idnits 2.17.1 draft-ietf-ippm-alt-mark-06.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 (July 25, 2017) is 2466 days in the past. Is this intentional? Checking references for intended status: Experimental ---------------------------------------------------------------------------- == Outdated reference: A later version (-12) exists of draft-ietf-bier-mpls-encapsulation-07 == Outdated reference: A later version (-15) exists of draft-ietf-bier-pmmm-oam-02 == Outdated reference: A later version (-07) exists of draft-ietf-mpls-flow-ident-04 == Outdated reference: A later version (-10) exists of draft-ietf-mpls-rfc6374-sfl-00 == Outdated reference: A later version (-12) exists of draft-ietf-nvo3-encap-00 == Outdated reference: A later version (-14) exists of draft-mirsky-sfc-pmamm-01 Summary: 0 errors (**), 0 flaws (~~), 7 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group G. Fioccola, Ed. 3 Internet-Draft A. Capello, Ed. 4 Intended status: Experimental M. Cociglio 5 Expires: January 26, 2018 L. Castaldelli 6 Telecom Italia 7 M. Chen, Ed. 8 L. Zheng, Ed. 9 Huawei Technologies 10 G. Mirsky, Ed. 11 ZTE 12 T. Mizrahi, Ed. 13 Marvell 14 July 25, 2017 16 Alternate Marking method for passive and hybrid performance monitoring 17 draft-ietf-ippm-alt-mark-06 19 Abstract 21 This document describes a method to perform packet loss, delay and 22 jitter measurements on live traffic. This method is based on 23 Alternate Marking (Coloring) technique. A report on the operational 24 experiment done at Telecom Italia is explained in order to give an 25 example and show the method applicability. This technique can be 26 applied in various situations as detailed in this document and could 27 be considered passive or hybrid depending on the application. 29 Requirements Language 31 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 32 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 33 "OPTIONAL" in this document are to be interpreted as described in BCP 34 14 [RFC2119] [RFC8174] when, and only when, they appear in all 35 capitals, as shown here. 37 Status of This Memo 39 This Internet-Draft is submitted in full conformance with the 40 provisions of BCP 78 and BCP 79. 42 Internet-Drafts are working documents of the Internet Engineering 43 Task Force (IETF). Note that other groups may also distribute 44 working documents as Internet-Drafts. The list of current Internet- 45 Drafts is at http://datatracker.ietf.org/drafts/current/. 47 Internet-Drafts are draft documents valid for a maximum of six months 48 and may be updated, replaced, or obsoleted by other documents at any 49 time. It is inappropriate to use Internet-Drafts as reference 50 material or to cite them other than as "work in progress." 52 This Internet-Draft will expire on January 26, 2018. 54 Copyright Notice 56 Copyright (c) 2017 IETF Trust and the persons identified as the 57 document authors. All rights reserved. 59 This document is subject to BCP 78 and the IETF Trust's Legal 60 Provisions Relating to IETF Documents 61 (http://trustee.ietf.org/license-info) in effect on the date of 62 publication of this document. Please review these documents 63 carefully, as they describe your rights and restrictions with respect 64 to this document. Code Components extracted from this document must 65 include Simplified BSD License text as described in Section 4.e of 66 the Trust Legal Provisions and are provided without warranty as 67 described in the Simplified BSD License. 69 Table of Contents 71 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 72 2. Overview of the method . . . . . . . . . . . . . . . . . . . 4 73 3. Detailed description of the method . . . . . . . . . . . . . 6 74 3.1. Packet loss measurement . . . . . . . . . . . . . . . . . 6 75 3.2. Timing aspects . . . . . . . . . . . . . . . . . . . . . 10 76 3.3. One-way delay measurement . . . . . . . . . . . . . . . . 11 77 3.3.1. Single marking methodology . . . . . . . . . . . . . 11 78 3.3.2. Double marking methodology . . . . . . . . . . . . . 13 79 3.4. Delay variation measurement . . . . . . . . . . . . . . . 14 80 4. Considerations . . . . . . . . . . . . . . . . . . . . . . . 15 81 4.1. Synchronization . . . . . . . . . . . . . . . . . . . . . 15 82 4.2. Data Correlation . . . . . . . . . . . . . . . . . . . . 15 83 4.3. Packet Re-ordering . . . . . . . . . . . . . . . . . . . 16 84 5. Implementation and deployment . . . . . . . . . . . . . . . . 17 85 5.1. Report on the operational experiment at Telecom Italia . 17 86 5.1.1. Coloring the packets . . . . . . . . . . . . . . . . 19 87 5.1.2. Counting the packets . . . . . . . . . . . . . . . . 20 88 5.1.3. Collecting data and calculating packet loss . . . . . 21 89 5.1.4. Metric transparency . . . . . . . . . . . . . . . . . 22 90 5.2. IP flow performance measurement (IPFPM) . . . . . . . . . 22 91 5.3. OAM Passive Performance Measurement . . . . . . . . . . . 22 92 5.4. RFC6374 Use Case . . . . . . . . . . . . . . . . . . . . 22 93 5.5. Application to active performance measurement . . . . . . 23 94 6. Hybrid measurement . . . . . . . . . . . . . . . . . . . . . 23 95 7. Compliance with RFC6390 guidelines . . . . . . . . . . . . . 23 96 8. Security Considerations . . . . . . . . . . . . . . . . . . . 25 97 9. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 26 98 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 27 99 11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 27 100 12. References . . . . . . . . . . . . . . . . . . . . . . . . . 27 101 12.1. Normative References . . . . . . . . . . . . . . . . . . 27 102 12.2. Informative References . . . . . . . . . . . . . . . . . 28 103 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 30 105 1. Introduction 107 Nowadays, most of the traffic in Service Providers' networks carries 108 contents that are highly sensitive to packet loss [RFC7680], delay 109 [RFC7679], and jitter [RFC3393]. 111 In view of this scenario, Service Providers need methodologies and 112 tools to monitor and measure network performances with an adequate 113 accuracy, in order to constantly control the quality of experience 114 perceived by their customers. On the other hand, performance 115 monitoring provides useful information for improving network 116 management (e.g. isolation of network problems, troubleshooting, 117 etc.). 119 A lot of work related to OAM, that includes also performance 120 monitoring techniques, has been done by Standards Developing 121 Organizations(SDOs): [RFC7276] provides a good overview of existing 122 OAM mechanisms defined in IETF, ITU-T and IEEE. Considering IETF, a 123 lot of work has been done on fault detection and connectivity 124 verification, while a minor effort has been dedicated so far to 125 performance monitoring. The IPPM WG has defined standard metrics to 126 measure network performance; however, the methods developed in this 127 WG mainly refer to focus on active measurement techniques. More 128 recently, the MPLS WG has defined mechanisms for measuring packet 129 loss, one-way and two-way delay, and delay variation in MPLS 130 networks[RFC6374], but their applicability to passive measurements 131 has some limitations, especially for pure connection-less networks. 133 The lack of adequate tools to measure packet loss with the desired 134 accuracy drove an effort to design a new method for the performance 135 monitoring of live traffic, possibly easy to implement and deploy. 136 The effort led to the method described in this document: basically, 137 it is a passive performance monitoring technique, potentially 138 applicable to any kind of packet based traffic, including Ethernet, 139 IP, and MPLS, both unicast and multicast. The method addresses 140 primarily packet loss measurement, but it can be easily extended to 141 one-way delay and delay variation measurements as well. 143 The method has been explicitly designed for passive measurements but 144 it can also be used with active probes. Passive measurements are 145 usually more easily understood by customers and provide a much better 146 accuracy, especially for packet loss measurements. 148 RFC 7799 [RFC7799] defines passive and hybrid methods of measurement. 149 In particular, Passive Methods of Measurement are based solely on 150 observations of an undisturbed and unmodified packet stream of 151 interest; Hybrid Methods are Methods of Measurement that use a 152 combination of Active Methods and Passive Methods. 154 Taking into consideration these definitions, Alternate Marking Method 155 could be considered Hybrid or Passive depending on the case. In case 156 the marking field is obtained by changing existing field values of 157 the packets (e.g. DSCP field), the technique is Hybrid. In case the 158 marking field is dedicated, reserved and is included in the protocol 159 specification Alternate Marking technique can be considered as 160 Passive (e.g. RFC6374 Synonymous Flow Label or OAM Marking Bits in 161 BIER Header). 