idnits 2.17.1 draft-ietf-6man-ipv6-alt-mark-07.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 (June 22, 2021) is 1032 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-06) exists of draft-chen-pce-pcep-ifit-02 == Outdated reference: A later version (-08) exists of draft-fz-spring-srv6-alt-mark-00 == Outdated reference: A later version (-01) exists of draft-hinden-6man-hbh-processing-00 == Outdated reference: A later version (-08) exists of draft-ietf-idr-sr-policy-ifit-01 == Outdated reference: A later version (-06) exists of draft-peng-v6ops-hbh-03 -- Obsolete informational reference (is this intentional?): RFC 8321 (Obsoleted by RFC 9341) -- Obsolete informational reference (is this intentional?): RFC 8889 (Obsoleted by RFC 9342) Summary: 0 errors (**), 0 flaws (~~), 6 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 6MAN Working Group G. Fioccola 3 Internet-Draft T. Zhou 4 Intended status: Standards Track Huawei 5 Expires: December 24, 2021 M. Cociglio 6 Telecom Italia 7 F. Qin 8 China Mobile 9 R. Pang 10 China Unicom 11 June 22, 2021 13 IPv6 Application of the Alternate Marking Method 14 draft-ietf-6man-ipv6-alt-mark-07 16 Abstract 18 This document describes how the Alternate Marking Method can be used 19 as a passive performance measurement tool in an IPv6 domain. It 20 defines a new Extension Header Option to encode Alternate Marking 21 information in both the Hop-by-Hop Options Header and Destination 22 Options Header. 24 Status of This Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at https://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on December 24, 2021. 41 Copyright Notice 43 Copyright (c) 2021 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (https://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 59 1.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 3 60 1.2. Requirements Language . . . . . . . . . . . . . . . . . . 3 61 2. Alternate Marking application to IPv6 . . . . . . . . . . . . 3 62 2.1. Controlled Domain . . . . . . . . . . . . . . . . . . . . 5 63 3. Definition of the AltMark Option . . . . . . . . . . . . . . 6 64 3.1. Data Fields Format . . . . . . . . . . . . . . . . . . . 6 65 4. Use of the AltMark Option . . . . . . . . . . . . . . . . . . 7 66 5. Alternate Marking Method Operation . . . . . . . . . . . . . 9 67 5.1. Packet Loss Measurement . . . . . . . . . . . . . . . . . 9 68 5.2. Packet Delay Measurement . . . . . . . . . . . . . . . . 11 69 5.3. Flow Monitoring Identification . . . . . . . . . . . . . 12 70 5.3.1. Uniqueness of FlowMonID . . . . . . . . . . . . . . . 13 71 5.4. Multipoint and Clustered Alternate Marking . . . . . . . 14 72 5.5. Data Collection and Calculation . . . . . . . . . . . . . 14 73 6. Security Considerations . . . . . . . . . . . . . . . . . . . 14 74 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 17 75 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 18 76 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 18 77 9.1. Normative References . . . . . . . . . . . . . . . . . . 18 78 9.2. Informative References . . . . . . . . . . . . . . . . . 18 79 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 20 81 1. Introduction 83 [RFC8321] and [RFC8889] describe a passive performance measurement 84 method, which can be used to measure packet loss, latency and jitter 85 on live traffic. Since this method is based on marking consecutive 86 batches of packets, the method is often referred to as the Alternate 87 Marking Method. 89 This document defines how the Alternate Marking Method can be used to 90 measure performance metrics in IPv6. The rationale is to apply the 91 Alternate Marking methodology to IPv6 and therefore allow detailed 92 packet loss, delay and delay variation measurements both hop-by-hop 93 and end-to-end to exactly locate the issues in an IPv6 network. 95 The Alternate Marking is an on-path telemetry technique and consists 96 in synchronizing the measurements in different points of a network by 97 switching the value of a marking bit and therefore divide the packet 98 flow into batches. Each batch represents a measurable entity 99 unambiguously recognizable by all network nodes along the path. By 100 counting the number of packets in each batch and comparing the values 101 measured by different nodes, it is possible to precisely measure the 102 packet loss. In a similar way the alternation of the values of the 103 marking bits can be used as a time reference to calculate the delay 104 and delay variation. The Alternate Marking operation is further 105 described in Section 5. 107 The format of IPv6 addresses is defined in [RFC4291] while [RFC8200] 108 defines the IPv6 Header, including a 20-bit Flow Label and the IPv6 109 Extension Headers. 111 [I-D.fioccola-v6ops-ipv6-alt-mark] summarizes the possible 112 implementation options for the application of the Alternate Marking 113 Method in an IPv6 domain. This document, starting from the outcome 114 of [I-D.fioccola-v6ops-ipv6-alt-mark], introduces a new TLV (type- 115 length-value) that can be encoded in the Options Headers (Hop-by-Hop 116 or Destination) for the purpose of the Alternate Marking Method 117 application in an IPv6 domain. While the case of Segment Routing 118 Header (SRH), defined in [RFC8754], is also discussed, it is valid 119 for all the types of Routing Header (RH). 121 The threat model for the application of the Alternate Marking Method 122 in an IPv6 domain is reported in Section 6. As for all the on-path 123 telemetry technique, the only definitive solution is that this 124 methodology MUST be applied in a controlled domain and therefore the 125 application to untrusted domain is NOT RECOMMENDED. 