. Carrying RFC6374 Packets over an LSP using an SFL 536 The packet format of an RFC6374 Query message using SFLs is shown in 537 Figure 5. 539 +-------------------------------+ 540 | | 541 | LSP | 542 | Label | 543 +-------------------------------+ 544 | | 545 | Synonymous Flow | 546 | Label | 547 +-------------------------------+ 548 | | 549 | GAL | 550 | | 551 +-------------------------------+ 552 | | 553 | ACH Type = 0xA | 554 | | 555 +-------------------------------+ 556 | | 557 | RFC6374 Measurement Message | 558 | | 559 | +-------------------------+ | 560 | | | | 561 | | RFC6374 Fixed | | 562 | | Header | | 563 | | | | 564 | +-------------------------+ | 565 | | | | 566 | | Optional SFL TLV | | 567 | | | | 568 | +-------------------------+ | 569 | | | | 570 | | Optional Return | | 571 | | Information | | 572 | | | | 573 | +-------------------------+ | 574 | | 575 +-------------------------------+ 577 Figure 5: RFC6734 Query Packet with SFL 579 The MPLS label stack is exactly the same as that used for the user 580 data service packets being instrumented except for the inclusion of 581 the GAL [RFC5586] to allow the receiver to distinguish between normal 582 data packets and OAM packets. Since the packet loss measurements are 583 being made on the data service packets, an RFC6374 direct loss 584 measurement is being made, and which is indicated by the type field 585 in the ACH (Type = 0x000A). 587 The RFC6374 measurement message consists of the three components, the 588 RFC6374 fixed header as specified in [RFC6374] carried over the ACH 589 channel type specified the type of measurement being made (currently: 590 loss, delay or loss and delay) as specified in RFC6374. 592 Two optional TLVs MAY also be carried if needed. The first is the 593 SFL TLV specified in Section 9.1. This is used to provide the 594 implementation with a reminder of the SFL that was used to carry the 595 RFC6374 message. This is needed because a number of MPLS 596 implementations do not provide the MPLS label stack to the MPLS OAM 597 handler. This TLV is required if RFC6374 messages are sent over UDP 598 [RFC7876]. This TLV MUST be included unless, by some method outside 599 the scope of this document, it is known that this information is not 600 needed by the RFC6374 Responder. 602 The second set of information that may be needed is the return 603 information that allows the responder send the RFC6374 response to 604 the Querier. This is not needed if the response is requested in-band 605 and the MPLS construct being measured is a point to point LSP, but 606 otherwise MUST be carried. The return address TLV is defined in 607 RFC6378 and the optional UDP Return Object is defined in [RFC7876]. 609 9.1. RFC6374 SFL TLV 611 Editor's Note we need to review the following in the light of further 612 thoughts on the associated signaling protocol(s). I am fairly 613 confident that we need all the fields other than SFL Batch and SFL 614 Index. The Index is useful in order to map between the label and 615 information associated with the FEC. The batch is part of the 616 lifetime management process. 618 The required RFC6374 SFL TLV is shown in Figure 6. This contains the 619 SFL that was carried in the label stack, the FEC that was used to 620 allocate the SFL and the index into the batch of SLs that were 621 allocated for the FEC that corresponds to this SFL. 623 0 1 2 3 624 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 625 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 626 | Type | Length |MBZ| SFL Batch | SFL Index | 627 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 628 | SFL | Reserved | 629 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 630 | FEC | 631 . . 632 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 634 Figure 6: SFL TLV 636 Where: 638 Type Type is set to Synonymous Flow Label (SFL-TLV). 640 Length The length of the TLV as specified in RFC6374. 642 MBZ MUST be sent as zero and ignored on receive. 644 SFL Batch The SFL batch that this SFL was allocated as part 645 of see [I-D.bryant-mpls-sfl-control] 647 SPL Index The index into the list of SFLs that were assigned 648 against the FEC that corresponds to the SFL. 650 SFL The SFL used to deliver this packet. This is an MPLS 651 label which is a component of a label stack entry as 652 defined in Section 2.1 of [RFC3032]. 654 Reserved MUST be sent as zero and ignored on receive. 656 FEC The Forwarding Equivalence Class that was used to 657 request this SFL. This is encoded as per 658 Section 3.4.1 of TBD 660 This information is needed to allow for operation with hardware that 661 discards the MPLS label stack before passing the remainder of the 662 stack to the OAM handler. By providing both the SFL and the FEC plus 663 index into the array of allocated SFLs a number of implementation 664 types are supported. 