163 This document is organized as follows: 165 o Section 2 gives an overview of the method, including a comparison 166 with different measurement strategies; 168 o Section 3 describes the method in detail; 170 o Section 4 reports considerations about synchronization, data 171 correlation and packet re-ordering; 173 o Section 5 reports examples of implementation and deployment of the 174 method. Furthermore the operational experiment done at Telecom 175 Italia is described; 177 o Section 6 introduces Hybrid measurement aspects; 179 o Section 7 is about the Compliance with RFC6390 guidelines; 181 o Section 8 includes some security aspects; 183 o Section 9 finally summarizes some concluding remarks. 185 2. Overview of the method 187 In order to perform packet loss measurements on a live traffic flow, 188 different approaches exist. The most intuitive one consists in 189 numbering the packets, so that each router that receives the flow can 190 immediately detect a packet missing. This approach, though very 191 simple in theory, is not simple to achieve: it requires the insertion 192 of a sequence number into each packet and the devices must be able to 193 extract the number and check it in real time. Such a task can be 194 difficult to implement on live traffic: if UDP is used as the 195 transport protocol, the sequence number is not available; on the 196 other hand, if a higher layer sequence number (e.g. in the RTP 197 header) is used, extracting that information from each packet and 198 process it in real time could overload the device. 200 An alternate approach is to count the number of packets sent on one 201 end, the number of packets received on the other end, and to compare 202 the two values. This operation is much simpler to implement, but 203 requires that the devices performing the measurement are in sync: in 204 order to compare two counters it is required that they refer exactly 205 to the same set of packets. Since a flow is continuous and cannot be 206 stopped when a counter has to be read, it could be difficult to 207 determine exactly when to read the counter. A possible solution to 208 overcome this problem is to virtually split the flow in consecutive 209 blocks by inserting periodically a delimiter so that each counter 210 refers exactly to the same block of packets. The delimiter could be 211 for example a special packet inserted artificially into the flow. 212 However, delimiting the flow using specific packets has some 213 limitations. First, it requires generating additional packets within 214 the flow and requires the equipment to be able to process those 215 packets. In addition, the method is vulnerable to out of order 216 reception of delimiting packets and, to a lesser extent, to their 217 loss. 219 The method proposed in this document follows the second approach, but 220 it doesn't use additional packets to virtually split the flow in 221 blocks. Instead, it "colors" the packets so that the packets 222 belonging to the same block will have the same color, whilst 223 consecutive blocks will have different colors. Each change of color 224 represents a sort of auto-synchronization signal that guarantees the 225 consistency of measurements taken by different devices along the 226 path. 228 Figure 1 represents a very simple network and shows how the method 229 can be used to measure packet loss on different network segments: by 230 enabling the measurement on several interfaces along the path, it is 231 possible to perform link monitoring, node monitoring or end-to-end 232 monitoring. The method is flexible enough to measure packet loss on 233 any segment of the network and can be used to isolate the faulty 234 element. 236 Traffic flow 237 ========================================================> 238 +------+ +------+ +------+ +------+ 239 ---<> R1 <>-----<> R2 <>-----<> R3 <>-----<> R4 <>--- 240 +------+ +------+ +------+ +------+ 241 . . . . . . 242 . . . . . . 243 . <------> <-------> . 244 . Node Packet Loss Link Packet Loss . 245 . . 246 <---------------------------------------------------> 247 End-to-End Packet loss 249 Figure 1: Available measurements 251 3. Detailed description of the method 253 This section describes in detail how the method operate. A special 254 emphasis is given to the measurement of packet loss, that represents 255 the core application of the method, but applicability to delay and 256 jitter measurements is also considered. 258 3.1. Packet loss measurement 260 The basic idea is to virtually split traffic flows into consecutive 261 blocks: each block represents a measurable entity unambiguously 262 recognizable by all network devices along the path. By counting the 263 number of packets in each block and comparing the values measured by 264 different network devices along the path, it is possible to measure 265 packet loss occurred in any single block between any two points. 267 As discussed in the previous section, a simple way to create the 268 blocks is to "color" the traffic (two colors are sufficient) so that 269 packets belonging to different consecutive blocks will have different 270 colors. Whenever the color changes, the previous block terminates 271 and the new one begins. Hence, all the packets belonging to the same 272 block will have the same color and packets of different consecutive 273 blocks will have different colors. The number of packets in each 274 block depends on the criterion used to create the blocks: if the 275 color is switched after a fixed number of packets, then each block 276 will contain the same number of packets (except for any losses); but 277 if the color is switched according to a fixed timer, then the number 278 of packets may be different in each block depending on the packet 279 rate. 281 The following figure shows how a flow looks like when it is split in 282 traffic blocks with colored packets. 284 A: packet with A coloring 285 B: packet with B coloring 287 | | | | | 288 | | Traffic flow | | 289 -------------------------------------------------------------------> 290 BBBBBBB AAAAAAAAAAA BBBBBBBBBBB AAAAAAAAAAA BBBBBBBBBBB AAAAAAA 291 -------------------------------------------------------------------> 292 ... | Block 5 | Block 4 | Block 3 | Block 2 | Block 1 293 | | | | | 295 Figure 2: Traffic coloring 297 Figure 3 shows how the method can be used to measure link packet loss 298 between two adjacent nodes. 300 Referring to the figure, let's assume we want to monitor the packet 301 loss on the link between two routers: router R1 and router R2. 302 According to the method, the traffic is colored alternatively with 303 two different colors, A and B. Whenever the color changes, the 304 transition generates a sort of square-wave signal, as depicted in the 305 following figure. 307 Color A ----------+ +-----------+ +---------- 308 | | | | 309 Color B +-----------+ +-----------+ 310 Block n ... Block 3 Block 2 Block 1 311 <---------> <---------> <---------> <---------> <---------> 313 Traffic flow 314 ===========================================================> 315 Color ...AAAAAAAAAAA BBBBBBBBBBB AAAAAAAAAAA BBBBBBBBBBB AAAAAAA... 316 ===========================================================> 318 Figure 3: Computation of link packet loss 320 Traffic coloring could be done by R1 itself or by an upward router. 321 R1 needs two counters, C(A)R1 and C(B)R1, on its egress interface: 322 C(A)R1 counts the packets with color A and C(B)R1 counts those with 323 color B. As long as traffic is colored A, only counter C(A)R1 will 324 be incremented, while C(B)R1 is not incremented; vice versa, when the 325 traffic is colored as B, only C(B)R1 is incremented. C(A)R1 and 326 C(B)R1 can be used as reference values to determine the packet loss 327 from R1 to any other measurement point down the path. Router R2, 328 similarly, will need two counters on its ingress interface, C(A)R2 329 and C(B)R2, to count the packets received on that interface and 330 colored with color A and B respectively. When an A block ends, it is 331 possible to compare C(A)R1 and C(A)R2 and calculate the packet loss 332 within the block; similarly, when the successive B block terminates, 333 it is possible to compare C(B)R1 with C(B)R2, and so on for every 334 successive block. 336 Likewise, by using two counters on R2 egress interface it is possible 337 to count the packets sent out of R2 interface and use them as 338 reference values to calculate the packet loss from R2 to any 339 measurement point down R2. 