127 1.1. Terminology 129 This document uses the terms related to the Alternate Marking Method 130 as defined in [RFC8321] and [RFC8889]. 132 1.2. Requirements Language 134 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 135 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 136 "OPTIONAL" in this document are to be interpreted as described in BCP 137 14 [RFC2119] [RFC8174] when, and only when, they appear in all 138 capitals, as shown here. 140 2. Alternate Marking application to IPv6 142 The Alternate Marking Method requires a marking field. As mentioned, 143 several alternatives have been analysed in 145 [I-D.fioccola-v6ops-ipv6-alt-mark] such as IPv6 Extension Headers, 146 IPv6 Address and Flow Label. 148 [I-D.fioccola-v6ops-ipv6-alt-mark] analyzed and discussed all the 149 available possibilities and the drawbacks: 151 Reusing existing Extension Header for Alternate Marking leads to a 152 non-optimized implementation; 154 Using the IPv6 destination address to encode the Alternate Marking 155 processing is very expensive; 157 Using the IPv6 Flow Label for Alternate Marking conflicts with the 158 utilization of the Flow Label for load distribution purpose 159 ([RFC6438]). 161 In the end, [I-D.fioccola-v6ops-ipv6-alt-mark] demonstrated that a 162 new Hop-by-Hop or a new Destination Option was the best approach. 164 The approach for the Alternate Marking application to IPv6 specified 165 in this memo is compliant with [RFC8200]. It involves the following 166 operations: 168 o The source node is the only one that writes the Option Header to 169 mark alternately the flow (for both Hop-by-Hop and Destination 170 Option). The intermediate nodes and destination node MUST only 171 read the marking values of the option without modifying the Option 172 Header. 174 o In case of Hop-by-Hop Option Header carrying Alternate Marking 175 bits, it is not inserted or deleted, but can be read by any node 176 along the path. The intermediate nodes may be configured to 177 support this Option or not and the measurement can be done only 178 for the nodes configured to read the Option. As further discussed 179 in Section 4, the presence of the hop-by-hop option should not 180 affect the traffic throughput both on nodes that do not recognize 181 this option and on the nodes that support it. However it is 182 important to mention that there is a difference between the theory 183 and the implementation and it can happen that packets with hop-by- 184 hop option could also be skipped or processed in the slow path. 185 While some proposals are trying to address this problem 186 ([I-D.peng-v6ops-hbh], [I-D.hinden-6man-hbh-processing]), these 187 aspects are out of the scope for this document. 189 o In case of Destination Option Header carrying Alternate Marking 190 bits, it is not processed, inserted, or deleted by any node along 191 the path until the packet reaches the destination node. Note 192 that, if there is also a Routing Header (RH), any visited 193 destination in the route list can process the Option Header. 195 Hop-by-Hop Option Header is also useful to signal to routers on the 196 path to process the Alternate Marking. However, as said, routers 197 will examine this option if properly configured. 199 The optimization of both implementation and scaling of the Alternate 200 Marking Method is also considered and a way to identify flows is 201 required. The Flow Monitoring Identification field (FlowMonID), as 202 introduced in Section 5.3, goes in this direction and it is used to 203 identify a monitored flow. 205 The FlowMonID is different from the Flow Label field of the IPv6 206 Header ([RFC6437]). The Flow Label field in the IPv6 header is used 207 by a source to label sequences of packets to be treated in the 208 network as a single flow and, as reported in [RFC6438], it can be 209 used for load-balancing/equal cost multi-path (LB/ECMP). The reuse 210 of Flow Label field for identifying monitored flows is not considered 211 since it may change the application intent and forwarding behaviour. 212 Furthermore the Flow Label may be changed en route and this may also 213 violate the measurement task. Also, since the Flow Label is pseudo- 214 random, there is always a finite probability of collision. Those 215 reasons make the definition of the FlowMonID necessary for IPv6. 216 Indeed, the FlowMonID is designed and only used to identify the 217 monitored flow. Flow Label and FlowMonID within the same packet are 218 totally disjoint, have different scope, identify different flows, and 219 are intended for different use cases. 221 The rationale for the FlowMonID is further discussed in Section 5.3. 222 This 20 bit field allows easy and flexible identification of the 223 monitored flow and enables a finer granularity and improved 224 measurement correlation. An important point that will be discussed 225 in Section 5.3.1 is the uniqueness of the FlowMonID and how to allow 226 disambiguation of the FlowMonID in case of collision. 228 The following section highlights an important requirement for the 229 application of the Alternate Marking to IPv6. The concept of the 230 controlled domain is explained and it is considered an essential 231 precondition, as also highlighted in Section 6. 233 2.1. Controlled Domain 235 [RFC8799] introduces the concept of specific limited domain solutions 236 and, in this regard, it is reported the IPv6 Application of the 237 Alternate Marking Method as an example. 239 IPv6 has much more flexibility than IPv4 and innovative applications 240 have been proposed, but for a number of reasons, such as the 241 policies, options supported, the style of network management and 242 security requirements, it is suggested to limit some of these 243 applications to a controlled domain. This is also the case of the 244 Alternate Marking application to IPv6 as assumed hereinafter. 