666 10. RFC6374 Combined Loss-Delay Measurement 668 This mode of operation is not currently supported by this 669 specification. 671 11. Privacy Considerations 673 The inclusion of originating and/or flow information in a packet 674 provides more identity information and hence potentially degrades the 675 privacy of the communication. Whilst the inclusion of the additional 676 granularity does allow greater insight into the flow characteristics 677 it does not specifically identify which node originated the packet 678 other than by inspection of the network at the point of ingress, or 679 inspection of the control protocol packets. This privacy threat may 680 be mitigated by encrypting the control protocol packets, regularly 681 changing the synonymous labels and by concurrently using a number of 682 such labels. 684 12. Security Considerations 686 The issue noted in Section 5 is a security consideration. There are 687 no other new security issues associated with the MPLS dataplane. Any 688 control protocol used to request SFLs will need to ensure the 689 legitimacy of the request. 691 13. IANA Considerations 693 13.1. Allocation of PW Associated Channel Type 695 As per the IANA considerations in [RFC5586], IANA is requested to 696 allocate the following Channel Type in the "PW Associated Channel 697 Type" registry: 699 Value Description TLV Follows Reference 700 ----- ----------------------------- ----------- --------- 701 TBD Description MPLS Multi-Packet No This 702 Delay Measurement 704 13.2. MPLS Loss/Delay TLV Object 706 IANA is request to allocate a new TLV from the 0-127 range on the 707 MPLS Loss/Delay Measurement TLV Object Registry: 709 Type Description Reference 710 ---- --------------------------------- --------- 711 TBD Synonymous Flow Label This 713 A value of 4 is recommended. 715 14. References 717 14.1. Normative References 719 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 720 Requirement Levels", BCP 14, RFC 2119, 721 DOI 10.17487/RFC2119, March 1997, 722 . 724 [RFC3032] Rosen, E., Tappan, D., Fedorkow, G., Rekhter, Y., 725 Farinacci, D., Li, T., and A. Conta, "MPLS Label Stack 726 Encoding", RFC 3032, DOI 10.17487/RFC3032, January 2001, 727 . 729 [RFC5586] Bocci, M., Ed., Vigoureux, M., Ed., and S. Bryant, Ed., 730 "MPLS Generic Associated Channel", RFC 5586, 731 DOI 10.17487/RFC5586, June 2009, 732 . 734 [RFC7876] Bryant, S., Sivabalan, S., and S. Soni, "UDP Return Path 735 for Packet Loss and Delay Measurement for MPLS Networks", 736 RFC 7876, DOI 10.17487/RFC7876, July 2016, 737 . 739 14.2. Informative References 741 [I-D.bryant-mpls-sfl-control] 742 Bryant, S., Swallow, G., and S. Sivabalan, "A Control 743 Protocol for Synonymous Flow Labels", draft-bryant-mpls- 744 sfl-control-00 (work in progress), March 2015. 746 [I-D.bryant-mpls-sfl-framework] 747 Bryant, S., Chen, M., Li, Z., Swallow, G., Sivabalan, S., 748 and G. Mirsky, "Synonymous Flow Label Framework", draft- 749 bryant-mpls-sfl-framework-02 (work in progress), October 750 2016. 752 [I-D.chen-ippm-coloring-based-ipfpm-framework] 753 Chen, M., Zheng, L., Mirsky, G., Fioccola, G., and T. 754 Mizrahi, "IP Flow Performance Measurement Framework", 755 draft-chen-ippm-coloring-based-ipfpm-framework-06 (work in 756 progress), March 2016. 758 [I-D.ietf-ippm-alt-mark] 759 Fioccola, G., Capello, A., Cociglio, M., Castaldelli, L., 760 Chen, M., Zheng, L., Mirsky, G., and T. Mizrahi, 761 "Alternate Marking method for passive performance 762 monitoring", draft-ietf-ippm-alt-mark-01 (work in 763 progress), July 2016. 765 [I-D.ietf-mpls-flow-ident] 766 Bryant, S., Chen, M., Li, Z., Pignataro, C., and G. 767 Mirsky, "MPLS Flow Identification Considerations", draft- 768 ietf-mpls-flow-ident-02 (work in progress), August 2016. 770 [I-D.tempia-ippm-p3m] 771 Capello, A., Cociglio, M., Fioccola, G., Castaldelli, L., 772 and A. Bonda, "A packet based method for passive 773 performance monitoring", draft-tempia-ippm-p3m-03 (work in 774 progress), March 2016. 776 [RFC3270] Le Faucheur, F., Wu, L., Davie, B., Davari, S., Vaananen, 777 P., Krishnan, R., Cheval, P., and J. Heinanen, "Multi- 778 Protocol Label Switching (MPLS) Support of Differentiated 779 Services", RFC 3270, DOI 10.17487/RFC3270, May 2002, 780 . 782 [RFC6374] Frost, D. and S. Bryant, "Packet Loss and Delay 783 Measurement for MPLS Networks", RFC 6374, 784 DOI 10.17487/RFC6374, September 2011, 785 . 787 Authors' Addresses 789 Stewart Bryant 790 Huawei 792 Email: stewart.bryant@gmail.com 794 Mach Chen 795 Huawei 797 Email: mach.chen@huawei.com 799 Zhenbin Li 800 Huawei 802 Email: lizhenbin@huawei.com 804 George Swallow 805 Cisco Systems 807 Email: swallow.ietf@gmail.com 809 Siva Sivabalan 810 Cisco Systems 812 Email: msiva@cisco.com 814 Gregory Mirsky 815 Ericsson 817 Email: gregory.mirsky@eicsson.com 818 Giuseppe Fioccola 819 Telecom Italia 821 Email: giuseppe.fioccola@telecomitalia.it