341 Using a fixed timer for color switching offers a better control over 342 the method: the (time) length of the blocks can be chosen large 343 enough to simplify the collection and the comparison of measures 344 taken by different network devices. It's preferable to read the 345 value of the counters not immediately after the color switch: some 346 packets could arrive out of order and increment the counter 347 associated to the previous block (color), so it is worth waiting for 348 some time. A safe choice is to wait L/2 time units (where L is the 349 duration for each block) after the color switch, to read the still 350 counter of the previous color, so the possibility to read a running 351 counter instead of a still one is minimized. The drawback is that 352 the longer the duration of the block, the less frequent the 353 measurement can be taken. 355 The following table shows how the counters can be used to calculate 356 the packet loss between R1 and R2. The first column lists the 357 sequence of traffic blocks while the other columns contain the 358 counters of A-colored packets and B-colored packets for R1 and R2. 359 In this example, we assume that the values of the counters are reset 360 to zero whenever a block ends and its associated counter has been 361 read: with this assumption, the table shows only relative values, 362 that is the exact number of packets of each color within each block. 363 If the values of the counters were not reset, the table would contain 364 cumulative values, but the relative values could be determined simply 365 by difference from the value of the previous block of the same color. 367 The color is switched on the basis of a fixed timer (not shown in the 368 table), so the number of packets in each block is different. 370 +-------+--------+--------+--------+--------+------+ 371 | Block | C(A)R1 | C(B)R1 | C(A)R2 | C(B)R2 | Loss | 372 +-------+--------+--------+--------+--------+------+ 373 | 1 | 375 | 0 | 375 | 0 | 0 | 374 | | | | | | | 375 | 2 | 0 | 388 | 0 | 388 | 0 | 376 | | | | | | | 377 | 3 | 382 | 0 | 381 | 0 | 1 | 378 | | | | | | | 379 | 4 | 0 | 377 | 0 | 374 | 3 | 380 | | | | | | | 381 | ... | ... | ... | ... | ... | ... | 382 | | | | | | | 383 | 2n | 0 | 387 | 0 | 387 | 0 | 384 | | | | | | | 385 | 2n+1 | 379 | 0 | 377 | 0 | 2 | 386 +-------+--------+--------+--------+--------+------+ 388 Table 1: Evaluation of counters for packet loss measurements 390 During an A block (blocks 1, 3 and 2n+1), all the packets are 391 A-colored, therefore the C(A) counters are incremented to the number 392 seen on the interface, while C(B) counters are zero. Vice versa, 393 during a B block (blocks 2, 4 and 2n), all the packets are B-colored: 394 C(A) counters are zero, while C(B) counters are incremented. 396 When a block ends (because of color switching) the relative counters 397 stop incrementing and it is possible to read them, compare the values 398 measured on router R1 and R2 and calculate the packet loss within 399 that block. 401 For example, looking at the table above, during the first block 402 (A-colored), C(A)R1 and C(A)R2 have the same value (375), which 403 corresponds to the exact number of packets of the first block (no 404 loss). Also during the second block (B-colored) R1 and R2 counters 405 have the same value (388), which corresponds to the number of packets 406 of the second block (no loss). During blocks three and four, R1 and 407 R2 counters are different, meaning that some packets have been lost: 408 in the example, one single packet (382-381) was lost during block 409 three and three packets (377-374) were lost during block four. 411 The method applied to R1 and R2 can be extended to any other router 412 and applied to more complex networks, as far as the measurement is 413 enabled on the path followed by the traffic flow(s) being observed. 415 3.2. Timing aspects 417 This document introduces two color switching method: one is based on 418 fixed number of packet, the other is based on fixed timer. But the 419 method based on fixed timer is preferable because is more 420 deterministic, and will be considered in the rest of the dcoument. 422 By considering the clock error between network devices R1 and R2, 423 they must be synchronized to the same clock reference with an 424 accuracy of +/- L/2 time units, where L is the time duration of the 425 block. So each colored packet can be assigned to the right batch by 426 each router. This is because the minimum time distance between two 427 packets of the same color but belonging to different batches is L 428 time units. 430 In practice, there are also out of order at batch boundaries, 431 strictly related to the delay between measurement points. This means 432 that, without considering clock error, we wait L/2 after color 433 switching to be sure to take a still counter. 435 In summary we need to take into account two contributions: clock 436 error between network devices and the interval we need to wait to 437 avoid out of order because of network delay. 439 The following figure explains both issues. 441 ...BBBBBBBBB | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA | BBBBBBBBB... 442 |<======================================>| 443 | L | 444 ...=========>|<==================><==================>|<==========... 445 | L/2 L/2 | 446 |<===>| |<===>| 447 d | | d 448 |<==========================>| 449 available counting interval 451 Figure 4: Timing aspects 453 It is assumed that all network devices are synchronized to a common 454 reference time with an accuracy of +/- A/2. Thus, the difference 455 between the clock values of any two network devices is bounded by A. 457 The guardband d is given by: 459 d = A + D_max - D_min, 460 where A is the clock accuracy, D_max is an upper bound on the network 461 delay between the network devices, and D_min is a lower bound on the 462 delay. 464 The available counting interval is L - 2d that must be > 0. 466 The condition that must be satisfied and is a requirement on the 467 synchronization accuracy is: 469 d < L/2. 471 3.3. One-way delay measurement 473 The same principle used to measure packet loss can be applied also to 474 one-way delay measurement. There are three alternatives, as 475 described hereinafter. 477 3.3.1. Single marking methodology 479 The alternation of colors can be used as a time reference to 480 calculate the delay. Whenever the color changes (that means that a 481 new block has started) a network device can store the timestamp of 482 the first packet of the new block; that timestamp can be compared 483 with the timestamp of the same packet on a second router to compute 484 packet delay. Considering Figure 2, R1 stores a timestamp TS(A1)R1 485 when it sends the first packet of block 1 (A-colored), a timestamp 486 TS(B2)R1 when it sends the first packet of block 2 (B-colored) and so 487 on for every other block. R2 performs the same operation on the 488 receiving side, recording TS(A1)R2, TS(B2)R2 and so on. Since the 489 timestamps refer to specific packets (the first packet of each block) 490 we are sure that timestamps compared to compute delay refer to the 491 same packets. By comparing TS(A1)R1 with TS(A1)R2 (and similarly 492 TS(B2)R1 with TS(B2)R2 and so on) it is possible to measure the delay 493 between R1 and R2. In order to have more measurements, it is 494 possible to take and store more timestamps, referring to other 495 packets within each block. 497 In order to coherently compare timestamps collected on different 498 routers, the network nodes must be in sync. Furthermore, a 499 measurement is valid only if no packet loss occurs and if packet 500 misordering can be avoided, otherwise the first packet of a block on 501 R1 could be different from the first packet of the same block on R2 502 (f.i. if that packet is lost between R1 and R2 or it arrives after 503 the next one). 505 The following table shows how timestamps can be used to calculate the 506 delay between R1 and R2. The first column lists the sequence of 507 blocks while other columns contain the timestamp referring to the 508 first packet of each block on R1 and R2. The delay is computed as a 509 difference between timestamps. For the sake of simplicity, all the 510 values are expressed in milliseconds. 512 +-------+---------+---------+---------+---------+-------------+ 513 | Block | TS(A)R1 | TS(B)R1 | TS(A)R2 | TS(B)R2 | Delay R1-R2 | 514 +-------+---------+---------+---------+---------+-------------+ 515 | 1 | 12.483 | - | 15.591 | - | 3.108 | 516 | | | | | | | 517 | 2 | - | 6.263 | - | 9.288 | 3.025 | 518 | | | | | | | 519 | 3 | 27.556 | - | 30.512 | - | 2.956 | 520 | | | | | | | 521 | | - | 18.113 | - | 21.269 | 3.156 | 522 | | | | | | | 523 | ... | ... | ... | ... | ... | ... | 524 | | | | | | | 525 | 2n | 77.463 | - | 80.501 | - | 3.038 | 526 | | | | | | | 527 | 2n+1 | - | 24.333 | - | 27.433 | 3.100 | 528 +-------+---------+---------+---------+---------+-------------+ 530 Table 2: Evaluation of timestamps for delay measurements 532 The first row shows timestamps taken on R1 and R2 respectively and 533 referring to the first packet of block 1 (which is A-colored). Delay 534 can be computed as a difference between the timestamp on R2 and the 535 timestamp on R1. Similarly, the second row shows timestamps (in 536 milliseconds) taken on R1 and R2 and referring to the first packet of 537 block 2 (which is B-colored). Comparing timestamps taken on 538 different nodes in the network and referring to the same packets 539 (identified using the alternation of colors) it is possible to 540 measure delay on different network segments. 