246 Therefore, the IPv6 application of the Alternate Marking Method MUST 247 NOT be deployed outside a controlled domain. It is RECOMMENDED that 248 an implementation can be able to reject packets that carry Alternate 249 Marking data and are entering or leaving the controlled domains. The 250 security considerations clarify this requirement and are reported in 251 Section 6. 253 3. Definition of the AltMark Option 255 The definition of a new TLV for the Options Extension Headers, 256 carrying the data fields dedicated to the Alternate Marking method, 257 is reported below. 259 3.1. Data Fields Format 261 The following figure shows the data fields format for enhanced 262 Alternate Marking TLV. This AltMark data can be encapsulated in the 263 IPv6 Options Headers (Hop-by-Hop or Destination Option). 265 0 1 2 3 266 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 1 267 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 268 | Option Type | Opt Data Len | 269 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 270 | FlowMonID |L|D| Reserved | 271 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 273 where: 275 o Option Type: 8 bit identifier of the type of Option that needs to 276 be allocated. Unrecognized Types MUST be ignored on receipt. For 277 Hop-by-Hop Options Header or Destination Options Header, [RFC8200] 278 defines how to encode the three high-order bits of the Option Type 279 field. The two high-order bits specify the action that must be 280 taken if the processing IPv6 node does not recognize the Option 281 Type; for AltMark these two bits MUST be set to 00 (skip over this 282 Option and continue processing the header). The third-highest- 283 order bit specifies whether or not the Option Data can change en 284 route to the packet's final destination; for AltMark the value of 285 this bit MUST be set to 0 (Option Data does not change en route). 286 In this way, since the three high-order bits of the AltMark Option 287 are set to 000, it means that nodes can simply skip this Option if 288 they do not recognize and that the data of this Option do not 289 change en route, indeed the source is the only one that can write 290 it. 292 o Opt Data Len: 4. It is the length of the Option Data Fields of 293 this Option in bytes. 295 o FlowMonID: 20 bits unsigned integer. The FlowMon identifier is 296 described in Section 5.3. As further discussed below, it has been 297 picked 20 bit since it is a reasonable value and a good compromise 298 in relation to the chance of collision if it is set pseudo 299 randomly by the source node or set by a centralized controller. 301 o L: Loss flag for Packet Loss Measurement as described in 302 Section 5.1; 304 o D: Delay flag for Single Packet Delay Measurement as described in 305 Section 5.2; 307 o Reserved: is reserved for future use. These bits MUST be set to 308 zero on transmission and ignored on receipt. 310 4. Use of the AltMark Option 312 The AltMark Option is the best way to implement the Alternate Marking 313 method and it is carried by the Hop-by-Hop Options header and the 314 Destination Options header. In case of Destination Option, it is 315 processed only by the source and destination nodes: the source node 316 inserts and the destination node removes it. While, in case of Hop- 317 by-Hop Option, it may be examined by any node along the path, if 318 explicitly configured to do so. 320 It is important to highlight that the Option Layout can be used both 321 as Destination Option and as Hop-by-Hop Option depending on the Use 322 Cases and it is based on the chosen type of performance measurement. 323 In general, it is needed to perform both end to end and hop by hop 324 measurements, and the Alternate Marking methodology allows, by 325 definition, both performance measurements. But, in many cases the 326 end-to-end measurement is not enough and it is required also the hop- 327 by-hop measurement, so the most complete choice is the Hop-by-Hop 328 Options Header. 330 IPv6, as specified in [RFC8200], allows nodes to optionally process 331 Hop-by-Hop headers. Specifically the Hop-by-Hop Options header is 332 not inserted or deleted, but may be examined or processed by any node 333 along a packet's delivery path, until the packet reaches the node (or 334 each of the set of nodes, in the case of multicast) identified in the 335 Destination Address field of the IPv6 header. Also, it is expected 336 that nodes along a packet's delivery path only examine and process 337 the Hop-by-Hop Options header if explicitly configured to do so. 339 The Hop-by-Hop Option defined in this document is designed to take 340 advantage of the property of how Hop-by-Hop options are processed. 341 Nodes that do not support this Option SHOULD ignore them. This can 342 mean that, in this case, the performance measurement does not account 343 for all links and nodes along a path. 345 Another application that can be mentioned is the presence of a 346 Routing Header, in particular it is possible to consider SRv6. A new 347 type of Routing Header, referred as SRH, has been defined for SRv6. 348 Like any other use case of IPv6, Hop-by-Hop and Destination Options 349 are useable when SRv6 header is present. Because SRv6 is implemented 350 through a Segment Routing Header (SRH), Destination Options before 351 the Routing Header are processed by each destination in the route 352 list, that means, in case of SRH, by every SR node that is identified 353 by the SR path. More details about the SRv6 application are 354 described in [I-D.fz-spring-srv6-alt-mark]. 356 In summary, it is possible to list the alternative possibilities: 358 o Destination Option not preceding a Routing Header => measurement 359 only by node in Destination Address. 361 o Hop-by-Hop Option => every router on the path with feature 362 enabled. 364 o Destination Option preceding a Routing Header => every destination 365 node in the route list. 367 In general, Hop-by-Hop and Destination Options are the most suitable 368 ways to implement Alternate Marking. 370 It is worth mentioning that new Hop-by-Hop Options are not strongly 371 recommended in [RFC7045] and [RFC8200], unless there is a clear 372 justification to standardize it, because nodes may be configured to 373 ignore the Options Header, drop or assign packets containing an 374 Options Header to a slow processing path. In case of the AltMark 375 data fields described in this document, the motivation to standardize 376 a new Hop-by-Hop Option is that it is needed for OAM (Operations, 377 Administration, and Maintenance). An intermediate node can read it 378 or not but this does not affect the packet behavior. The source node 379 is the only one that writes the Hop-by-Hop Option to mark alternately 380 the flow, so, the performance measurement can be done for those nodes 381 configured to read this Option, while the others are simply not 382 considered for the metrics. 384 It is important to highlight that the definition of the Hop-by-Hop 385 Options in this document is designed to minimize throughput impact 386 both on nodes that do not recognize the Option and on node that 387 support it. Indeed, the three high-order bits of the Options Header 388 defined in this draft are 000 and, in theory, as per [RFC8200] and 389 [I-D.hinden-6man-hbh-processing], this means "skip if do not 390 recognize and data do not change en route". [RFC8200] also mentions 391 that the nodes only examine and process the Hop-by-Hop Options header 392 if explicitly configured to do so. For these reasons, this HbH 393 Option should not affect the throughput. However, in practice, it is 394 important to be aware for the implementation that the things may be 395 different and it can happen that packets with Hop-by-Hop are forced 396 onto the slow path, but this is a general issue, as also explained in 397 [I-D.hinden-6man-hbh-processing]. 399 5. Alternate Marking Method Operation 401 This section describes how the method operates. [RFC8321] introduces 402 several alternatives but in this section the most applicable methods 403 are reported and a new field is introduced to facilitate the 404 deployment and improve the scalability. 406 5.1. Packet Loss Measurement 408 The measurement of the packet loss is really straightforward. The 409 packets of the flow are grouped into batches, and all the packets 410 within a batch are marked by setting the L bit (Loss flag) to a same 411 value. The source node can switch the value of the L bit between 0 412 and 1 after a fixed number of packets or according to a fixed timer, 413 and this depends on the implementation. The source node is the only 414 one that marks the packets to create the batches, while the 415 intermediate nodes only read the marking values and identify the 416 packet batches. By counting the number of packets in each batch and 417 comparing the values measured by different network nodes along the 418 path, it is possible to measure the packet loss occurred in any 419 single batch between any two nodes. Each batch represents a 420 measurable entity unambiguously recognizable by all network nodes 421 along the path. 423 Both fixed number of packets and fixed timer can be used by the 424 source node to create packet batches. But, as also explained in 425 [RFC8321], using a fixed timer for the switching offers better 426 control over the method, indeed the length of the batches can be 427 chosen large enough to simplify the collection and the comparison of 428 the measures taken by different network nodes. In the implementation 429 the counters can be sent out by each node to the controller that is 430 responsible for the calculation. It is also possible to exchange 431 this information by using other on-path techniques. But this is out 432 of scope for this document. 434 Packets with different L values may get swapped at batch boundaries, 435 and in this case, it is required that each marked packet can be 436 assigned to the right batch by each router. It is important to 437 mention that for the application of this method there are two 438 elements to consider: the clock error between network nodes and the 439 network delay. These can create offsets between the batches and out- 440 of-order of the packets. The mathematical formula on timing aspects, 441 explained in section 3.2 of [RFC8321], must be satisfied and it takes 442 into considerations the different causes of reordering such as clock 443 error and network delay. The assumption is to define the available 444 counting interval where to get stable counters and to avoid these 445 issues. Specifically, if the effects of network delay are ignored, 446 the condition to implement the methodology is that the clocks in 447 different nodes MUST be synchronized to the same clock reference with 448 an accuracy of +/- B/2 time units, where B is the fixed time duration 449 of the block. In this way each marked packet can be assigned to the 450 right batch by each node. Usually the counters can be taken in the 451 middle of the batch period to be sure to take still counters. In a 452 few words this implies that the length of the batches MUST be chosen 453 large enough so that the method is not affected by those factors. 