542 For the sake of simplicity, in the above example a single measurement 543 is provided within a block, taking into account only the first packet 544 of each block. The number of measurements can be easily increased by 545 considering multiple packets in the block: for instance, a timestamp 546 could be taken every N packets, thus generating multiple delay 547 measurements. Taking this to the limit, in principle the delay could 548 be measured for each packet, by taking and comparing the 549 corresponding timestamps (possible but impractical from an 550 implementation point of view). 552 3.3.1.1. Mean delay 554 As mentioned before, the method previously exposed for measuring the 555 delay is sensitive to out of order reception of packets. In order to 556 overcome this problem, a different approach has been considered: it 557 is based on the concept of mean delay. The mean delay is calculated 558 by considering the average arrival time of the packets within a 559 single block. The network device locally stores a timestamp for each 560 packet received within a single block: summing all the timestamps and 561 dividing by the total number of packets received, the average arrival 562 time for that block of packets can be calculated. By subtracting the 563 average arrival times of two adjacent devices it is possible to 564 calculate the mean delay between those nodes. When computing the 565 mean delay, measurement error could be augmented by accumulating 566 measurement error of a lot of packets. This method is robust to out 567 of order packets and also to packet loss (only a small error is 568 introduced). Moreover, it greatly reduces the number of timestamps 569 (only one per block for each network device) that have to be 570 collected by the management system. On the other hand, it only gives 571 one measure for the duration of the block (f.i. 5 minutes), and it 572 doesn't give the minimum, maximum and median delay values (RFC 6703 573 [RFC6703]). This limitation could be overcome by reducing the 574 duration of the block (f.i. from 5 minutes to a few seconds), that 575 implicates an highly optimized implementation of the method. 577 By summing the mean delays of the two directions of a path, it is 578 also possible to measure the two-way mean delay (round-trip delay). 580 3.3.2. Double marking methodology 582 The Single marking methodology for one-way delay measurement is 583 sensitive to out of order reception of packets. The first approach 584 to overcome this problem is described before and is based on the 585 concept of mean delay. But the limitation of mean delay is that it 586 doesn't give information about the delay values distribution for the 587 duration of the block. Additionally it may be useful to have not 588 only the mean delay but also the minimum, maximum and median delay 589 values and, in wider terms, to know more about the statistic 590 distribution of delay values. So in order to have more information 591 about the delay and to overcome out of order issues, a different 592 approach can be introduced: it is based on double marking 593 methodology. 595 Basically, the idea is to use the first marking to create the 596 alternate flow and, within this colored flow, a second marking to 597 select the packets for measuring delay/jitter. The first marking is 598 needed for packet loss and mean delay measurement. The second 599 marking creates a new set of marked packets that are fully identified 600 over the network, so that a network device can store the timestamps 601 of these packets; these timestamps can be compared with the 602 timestamps of the same packets on a second router to compute packet 603 delay values for each packet. The number of measurements can be 604 easily increased by changing the frequency of the second marking. 605 But the frequency of the second marking must be not too high in order 606 to avoid out of order issues. Between packets with the second 607 marking there should be a security time gap (e.g. this gap could be, 608 at the minimum, the mean network delay calculated with the previous 609 methodology) to avoid out of order issues and also to have a number 610 of measurement packets that is rate independent. If a second marking 611 packet is lost, the delay measurement for the considered block is 612 corrupted and should be discarded. 614 Mean delay is calculated on all the packets of a sample and is a 615 simple computation to be performed for single marking method. In 616 some cases the mean delay measure is not sufficient to characterize 617 the sample, and more statistics of delay extent data are needed, e.g. 618 percentiles, variance and median delay values. The conventional 619 range (maximum-minimum) should be avoided for several reasons, 620 including stability of the maximum delay due to the influence by 621 outliers. RFC 5481 [RFC5481] section 6.5 highlights how the 99.9th 622 percentile of delay and delay variation is more helpful to 623 performance planners. To overcome this drawback the idea is to 624 couple the mean delay measure for the entire batch with double 625 marking method, where a subset of batch packets are selected for 626 extensive delay calculation by using a second marking. In this way 627 it is possible to perform a detailed analysis on these double marked 628 packets. Please note that there are classic algorithms for median 629 and variance calculation, but are out of the scope of this document. 630 The comparison between the mean delay for the entire batch and the 631 mean delay on these double marked packets gives an useful information 632 since it is possible to understand if the double marking measurements 633 are actually representative of the delay trends. 635 3.4. Delay variation measurement 637 Similarly to one-way delay measurement (both for single marking and 638 double marking), the method can also be used to measure the inter- 639 arrival jitter. We refer to the definition in RFC 3393 [RFC3393]. 640 The alternation of colors, for single marking method, can be used as 641 a time reference to measure delay variations. In case of double 642 marking, the time reference is given by the second marked packets. 643 Considering the example depicted in Figure 2, R1 stores a timestamp 644 TS(A)R1 whenever it sends the first packet of a block and R2 stores a 645 timestamp TS(B)R2 whenever it receives the first packet of a block. 646 The inter-arrival jitter can be easily derived from one-way delay 647 measurement, by evaluating the delay variation of consecutive 648 samples. 650 The concept of mean delay can also be applied to delay variation, by 651 evaluating the average variation of the interval between consecutive 652 packets of the flow from R1 to R2. 654 4. Considerations 656 This section highlights some considerations about the methodology. 658 4.1. Synchronization 660 The Alternate Marking technique does not require a strong 661 synchronization, especially for packet loss and two-way delay 662 measurement. Only one-way delay measurement requires network devices 663 to have synchronized clocks. 665 The color switching is the reference for all the network devices, and 666 the only requirement to be achieved is that all network devices have 667 to recognize the right batch along the path. 669 If the length of the measurement period is L time units, then all 670 network devices must be synchronized to the same clock reference with 671 an accuracy of +/- L/2 time units (without considering network 672 delay). This level of accuracy guarantees that all network devices 673 consistently match the color bit to the correct block. For example, 674 if the color is toggeled every second (L = 1 second), then clocks 675 must be synchronized with an accuracy of +/- 0.5 second to a common 676 time reference. 678 This synchronization requirement can be satisfied even with a 679 relatively inaccurate synchronization method. This is true for 680 packet loss and two-way delay measurement, instead, for one-way delay 681 measurement clock synchronization must be accurate. 683 Therefore, a system that uses only packet loss and two-way delay 684 measurement does not require synchronization. This is because the 685 value of the clocks of network devices does not affect the 686 computation of the two-way delay measurement. 688 4.2. Data Correlation 690 Data Correlation is the mechanism to compare counters and timestamps 691 for packet loss, delay and delay variation calculation. It could be 692 performed in several ways depending on the alternate marking 693 application and use case. 695 o A possibility is to use a centralized solution using Network 696 Management System (NMS) to correlate data; 698 o Another possibility is to define a protocol based distributed 699 solution, by defining a new protocol or by extending the existing 700 protocols (e.g. RFC6374, TWAMP, OWAMP) in order to communicate 701 the counters and timestamps between nodes. 703 In the following paragraphs an example data correlation mechanism is 704 explained and could be use independently of the adopted solutions. 706 When data is collected on the upstream and downstream node, e.g., 707 packet counts for packet loss measurement or timestamps for packet 708 delay measurement, and periodically reported to or pulled by other 709 nodes or NMS, a certain data correlation mechanism SHOULD be in use 710 to help the nodes or NMS to tell whether any two or more packet 711 counts are related to the same block of markers, or any two 712 timestamps are related to the same marked packet. 714 The alternate marking method described in this document literally 715 split the packets of the measured flow into different measurement 716 blocks, in addition a Block Number could be assigned to each of such 717 measurement block. The BN is generated each time a node reads the 718 data (packet counts or timestamps), and is associated with each 719 packet count and timestamp reported to or pulled by other nodes or 720 NMS. The value of BN could be calculated as the modulo of the local 721 time (when the data are read) and the interval of the marking time 722 period. 724 When the nodes or NMS see, for example, same BNs associated with two 725 packet counts from an upstream and a downstream node respectively, it 726 considers that these two packet counts corresponding to the same 727 block, i.e. that these two packet counts belong to the same block of 728 markers from the upstream and downstream node. The assumption of 729 this BN mechanism is that the measurement nodes are time 730 synchronized. This requires the measurement nodes to have a certain 731 time synchronization capability (e.g., the Network Time Protocol 732 (NTP) RFC 5905 [RFC5905], or the IEEE 1588 Precision Time Protocol 733 (PTP) [IEEE-1588]). Synchronization aspects are further discussed in 734 Section 4. 736 4.3. Packet Re-ordering 738 Due to ECMP, packet re-ordering is very common in IP network. The 739 accuracy of marking based PM, especially packet loss measurement, may 740 be affected by packet re-ordering. Take a look at the following 741 example: 743 Block : 1 | 2 | 3 | 4 | 5 |... 744 --------|---------|---------|---------|---------|---------|--- 745 Node R1 : AAAAAAA | BBBBBBB | AAAAAAA | BBBBBBB | AAAAAAA |... 746 Node R2 : AAAAABB | AABBBBA | AAABAAA | BBBBBBA | ABAAABA |... 748 Figure 5: Packet Reordering 750 In the following paragraphs an example of data correlation mechanism 751 is explained and could be use independently of the adopted solutions. 753 Most of the packet re-ordering occur at the edge of adjacent blocks, 754 and they are easy to handle if the interval of each block is 755 sufficient large. Then, it can assume that the packets with 756 different marker belong to the block that they are more close to. If 757 the interval is small, it is difficult and sometime impossible to 758 determine to which block a packet belongs. See above example, the 759 packet with the marker of "B" in block 3, there is no safe way to 760 tell whether the packet belongs to block 2 or block 4. 762 To choose a proper interval is important and how to choose a proper 763 interval is out of the scope of this document. But an implementation 764 SHOULD provide a way to configure the interval and allow a certain 765 degree of packet re-ordering. 767 5. Implementation and deployment 769 The methodology described in the previous sections can be applied in 770 various situations. Basically Alternate Marking technique could be 771 used in many cases for performance measurement. The only requirement 772 is to select and mark the flow to be monitored; in this way packets 773 are batched by the sender and each batch is alternately marked such 774 that can be easily recognized by the receiver. 776 An example of implementation and deployment is explained in the next 777 section, just to clarify how the method can work. 779 5.1. Report on the operational experiment at Telecom Italia 781 The method described in this document, also called PNPM (Packet 782 Network Performance Monitoring), has been invented and engineered in 783 Telecom Italia and it's currently being used in Telecom Italia's 784 network. The methodology has been applied by leveraging functions 785 and tools available on IP routers and it's currently being used to 786 monitor packet loss in some portions of Telecom Italia's network. 787 The application of the method to delay measurement is currently being 788 evaluated in Telecom Italia's labs. This section describes how the 789 features currently available on existing routing platforms can be 790 used to apply the method, in order to give an example of 791 implementation and deployment. 793 The fundamental steps for this implementation of the method can be 794 summarized in the following items: 796 o coloring the packets; 798 o counting the packets; 800 o collecting data and calculating the packet loss. 802 o metric transparency. 804 Before going deeper into the implementation details, it's worth 805 mentioning two different strategies that can be used when 806 implementing the method: 808 o flow-based: the flow-based strategy is used when only a limited 809 number of traffic flows need to be monitored. This could be the 810 case, for example, of IPTV channels or other specific applications 811 traffic with high QoS requirements (i.e. Mobile Backhauling 812 traffic). According to this strategy, only a subset of the flows 813 is colored. Counters for packet loss measurements can be 814 instantiated for each single flow, or for the set as a whole, 815 depending on the desired granularity. A relevant problem with 816 this approach is the necessity to know in advance the path 817 followed by flows that are subject to measurement. Path rerouting 818 and traffic load-balancing increase the issue complexity, 819 especially for unicast traffic. The problem is easier to solve 820 for multicast traffic where load balancing is seldom used, 821 especially for IPTV traffic where static joins are frequently used 822 to force traffic forwarding and replication. Another application 823 is on Mobile Backhauling, implemented with a VPN MPLS in Telecom 824 Italia's network; where the monitoring is between the Provider 825 Edge nodes of the VPN MPLS. 827 o link-based: measurements are performed on all the traffic on a 828 link by link basis. The link could be a physical link or a 829 logical link (for instance an Ethernet VLAN or a MPLS PW). 830 Counters could be instantiated for the traffic as a whole or for 831 each traffic class (in case it is desired to monitor each class 832 separately), but in the second case a couple of counters is needed 833 for each class. 835 The current implementation in Telecom Italia uses the first strategy. 836 As mentioned, the flow-based measurement requires the identification 837 of the flow to be monitored and the discovery of the path followed by 838 the selected flow. It is possible to monitor a single flow or 839 multiple flows grouped together, but in this case measurement is 840 consistent only if all the flows in the group follow the same path. 841 Moreover, a Service Provider should be aware that, if a measurement 842 is performed by grouping many flows, it is not possible to determine 843 exactly which flow was affected by packets loss. In order to have 844 measures per single flow it is necessary to configure counters for 845 each specific flow. Once the flow(s) to be monitored have been 846 identified, it is necessary to configure the monitoring on the proper 847 nodes. Configuring the monitoring means configuring the policy to 848 intercept the traffic and configuring the counters to count the 849 packets. To have just an end-to-end monitoring, it is sufficient to 850 enable the monitoring on the first and the last hop routers of the 851 path: the mechanism is completely transparent to intermediate nodes 852 and independent from the path followed by traffic flows. On the 853 contrary, to monitor the flow on a hop-by-hop basis along its whole 854 path it is necessary to enable the monitoring on every node from the 855 source to the destination. In case the exact path followed by the 856 flow is not known a priori (i.e. the flow has multiple paths to reach 857 the destination) it is necessary to enable the monitoring system on 858 every path: counters on interfaces traversed by the flow will report 859 packet count, counters on other interfaces will be null. 861 5.1.1. Coloring the packets 863 The coloring operation is fundamental in order to create packet 864 blocks. This implies choosing where to activate the coloring and how 865 to color the packets. 