454 The length of the batches can be determined based on the specific 455 deployment scenario. 457 L bit=1 ----------+ +-----------+ +---------- 458 | | | | 459 L bit=0 +-----------+ +-----------+ 460 Batch n ... Batch 3 Batch 2 Batch 1 461 <---------> <---------> <---------> <---------> <---------> 463 Traffic Flow 464 ===========================================================> 465 L bit ...1111111111 0000000000 11111111111 00000000000 111111111... 466 ===========================================================> 468 Figure 1: Packet Loss Measurement and Single-Marking Methodology 469 using L bit 471 It is worth mentioning that the length of the batches is considered 472 stable over time in the previous figure. In theory, it is possible 473 to change the length of batches over time and and among different 474 flows for more flexibility. But, in practice, it could complicate 475 the correlation of the information. 477 5.2. Packet Delay Measurement 479 The same principle used to measure packet loss can be applied also to 480 one-way delay measurement. Delay metrics MAY be calculated using the 481 two possibilities: 483 1. Single-Marking Methodology: This approach uses only the L bit to 484 calculate both packet loss and delay. In this case, the D flag 485 MUST be set to zero on transmit and ignored by the monitoring 486 points. The alternation of the values of the L bit can be used 487 as a time reference to calculate the delay. Whenever the L bit 488 changes and a new batch starts, a network node can store the 489 timestamp of the first packet of the new batch, that timestamp 490 can be compared with the timestamp of the first packet of the 491 same batch on a second node to compute packet delay. But this 492 measurement is accurate only if no packet loss occurs and if 493 there is no packet reordering at the edges of the batches. A 494 different approach can also be considered and it is based on the 495 concept of the mean delay. The mean delay for each batch is 496 calculated by considering the average arrival time of the packets 497 for the relative batch. There are limitations also in this case 498 indeed, each node needs to collect all the timestamps and 499 calculate the average timestamp for each batch. In addition the 500 information is limited to a mean value. 502 2. Double-Marking Methodology: This approach is more complete and 503 uses the L bit only to calculate packet loss and the D bit (Delay 504 flag) is fully dedicated to delay measurements. The idea is to 505 use the first marking with the L bit to create the alternate flow 506 and, within the batches identified by the L bit, a second marking 507 is used to select the packets for measuring delay. The D bit 508 creates a new set of marked packets that are fully identified 509 over the network, so that a network node can store the timestamps 510 of these packets; these timestamps can be compared with the 511 timestamps of the same packets on a second node to compute packet 512 delay values for each packet. The most efficient and robust mode 513 is to select a single double-marked packet for each batch, in 514 this way there is no time gap to consider between the double- 515 marked packets to avoid their reorder. Regarding the rule for 516 the selection of the packet to be double-marked, the same 517 considerations in Section 5.1 apply also here and the double- 518 marked packet can be chosen within the available counting 519 interval that is not affected by factors such as clock errors. 520 If a double-marked packet is lost, the delay measurement for the 521 considered batch is simply discarded, but this is not a big 522 problem because it is easy to recognize the problematic batch and 523 skip the measurement just for that one. So in order to have more 524 information about the delay and to overcome out-of-order issues 525 this method is preferred. 527 In summary the approach with double marking is better than the 528 approach with single marking. Moreover the two approaches can also 529 be combined to have even more information and statistics on delay. 531 Similar to what said in Section 5.1 for the packet counters, in the 532 implementation the timestamps can be sent out to the controller that 533 is responsible for the calculation or could also be exchanged using 534 other on-path techniques. But this is out of scope for this 535 document. 537 L bit=1 ----------+ +-----------+ +---------- 538 | | | | 539 L bit=0 +-----------+ +-----------+ 541 D bit=1 + + + + + 542 | | | | | 543 D bit=0 ------+----------+----------+----------+------------+----- 545 Traffic Flow 546 ===========================================================> 547 L bit ...1111111111 0000000000 11111111111 00000000000 111111111... 549 D bit ...0000010000 0000010000 00000100000 00001000000 000001000... 550 ===========================================================> 552 Figure 2: Double-Marking Methodology using L bit and D bit 554 Likewise to packet delay measurement (both for Single Marking and 555 Double Marking), the method can also be used to measure the inter- 556 arrival jitter. 558 5.3. Flow Monitoring Identification 560 The Flow Monitoring Identification (FlowMonID) is required for some 561 general reasons: 563 o First, it helps to reduce the per node configuration. Otherwise, 564 each node needs to configure an access-control list (ACL) for each 565 of the monitored flows. Moreover, using a flow identifier allows 566 a flexible granularity for the flow definition. 568 o Second, it simplifies the counters handling. Hardware processing 569 of flow tuples (and ACL matching) is challenging and often incurs 570 into performance issues, especially in tunnel interfaces. 