867 In case of flow-based measurements, it is desirable, in general, to 868 have a single coloring node because it is easier to manage and 869 doesn't rise any risk of conflict (consider the case where two nodes 870 color the same flow). Thus it is advantageous to color the flow as 871 close as possible to the source. In addition, coloring a flow close 872 to the source allows an end-to-end measure if a measurement point is 873 enabled on the last-hop router as well. The only requirement is that 874 the coloring must change periodically and every node along the path 875 must be able to identify unambiguously the colored packets. For 876 link-based measurements, all traffic needs to be colored when 877 transmitted on the link. If the traffic had already been colored, 878 then it has to be re-colored because the color must be consistent on 879 the link. This means that each hop along the path must (re-)color 880 the traffic; the color is not required to be consistent along 881 different links. 883 Traffic coloring can be implemented by setting a specific bit in the 884 packet header and changing the value of that bit periodically. With 885 current router implementations, only QoS related fields and features 886 offer the required flexibility to set bits in the packet header. In 887 case a Service Provider only uses the three most significant bits of 888 the DSCP field (corresponding to IP Precedence) for QoS 889 classification and queuing, it is possible to use the two less 890 significant bits of the DSCP field (bit 0 and bit 1) to implement the 891 method without affecting QoS policies. One of the two bits (bit 0) 892 could be used to identify flows subject to traffic monitoring (set to 893 1 if the flow is under monitoring, otherwise it is set to 0), while 894 the second (bit 1) can be used for coloring the traffic (switching 895 between values 0 and 1, corresponding to color A and B) and creating 896 the blocks. 898 In practice, coloring the traffic using the DSCP field can be 899 implemented by configuring on the router output interface an access 900 list that intercepts the flow(s) to be monitored and applies to them 901 a policy that sets the DSCP field accordingly. Since traffic 902 coloring has to be switched between the two values over time, the 903 policy needs to be modified periodically: an automatic script can be 904 used perform this task on the basis of a fixed timer. In Telecom 905 Italia's implementation this timer is set to 5 minutes: this value 906 showed to be a good compromise between measurement frequency and 907 stability of the measurement (i.e. possibility to collect all the 908 measures referring to the same block). 910 5.1.2. Counting the packets 912 Assuming that the coloring of the packets is performed only by the 913 source node, the nodes between source and destination (included) have 914 to count the colored packets that they receive and forward: this 915 operation can be enabled on every router along the path or only on a 916 subset, depending on which network segment is being monitored (a 917 single link, a particular metro area, the backbone, the whole path). 919 Since the color switches periodically between two values, two 920 counters (one for each value) are needed: one counter for packets 921 with color A and one counter for packets with color B. For each flow 922 (or group of flows) being monitored and for every interface where the 923 monitoring is active, a couple of counters is needed. For example, 924 in order to monitor separately 3 flows on a router with 4 interfaces 925 involved, 24 counters are needed (2 counters for each of the 3 flows 926 on each of the 4 interfaces). If traffic is colored using the DSCP 927 field, as in Telecom Italia's implementation, an access-list that 928 matches specific DSCP values can be used to count the packets of the 929 flow(s) being monitored. 931 In case of link-based measurements the behaviour is similar except 932 that coloring and counting operations are performed on a link by link 933 basis at each endpoint of the link. 935 Another important aspect to take into consideration is when to read 936 the counters: in order to count the exact number of packets of a 937 block the routers must perform this operation when that block has 938 ended: in other words, the counter for color A must be read when the 939 current block has color B, in order to be sure that the value of the 940 counter is stable. This task can be accomplished in two ways. The 941 general approach suggests to read the counters periodically, many 942 times during a block duration, and to compare these successive 943 readings: when the counter stops incrementing means that the current 944 block has ended and its value can be elaborated safely. 945 Alternatively, if the coloring operation is performed on the basis of 946 a fixed timer, it is possible to configure the reading of the 947 counters according to that timer: for example, if each block is 5 948 minutes long, reading the counter for color A every 5 minute in the 949 middle of the subsequent block (with color B) is a safe choice. A 950 sufficient margin should be considered between the end of a block and 951 the reading of the counter, in order to take into account any out-of- 952 order packets. The choice of a 5 minutes timer for colore switching 953 was also inspired by these considerations. 955 5.1.3. Collecting data and calculating packet loss 957 The nodes enabled to perform performance monitoring collect the value 958 of the counters, but they are not able to directly use this 959 information to measure packet loss, because they only have their own 960 samples. For this reason, an external Network Management System 961 (NMS) is required to collect and elaborate data and to perform packet 962 loss calculation. The NMS compares the values of counters from 963 different nodes and can calculate if some packets were lost (even a 964 single packet) and also where packets were lost. 966 The value of the counters needs to be transmitted to the NMS as soon 967 as it has been read. This can be accomplished by using SNMP or FTP 968 and can be done in Push Mode or Polling Mode. In the first case, 969 each router periodically sends the information to the NMS, in the 970 latter case it is the NMS that periodically polls routers to collect 971 information. In any case, the NMS has to collect all the relevant 972 values from all the routers within one cycle of the timer (5 973 minutes). 975 If link-based measurement is used, it would be possible to use a 976 protocol to exchange values of counters between the two endpoints in 977 order to let them perform the packet loss calculation for each 978 traffic direction. A similar approach could be complicated if 979 applied to a flow-based measurement. 981 5.1.4. Metric transparency 983 In Telecom Italia's implementation the source node colors the packets 984 with a policy that is modified periodically via an automatic script 985 in order to alternate the DSCP field of the packets. The nodes 986 between source and destination (included) have to count with an 987 access-list the colored packets that they receive and forward. 989 Moreover the destination node has an important role: the colored 990 packets are intercepted and a policy restores and sets the DSCP field 991 of all the packets to the initial value. In this way the metric is 992 transparent because outside the section of the network under 993 monitoring the traffic flow is unchanged. 995 In such a case, thanks to this restoring technique, network elements 996 outside the Alternate Marking monitoring domain (e.g. the two 997 Provider Edge nodes of the Mobile Backhauling VPN MPLS) are totally 998 anaware that packets were marked. So this restoring technique makes 999 Alternate Marking completely transparent outside its monitoring 1000 domain. 1002 5.2. IP flow performance measurement (IPFPM) 1004 This application of marking method is described in 1005 [I-D.chen-ippm-coloring-based-ipfpm-framework]. 1007 5.3. OAM Passive Performance Measurement 1009 In [I-D.ietf-bier-mpls-encapsulation] two OAM bits from Bit Index 1010 Explicit Replication (BIER) Header are reserved for the passive 1011 performance measurement marking method. [I-D.ietf-bier-pmmm-oam] 1012 details the measurement for multicast service over BIER domain. 1014 In addition, the alternate marking method could also be used in a 1015 Service Function Chaining (SFC) domain. 1017 The application of the marking method to Network Virtualization 1018 Overlays (NVO3) protocols is a work in progress (see 1019 [I-D.ietf-nvo3-encap]). 