572 o Third, it eases the data export encapsulation and correlation for 573 the collectors. 575 The FlowMon identifier field is to uniquely identify a monitored flow 576 within the measurement domain. The field is set at the source node. 577 The FlowMonID can be set in two ways: 579 * It can be uniformly assigned by the central controller. Since 580 the controller knows the network topology, it can set the value 581 properly to avoid or minimize ambiguity and guarantee the 582 uniqueness. 584 * It can be algorithmically generated by the source node, that can 585 set it pseudo-randomly with some chance of collision. This 586 approach cannot guarantee the uniqueness of FlowMonID but it may 587 be preferred for local or private networks, where the conflict 588 probability is small due to the large FlowMonID space. 590 The value of 20 bits has been selected for the FlowMonID since it is 591 a good compromise and implies a low rate of ambiguous FlowMonIDs that 592 can be considered acceptable in most of the applications. Indeed 593 with 20 bits the number of combinations is 1048576. 595 if the FlowMonID is set by the source node, the intermediate nodes 596 can read the FlowMonIDs from the packets in flight and act 597 accordingly. While, if the FlowMonID is set by the controller, both 598 possibilities are feasible for the intermediate nodes which can learn 599 by reading the packets or can be instructed by the controller. 601 When all values in the FlowMonID space are consumed, the centralized 602 controller can keep track and reassign the values that are not used 603 any more by old flows, while if the FlowMonID is pseudo randomly 604 generated by the source, conflicts and collisions are possible. 606 5.3.1. Uniqueness of FlowMonID 608 It is important to note that if the 20 bit FlowMonID is set 609 independently and pseudo randomly there is a chance of collision. 610 Indeed, by using the well-known birthday problem in probability 611 theory, if the 20 bit FlowMonID is set independently and pseudo 612 randomly without any additional input entropy, there is a 50% chance 613 of collision for 1206 flows. So, for more entropy, FlowMonID can 614 either be combined with other identifying flow information in a 615 packet (e.g. it is possible to consider the hashed 3-tuple Flow 616 Label, Source and Destination addresses) or the FlowMonID size could 617 be increased. 619 This issue is more visible when the FlowMonID is pseudo randomly 620 generated by the source node and there needs to tag it with 621 additional flow information to allow disambiguation. While, in case 622 of a centralized controller, the controller should set FlowMonID by 623 considering these aspects and instruct the nodes properly in order to 624 guarantee its uniqueness. 626 5.4. Multipoint and Clustered Alternate Marking 628 The Alternate Marking method can also be extended to any kind of 629 multipoint to multipoint paths, and the network clustering approach 630 allows a flexible and optimized performance measurement, as described 631 in [RFC8889]. 633 The Cluster is the smallest identifiable subnetwork of the entire 634 Network graph that still satisfies the condition that the number of 635 packets that goes in is the same that goes out. With network 636 clustering, it is possible to use the partition of the network into 637 clusters at different levels in order to perform the needed degree of 638 detail. So, for Multipoint Alternate Marking, FlowMonID can identify 639 in general a multipoint-to-multipoint flow and not only a point-to- 640 point flow. 642 5.5. Data Collection and Calculation 644 The nodes enabled to perform performance monitoring collect the value 645 of the packet counters and timestamps. There are several 646 alternatives to implement Data Collection and Calculation, but this 647 is not specified in this document. 649 There are documents on the control plane mechanisms of Alternate 650 Marking, e.g. [I-D.ietf-idr-sr-policy-ifit], 651 [I-D.chen-pce-pcep-ifit]. 653 6. Security Considerations 655 This document aims to apply a method to perform measurements that 656 does not directly affect Internet security nor applications that run 657 on the Internet. However, implementation of this method must be 658 mindful of security and privacy concerns. 660 There are two types of security concerns: potential harm caused by 661 the measurements and potential harm to the measurements. 663 Harm caused by the measurement: Alternate Marking implies 664 modifications on the fly to an Option Header of IPv6 packets by the 665 source node but this must be performed in a way that does not alter 666 the quality of service experienced by the packets and that preserves 667 stability and performance of routers doing the measurements. As 668 already discussed in Section 4, it is RECOMMENDED that the AltMark 669 Option does not affect the throughput and therefore the user 670 experience. 672 Harm to the measurement: Alternate Marking measurements could be 673 harmed by routers altering the fields of the AltMark Option (e.g. 674 marking of the packets, FlowMonID) or by a malicious attacker adding 675 AltMark Option to the packets in order to consume the resources of 676 network devices and entities involved. As described above, the 677 source node is the only one that writes the Option Header while the 678 intermediate nodes and destination node only read it without 679 modifying the Option Header. But, for example, an on-path attacker 680 can modify the flags, whether intentionally or accidentally, or 681 insert deliberately a new option to the packet flow or delete the 682 option from the packet flow. The consequent effect could be to give 683 the appearance of loss or delay or invalidate the measurement by 684 modifying option identifiers, such as FlowMonID. The malicious 685 implication can be to cause actions from the network administrator 686 where an intervention is not necessary or to hide real issues in the 687 network. Since the measurement itself may be affected by network 688 nodes intentionally altering the bits of the AltMark Option or 689 injecting Options headers as a means for Denial of Service (DoS), the 690 Alternate Marking MUST be applied in the context of a controlled 691 domain, where the network nodes are locally administered and this 692 type of attack can be avoided. 