1021 5.4. RFC6374 Use Case 1023 RFC6374 [RFC6374] uses the LM packet as the packet accounting 1024 demarcation point. Unfortunately this gives rise to a number of 1025 problems that may lead to significant packet accounting errors in 1026 certain situations. [I-D.ietf-mpls-flow-ident] discusses the desired 1027 capabilities for MPLS flow identification in order to perform a 1028 better in-band performance monitoring of user data packets. A method 1029 of accomplishing identification is Synonymous Flow Labels (SFL) 1030 introduced in [I-D.bryant-mpls-sfl-framework], while 1031 [I-D.ietf-mpls-rfc6374-sfl] describes RFC6374 performance 1032 measurements with SFL. 1034 5.5. Application to active performance measurement 1036 [I-D.fioccola-ippm-alt-mark-active] describes how to extend the 1037 existing Active Measurement Protocol, in order to implement alternate 1038 marking methodology. [I-D.fioccola-ippm-rfc6812-alt-mark-ext] 1039 describes an extension to the Cisco SLA Protocol Measurement-Type 1040 UDP-Measurement. 1042 6. Hybrid measurement 1044 The method has been explicitly designed for passive measurements but 1045 it can also be used with active measurements. In order to have both 1046 end to end measurements and intermediate measurements (hybrid 1047 measurements) two end points can exchanges artificial traffic flows 1048 and apply alternate marking over these flows. In the intermediate 1049 points artificial traffic is managed in the same way as real traffic 1050 and measured as specified before. So the application of marking 1051 method can simplify also the active measurement, as explained in 1052 [I-D.fioccola-ippm-alt-mark-active]. 1054 7. Compliance with RFC6390 guidelines 1056 RFC6390 [RFC6390] defines a framework and a process for developing 1057 Performance Metrics for protocols above and below the IP layer (such 1058 as IP-based applications that operate over reliable or datagram 1059 transport protocols). 1061 This document doesn't aim to propose a new Performance Metric but a 1062 new method of measurement for a few Performance Metrics that have 1063 already been standardized. Nevertheless, it's worth applying 1064 [RFC6390] guidelines to the present document, in order to provide a 1065 more complete and coherent description of the proposed method. We 1066 used a subset of the Performance Metric Definition template defined 1067 by [RFC6390]. 1069 o Metric name and description: as already stated, this document 1070 doesn't propose any new Performance Metric. On the contrary, it 1071 describes a novel method for measuring packet loss [RFC7680]. The 1072 same concept, with small differences, can also be used to measure 1073 delay [RFC7679], and jitter [RFC3393]. The document mainly 1074 describes the applicability to packet loss measurement. 1076 o Method of Measurement or Calculation: according to the method 1077 described in the previous sections, the number of packets lost is 1078 calculated by subtracting the value of the counter on the source 1079 node from the value of the counter on the destination node. Both 1080 counters must refer to the same color. The calculation is 1081 performed when the value of the counters is in a steady state. 1083 o Units of Measurement: the method calculates and reports the exact 1084 number of packets sent by the source node and not received by the 1085 destination node. 1087 o Measurement Points: the measurement can be performed between 1088 adjacent nodes, on a per-link basis, or along a multi-hop path, 1089 provided that the traffic under measurement follows that path. In 1090 case of a multi-hop path, the measurements can be performed both 1091 end-to-end and hop-by-hop. 1093 o Measurement Timing: the method have a constraint on the frequency 1094 of measurements. In order to perform a measure, the counter must 1095 be in a steady state: this happens when the traffic is being 1096 colored with the alternate color; for example in the Telecom 1097 Italia application of the method the time interval is set to 5 1098 minutes. 1100 o Implementation: the Telecom Italia application of the method uses 1101 two encodings of the DSCP field to color the packets; this enables 1102 the use of policy configurations on the router to color the 1103 packets and accordingly configure the counter for each color. The 1104 path followed by traffic being measured should be known in advance 1105 in order to configure the counters along the path and be able to 1106 compare the correct values. 1108 o Use and Applications: the method can be used to measure packet 1109 loss with high precision on live traffic; moreover, by combining 1110 end-to-end and per-link measurements, the method is useful to 1111 pinpoint the single link that is experiencing loss events. 1113 o Reporting Model: the value of the counters has to be sent to a 1114 centralized management system that perform the calculations; such 1115 samples must contain a reference to the time interval they refer 1116 to, so that the management system can perform the correct 1117 correlation; the samples have to be sent while the corresponding 1118 counter is in a steady state (within a time interval), otherwise 1119 the value of the sample should be stored locally. 1121 o Dependencies: the values of the counters have to be correlated to 1122 the time interval they refer to; moreover, as far the Telecom 1123 Italia application of the method is based on DSCP values, there 1124 are significant dependencies on the usage of the DSCP field: it 1125 must be possible to rely on unused DSCP values without affecting 1126 QoS-related configuration and behavior; moreover, the intermediate 1127 nodes must not change the value of the DSCP field not to alter the 1128 measurement. 1130 o Organization of Results: the method of measurement produces 1131 singletons. 1133 o Parameters: currently, the main parameter of the method is the 1134 time interval used to alternate the colors and read the counters. 1136 8. Security Considerations 1138 This document specifies a method to perform measurements in the 1139 context of a Service Provider's network and has not been developed to 1140 conduct Internet measurements, so it does not directly affect 1141 Internet security nor applications which run on the Internet. 1142 However, implementation of this method must be mindful of security 1143 and privacy concerns. 1145 There are two types of security concerns: potential harm caused by 1146 the measurements and potential harm to the measurements. For what 1147 concerns the first point, the measurements described in this document 1148 are passive, so there are no packets injected into the network 1149 causing potential harm to the network itself and to data traffic. 1150 Nevertheless, the method implies modifications on the fly to the IP 1151 header of data packets: this must be performed in a way that doesn't 1152 alter the quality of service experienced by packets subject to 1153 measurements and that preserve stability and performance of routers 1154 doing the measurements. The measurements themselves could be harmed 1155 by routers altering the marking of the packets, or by an attacker 1156 injecting artificial traffic. Authentication techniques, such as 1157 digital signatures, may be used where appropriate to guard against 1158 injected traffic attacks. 1160 The privacy concerns of network measurement are limited because the 1161 method only relies on information contained in the IP header without 1162 any release of user data. 1164 The measurement itself may be affected by routers (or other network 1165 devices) along the path of IP packets intentionally altering the 1166 value of marking bits of packets. As mentioned above, the mechanism 1167 specified in this document is just in the context of one Service 1168 Provider's network, and thus the routers (or other network devices) 1169 are locally administered and this type of attack can be avoided. 1171 One of the main security threats in OAM protocols is network 1172 reconnaissance; an attacker can gather information about the network 1173 performance by passively eavesdropping to OAM messages. The 1174 advantage of the methods described in this document is that the 1175 marking bits are the only information that is exchanged between the 1176 network devices. Therefore, passive eavesdropping to data plane 1177 traffic does not allow attackers to gain information about the 1178 network performance. 