694 The flow identifier (FlowMonID) composes the AltMark Option together 695 with the two marking bits (L and D). As explained in Section 5.3.1, 696 there is a chance of collision if the FlowMonID is set pseudo 697 randomly and a solution exist. In general this may not be a problem 698 and a low rate of ambiguous FlowMonIDs can be acceptable, since this 699 does not cause significant harm to the operators or their clients and 700 this harm may not justify the complications of avoiding it. But, for 701 large scale measurements where it is possible to monitor a big number 702 of flows, the disambiguation of the FlowMonID field is something to 703 take into account. 705 The privacy concerns also need to be analyzed even if the method only 706 relies on information contained in the Option Header without any 707 release of user data. Indeed, from a confidentiality perspective, 708 although AltMark Option does not contain user data, the metadata can 709 be used for network reconnaissance to compromise the privacy of users 710 by allowing attackers to collect information about network 711 performance and network paths. AltMark Option contains two kind of 712 metadata: the marking bits (L and D bits) and the flow identifier 713 (FlowMonID). 715 The marking bits are the small information that is exchanged 716 between the network nodes. Therefore, due to this intrinsic 717 characteristic, network reconnaissance through passive 718 eavesdropping on data-plane traffic is difficult. Indeed an 719 attacker cannot gain information about network performance from a 720 single monitoring point. The only way for an attacker can be to 721 eavesdrop on multiple monitoring points at the same time, because 722 they have to do the same kind of calculation and aggregation as 723 Alternate Marking requires, and, after that, it can finally gain 724 information about the network performance, but this is not 725 immediate. 727 The FlowMonID field is used in the AltMark Option as identifier of 728 the monitored flow. It represents a more sensitive information 729 for network reconnaissance and may allow a flow tracking type of 730 attack because an attacker could collect information about network 731 paths. 733 Furthermore, in a pervasive surveillance attack, the information that 734 can be derived over time is more. But the application of the 735 Alternate Marking to a controlled domain helps to mitigate all the 736 above aspects of privacy concerns. 738 At the management plane, attacks can be set up by misconfiguring or 739 by maliciously configuring AltMark Option. Thus, AltMark Option 740 configuration MUST be secured in a way that authenticates authorized 741 users and verifies the integrity of configuration procedures. 742 Solutions to ensure the integrity of AltMark Option are outside the 743 scope of this document. 745 As stated above, the precondition for the application of the 746 Alternate Marking is that it MUST be applied in specific controlled 747 domains, thus confining the potential attack vectors within the 748 network domain. [RFC8799] analyzes and discusses the trend towards 749 network behaviors that can be applied only within a limited domain. 750 This is due to the specific set of requirements especially related to 751 security, network management, policies and options supported which 752 may vary between such limited domains. A limited administrative 753 domain provides the network administrator with the means to select, 754 monitor and control the access to the network, making it a trusted 755 domain. In this regard it is expected to enforce policies at the 756 domain boundaries to filter both external packets with AltMark Option 757 entering the domain and internal packets with AltMark Option leaving 758 the domain. Therefore the trusted domain is unlikely subject to 759 hijacking of packets since packets with AltMark Option are processed 760 and used only within the controlled domain. 762 Additionally, it is to be noted that the AltMark Option is carried by 763 the Options Header and it may have some impact on the packet sizes 764 for the monitored flow and on the path MTU, since some packets might 765 exceed the MTU. However the relative small size (48 bit in total) of 766 these Option Headers and its application to a controlled domain help 767 to mitigate the problem. 769 It is worth mentioning that the security concerns may change based on 770 the specific deployment scenario and related threat analysis, which 771 can lead to specific security solutions that are beyond the scope of 772 this document. As an example, the AltMark Option can be used as Hop- 773 by-Hop or Destination Option and, in case of Destination Option, 774 multiple domains may be traversed by the AltMark Option that is not 775 confined to a single domain. In this case, the user, aware of the 776 kind of risks, may still want to use Alternate Marking for telemetry 777 and test purposes but the inter-domain links need to be secured 778 (e.g., by IPsec) in order to avoid external threats. 780 The Alternate Marking application described in this document relies 781 on an time synchronization protocol. Thus, by attacking the time 782 protocol, an attacker can potentially compromise the integrity of the 783 measurement. A detailed discussion about the threats against time 784 protocols and how to mitigate them is presented in [RFC7384]. Also, 785 the time, which is distributed to the network nodes through the time 786 protocol, is centrally taken from an external accurate time source, 787 such as an atomic clock or a GPS clock, and by attacking the time 788 source it can be possible to compromise the integrity of the 789 measurement as well. There are security measures that can be taken 790 to mitigate the GPS spoofing attacks and a network administrator 791 should certainly employ solutions to secure the network domain. 793 7. IANA Considerations 795 The Option Type should be assigned in IANA's "Destination Options and 796 Hop-by-Hop Options" registry. 798 This draft requests the following IPv6 Option Type assignment from 799 the Destination Options and Hop-by-Hop Options sub-registry of 800 Internet Protocol Version 6 (IPv6) Parameters 801 (https://www.iana.org/assignments/ipv6-parameters/). 803 Hex Value Binary Value Description Reference 804 act chg rest 805 ---------------------------------------------------------------- 806 TBD 00 0 tbd AltMark [This draft] 808 8. Acknowledgements 810 The authors would like to thank Bob Hinden, Ole Troan, Stewart 811 Bryant, Christopher Wood, Yoshifumi Nishida, Tom Herbert, Stefano 812 Previdi, Brian Carpenter, Eric Vyncke, Greg Mirsky, Ron Bonica for 813 the precious comments and suggestions. 815 9. References 817 9.1. Normative References 819 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 820 Requirement Levels", BCP 14, RFC 2119, 821 DOI 10.17487/RFC2119, March 1997, 822 . 824 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 825 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 826 May 2017, . 828 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 829 (IPv6) Specification", STD 86, RFC 8200, 830 DOI 10.17487/RFC8200, July 2017, 831 . 833 9.2. Informative References 835 [I-D.chen-pce-pcep-ifit] 836 Chen, H., Yuan, H., Zhou, T., Li, W., Fioccola, G., and Y. 837 Wang, "Path Computation Element Communication Protocol 838 (PCEP) Extensions to Enable IFIT", draft-chen-pce-pcep- 839 ifit-02 (work in progress), February 2021. 841 [I-D.fioccola-v6ops-ipv6-alt-mark] 842 Fioccola, G., Velde, G. V. D., Cociglio, M., and P. Muley, 843 "IPv6 Performance Measurement with Alternate Marking 844 Method", draft-fioccola-v6ops-ipv6-alt-mark-01 (work in 845 progress), June 2018. 847 [I-D.fz-spring-srv6-alt-mark] 848 Fioccola, G., Zhou, T., and M. Cociglio, "Segment Routing 849 Header encapsulation for Alternate Marking Method", draft- 850 fz-spring-srv6-alt-mark-00 (work in progress), January 851 2021. 853 [I-D.hinden-6man-hbh-processing] 854 Hinden, R. M. and G. Fairhurst, "IPv6 Hop-by-Hop Options 855 Processing Procedures", draft-hinden-6man-hbh- 856 processing-00 (work in progress), December 2020. 858 [I-D.ietf-idr-sr-policy-ifit] 859 Qin, F., Yuan, H., Zhou, T., Fioccola, G., and Y. Wang, 860 "BGP SR Policy Extensions to Enable IFIT", draft-ietf-idr- 861 sr-policy-ifit-01 (work in progress), February 2021. 863 [I-D.peng-v6ops-hbh] 864 Peng, S., Li, Z., Xie, C., Qin, Z., and G. Mishra, 865 "Processing of the Hop-by-Hop Options Header", draft-peng- 866 v6ops-hbh-03 (work in progress), January 2021. 868 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 869 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 870 2006, . 872 [RFC6437] Amante, S., Carpenter, B., Jiang, S., and J. Rajahalme, 873 "IPv6 Flow Label Specification", RFC 6437, 874 DOI 10.17487/RFC6437, November 2011, 875 . 877 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 878 for Equal Cost Multipath Routing and Link Aggregation in 879 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 880 . 882 [RFC7045] Carpenter, B. and S. Jiang, "Transmission and Processing 883 of IPv6 Extension Headers", RFC 7045, 884 DOI 10.17487/RFC7045, December 2013, 885 . 887 [RFC7384] Mizrahi, T., "Security Requirements of Time Protocols in 888 Packet Switched Networks", RFC 7384, DOI 10.17487/RFC7384, 889 October 2014, . 891 [RFC8321] Fioccola, G., Ed., Capello, A., Cociglio, M., Castaldelli, 892 L., Chen, M., Zheng, L., Mirsky, G., and T. Mizrahi, 893 "Alternate-Marking Method for Passive and Hybrid 894 Performance Monitoring", RFC 8321, DOI 10.17487/RFC8321, 895 January 2018, . 897 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 898 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 899 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 900 . 902 [RFC8799] Carpenter, B. and B. Liu, "Limited Domains and Internet 903 Protocols", RFC 8799, DOI 10.17487/RFC8799, July 2020, 904 . 906 [RFC8889] Fioccola, G., Ed., Cociglio, M., Sapio, A., and R. Sisto, 907 "Multipoint Alternate-Marking Method for Passive and 908 Hybrid Performance Monitoring", RFC 8889, 909 DOI 10.17487/RFC8889, August 2020, 910 . 912 Authors' Addresses 914 Giuseppe Fioccola 915 Huawei 916 Riesstrasse, 25 917 Munich 80992 918 Germany 920 Email: giuseppe.fioccola@huawei.com 922 Tianran Zhou 923 Huawei 924 156 Beiqing Rd. 925 Beijing 100095 926 China 928 Email: zhoutianran@huawei.com 930 Mauro Cociglio 931 Telecom Italia 932 Via Reiss Romoli, 274 933 Torino 10148 934 Italy 936 Email: mauro.cociglio@telecomitalia.it 937 Fengwei Qin 938 China Mobile 939 32 Xuanwumenxi Ave. 940 Beijing 100032 941 China 943 Email: qinfengwei@chinamobile.com 945 Ran Pang 946 China Unicom 947 9 Shouti South Rd. 948 Beijing 100089 949 China 951 Email: pangran@chinaunicom.cn