1180 Delay attacks are another potential threat in the context of this 1181 document. Delay measurement is performed using a specific packet in 1182 each block, marked by a dedicated color bit. Therefore, a man-in- 1183 the-middle attacker can selectively induce synthetic delay only to 1184 delay-colored packets, causing systematic error in the delay 1185 measurements. As discussed in previous sections, the methods 1186 described in this document rely on an underlying time synchronization 1187 protocol. Thus, by attacking the time protocol an attacker can 1188 potentially compromise the integrity of the measurement. A detailed 1189 discussion about the threats against time protocols and how to 1190 mitigate them is presented in RFC 7384 [RFC7384]. 1192 9. Conclusions 1194 The advantages of the method described in this document are: 1196 o easy implementation: it can be implemented using features already 1197 available on major routing platforms; 1199 o low computational effort: the additional load on processing is 1200 negligible; 1202 o accurate packet loss measurement: single packet loss granularity 1203 is achieved with a passive measurement; 1205 o potential applicability to any kind of packet/frame -based 1206 traffic: Ethernet, IP, MPLS, etc., both unicast and multicast; 1208 o robustness: the method can tolerate out of order packets and it's 1209 not based on "special" packets whose loss could have a negative 1210 impact; 1212 o no interoperability issues: the features required to implement the 1213 method are available on all current routing platforms. 1215 The method doesn't raise any specific need for protocol extension, 1216 but it could be further improved by means of some extension to 1217 existing protocols. Specifically, the use of DiffServ bits for 1218 coloring the packets could not be a viable solution in some cases: a 1219 standard method to color the packets for this specific application 1220 could be beneficial. 1222 10. IANA Considerations 1224 There are no IANA actions required. 1226 11. Acknowledgements 1228 The previous IETF drafts about this technique were: 1229 [I-D.cociglio-mboned-multicast-pm] and [I-D.tempia-opsawg-p3m]. 1230 There are some references to this methodology in other IETF works 1231 (e.g. [I-D.ietf-mpls-flow-ident], [I-D.bryant-mpls-sfl-framework] 1232 [I-D.ietf-mpls-rfc6374-sfl], [I-D.ietf-bier-mpls-encapsulation], 1233 [I-D.ietf-bier-pmmm-oam], [I-D.mirsky-sfc-pmamm], 1234 [I-D.chen-ippm-coloring-based-ipfpm-framework]). 1236 In addition the authors would like to thank Alberto Tempia Bonda, 1237 Domenico Laforgia, Daniele Accetta and Mario Bianchetti for their 1238 contribution to the definition and the implementation of the method. 1240 12. References 1242 12.1. Normative References 1244 [IEEE-1588] 1245 IEEE 1588-2008, "IEEE Standard for a Precision Clock 1246 Synchronization Protocol for Networked Measurement and 1247 Control Systems", July 2008. 1249 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1250 Requirement Levels", BCP 14, RFC 2119, 1251 DOI 10.17487/RFC2119, March 1997, 1252 . 1254 [RFC3393] Demichelis, C. and P. Chimento, "IP Packet Delay Variation 1255 Metric for IP Performance Metrics (IPPM)", RFC 3393, 1256 DOI 10.17487/RFC3393, November 2002, 1257 . 1259 [RFC5905] Mills, D., Martin, J., Ed., Burbank, J., and W. Kasch, 1260 "Network Time Protocol Version 4: Protocol and Algorithms 1261 Specification", RFC 5905, DOI 10.17487/RFC5905, June 2010, 1262 . 1264 [RFC7679] Almes, G., Kalidindi, S., Zekauskas, M., and A. Morton, 1265 Ed., "A One-Way Delay Metric for IP Performance Metrics 1266 (IPPM)", STD 81, RFC 7679, DOI 10.17487/RFC7679, January 1267 2016, . 1269 [RFC7680] Almes, G., Kalidindi, S., Zekauskas, M., and A. Morton, 1270 Ed., "A One-Way Loss Metric for IP Performance Metrics 1271 (IPPM)", STD 82, RFC 7680, DOI 10.17487/RFC7680, January 1272 2016, . 1274 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1275 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1276 May 2017, . 1278 12.2. Informative References 1280 [I-D.bryant-mpls-sfl-framework] 1281 Bryant, S., Chen, M., Li, Z., Swallow, G., Sivabalan, S., 1282 and G. Mirsky, "Synonymous Flow Label Framework", draft- 1283 bryant-mpls-sfl-framework-05 (work in progress), June 1284 2017. 1286 [I-D.chen-ippm-coloring-based-ipfpm-framework] 1287 Chen, M., Zheng, L., Mirsky, G., Fioccola, G., and T. 1288 Mizrahi, "IP Flow Performance Measurement Framework", 1289 draft-chen-ippm-coloring-based-ipfpm-framework-06 (work in 1290 progress), March 2016. 1292 [I-D.cociglio-mboned-multicast-pm] 1293 Cociglio, M., Capello, A., Bonda, A., and L. Castaldelli, 1294 "A method for IP multicast performance monitoring", draft- 1295 cociglio-mboned-multicast-pm-01 (work in progress), 1296 October 2010. 1298 [I-D.fioccola-ippm-alt-mark-active] 1299 Fioccola, G., Clemm, A., Bryant, S., Cociglio, M., 1300 Chandramouli, M., and A. Capello, "Alternate Marking 1301 Extension to Active Measurement Protocol", draft-fioccola- 1302 ippm-alt-mark-active-01 (work in progress), March 2017. 1304 [I-D.fioccola-ippm-rfc6812-alt-mark-ext] 1305 Fioccola, G., Clemm, A., Cociglio, M., Chandramouli, M., 1306 and A. Capello, "Alternate Marking Extension to Cisco SLA 1307 Protocol RFC6812", draft-fioccola-ippm-rfc6812-alt-mark- 1308 ext-01 (work in progress), March 2016. 1310 [I-D.ietf-bier-mpls-encapsulation] 1311 Wijnands, I., Rosen, E., Dolganow, A., Tantsura, J., 1312 Aldrin, S., and I. Meilik, "Encapsulation for Bit Index 1313 Explicit Replication in MPLS and non-MPLS Networks", 1314 draft-ietf-bier-mpls-encapsulation-07 (work in progress), 1315 June 2017. 1317 [I-D.ietf-bier-pmmm-oam] 1318 Mirsky, G., Zheng, L., Chen, M., and G. Fioccola, 1319 "Performance Measurement (PM) with Marking Method in Bit 1320 Index Explicit Replication (BIER) Layer", draft-ietf-bier- 1321 pmmm-oam-02 (work in progress), July 2017. 1323 [I-D.ietf-mpls-flow-ident] 1324 Bryant, S., Pignataro, C., Chen, M., Li, Z., and G. 1325 Mirsky, "MPLS Flow Identification Considerations", draft- 1326 ietf-mpls-flow-ident-04 (work in progress), February 2017. 1328 [I-D.ietf-mpls-rfc6374-sfl] 1329 Bryant, S., Chen, M., Li, Z., Swallow, G., Sivabalan, S., 1330 Mirsky, G., and G. Fioccola, "RFC6374 Synonymous Flow 1331 Labels", draft-ietf-mpls-rfc6374-sfl-00 (work in 1332 progress), June 2017. 1334 [I-D.ietf-nvo3-encap] 1335 Boutros, S., Ganga, I., Garg, P., Manur, R., Mizrahi, T., 1336 Mozes, D., and E. Nordmark, "NVO3 Encapsulation 1337 Considerations", draft-ietf-nvo3-encap-00 (work in 1338 progress), June 2017. 1340 [I-D.mirsky-sfc-pmamm] 1341 Mirsky, G., Fioccola, G., and T. Mizrahi, "Performance 1342 Measurement (PM) with Alternate Marking Method in Service 1343 Function Chaining (SFC) Domain", draft-mirsky-sfc-pmamm-01 1344 (work in progress), June 2017. 1346 [I-D.tempia-opsawg-p3m] 1347 Capello, A., Cociglio, M., Castaldelli, L., and A. Bonda, 1348 "A packet based method for passive performance 1349 monitoring", draft-tempia-opsawg-p3m-04 (work in 1350 progress), February 2014. 1352 [RFC5481] Morton, A. and B. Claise, "Packet Delay Variation 1353 Applicability Statement", RFC 5481, DOI 10.17487/RFC5481, 1354 March 2009, . 1356 [RFC6374] Frost, D. and S. Bryant, "Packet Loss and Delay 1357 Measurement for MPLS Networks", RFC 6374, 1358 DOI 10.17487/RFC6374, September 2011, 1359 . 1361 [RFC6390] Clark, A. and B. Claise, "Guidelines for Considering New 1362 Performance Metric Development", BCP 170, RFC 6390, 1363 DOI 10.17487/RFC6390, October 2011, 1364 . 1366 [RFC6703] Morton, A., Ramachandran, G., and G. Maguluri, "Reporting 1367 IP Network Performance Metrics: Different Points of View", 1368 RFC 6703, DOI 10.17487/RFC6703, August 2012, 1369 . 1371 [RFC7276] Mizrahi, T., Sprecher, N., Bellagamba, E., and Y. 1372 Weingarten, "An Overview of Operations, Administration, 1373 and Maintenance (OAM) Tools", RFC 7276, 1374 DOI 10.17487/RFC7276, June 2014, 1375 . 1377 [RFC7384] Mizrahi, T., "Security Requirements of Time Protocols in 1378 Packet Switched Networks", RFC 7384, DOI 10.17487/RFC7384, 1379 October 2014, . 1381 [RFC7799] Morton, A., "Active and Passive Metrics and Methods (with 1382 Hybrid Types In-Between)", RFC 7799, DOI 10.17487/RFC7799, 1383 May 2016, . 1385 Authors' Addresses 1387 Giuseppe Fioccola (editor) 1388 Telecom Italia 1389 Via Reiss Romoli, 274 1390 Torino 10148 1391 Italy 1393 Email: giuseppe.fioccola@telecomitalia.it 1395 Alessandro Capello (editor) 1396 Telecom Italia 1397 Via Reiss Romoli, 274 1398 Torino 10148 1399 Italy 1401 Email: alessandro.capello@telecomitalia.it 1402 Mauro Cociglio 1403 Telecom Italia 1404 Via Reiss Romoli, 274 1405 Torino 10148 1406 Italy 1408 Email: mauro.cociglio@telecomitalia.it 1410 Luca Castaldelli 1411 Telecom Italia 1412 Via Reiss Romoli, 274 1413 Torino 10148 1414 Italy 1416 Email: luca.castaldelli@telecomitalia.it 1418 Mach(Guoyi) Chen (editor) 1419 Huawei Technologies 1421 Email: mach.chen@huawei.com 1423 Lianshu Zheng (editor) 1424 Huawei Technologies 1426 Email: vero.zheng@huawei.com 1428 Greg Mirsky (editor) 1429 ZTE 1430 USA 1432 Email: gregimirsky@gmail.com 1434 Tal Mizrahi (editor) 1435 Marvell 1436 6 Hamada st. 1437 Yokneam 1438 Israel 1440 Email: talmi@marvell.com