idnits 2.17.00 (12 Aug 2021) /tmp/idnits60941/draft-ietf-v6ops-ipv6-ehs-packet-drops-01.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 == Line 183 has weird spacing: '... length var...' -- The document date (October 14, 2020) is 584 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-10) exists of draft-ietf-opsec-ipv6-eh-filtering-06 -- Obsolete informational reference (is this intentional?): RFC 2460 (Obsoleted by RFC 8200) -- Obsolete informational reference (is this intentional?): RFC 5575 (Obsoleted by RFC 8955) Summary: 0 errors (**), 0 flaws (~~), 3 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 IPv6 Operations Working Group (v6ops) F. Gont 3 Internet-Draft SI6 Networks 4 Intended status: Informational N. Hilliard 5 Expires: April 17, 2021 INEX 6 G. Doering 7 SpaceNet AG 8 W. Kumari 9 Google 10 G. Huston 11 APNIC 12 W. Liu 13 Huawei Technologies 14 October 14, 2020 16 Operational Implications of IPv6 Packets with Extension Headers 17 draft-ietf-v6ops-ipv6-ehs-packet-drops-01 19 Abstract 21 This document summarizes the operational implications of IPv6 22 extension headers, and attempts to analyze reasons why packets with 23 IPv6 extension headers may be dropped in the public Internet. 25 Status of This Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at https://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on April 17, 2021. 42 Copyright Notice 44 Copyright (c) 2020 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (https://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 60 2. Disclaimer . . . . . . . . . . . . . . . . . . . . . . . . . 3 61 3. Background Information . . . . . . . . . . . . . . . . . . . 3 62 4. Previous Work on IPv6 Extension Headers . . . . . . . . . . . 5 63 5. Packet Forwarding Engine Constraints . . . . . . . . . . . . 7 64 5.1. Recirculation . . . . . . . . . . . . . . . . . . . . . . 8 65 6. Requirement to Process Layer-3/layer-4 information in 66 Intermediate Systems . . . . . . . . . . . . . . . . . . . . 8 67 6.1. ECMP and Hash-based Load-Sharing . . . . . . . . . . . . 8 68 6.2. Enforcing infrastructure ACLs . . . . . . . . . . . . . . 9 69 6.3. DDoS Management and Customer Requests for Filtering . . . 9 70 6.4. Network Intrusion Detection and Prevention . . . . . . . 10 71 6.5. Firewalling . . . . . . . . . . . . . . . . . . . . . . . 10 72 7. Operational Implications . . . . . . . . . . . . . . . . . . 11 73 7.1. Inability to Find Layer-4 Information . . . . . . . . . . 11 74 7.2. Route-Processor Protection . . . . . . . . . . . . . . . 11 75 7.3. Inability to Perform Fine-grained Filtering . . . . . . . 12 76 7.4. Security Concerns Associated with IPv6 Extension Headers 12 77 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13 78 9. Security Considerations . . . . . . . . . . . . . . . . . . . 13 79 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 13 80 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 14 81 11.1. Normative References . . . . . . . . . . . . . . . . . . 14 82 11.2. Informative References . . . . . . . . . . . . . . . . . 15 83 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 19 85 1. Introduction 87 IPv6 Extension Headers (EHs) allow for the extension of the IPv6 88 protocol, and provide support for core functionality such as IPv6 89 fragmentation. However, common implementation limitations suggest 90 that EHs present a challenge for IPv6 packet routing equipment and 91 middle-boxes, and evidence exists that IPv6 packets with EHs may be 92 intentionally dropped in the public Internet in some network 93 deployments. 95 The authors of this document have been involved in numerous 96 discussions about IPv6 extension headers (both within the IETF and in 97 other fora), and have noticed that the security and operational 98 implications associated with IPv6 EHs were unknown to the larger 99 audience participating in these discussions. 101 This document has the following goals: 103 o Raise awareness about the operational and security implications of 104 IPv6 Extension Headers, and presents reasons why some networks may 105 intentionally drop packets containing IPv6 Extension Headers. 107 o Highlight areas where current IPv6 support by networking devices 108 maybe sub-optimal, such that the aforementioned support is 109 improved. 111 o Highlight operational issues associated with IPv6 extension 112 headers, such that those issues are considered in IETF 113 standardization efforts. 115 Section 3 provides background information about the IPv6 packet 116 structure and associated implications. Section 4 of this document 117 summarizes the previous work that has been carried out in the area of 118 IPv6 extension headers. Section 5 discusses packet forwarding engine 119 constraints in contemporary routers. Section 6 discusses why 120 contemporary routers and middle-boxes may need to access Layer-4 121 information to make a forwarding decision. Finally, Section 7 122 discusses the operational implications of IPv6 EHs. 124 2. Disclaimer 126 This document analyzes the operational challenges represented by 127 packets that employ IPv6 Extension Headers, and documents some of the 128 operational reasons why these packets may be dropped in the public 129 Internet. This document is not a recommendation to drop such 130 packets, but rather an analysis of why they are dropped. 132 3. Background Information 134 It is useful to compare the basic structure of IPv6 packets against 135 that of IPv4 packets, and analyze the implications of the two 136 different packet structures. 138 IPv4 packets have a variable-length header size, that allows for the 139 use of IPv4 "options" -- optional information that may be of use by 140 nodes processing IPv4 packets. The IPv4 header length is specified 141 in the IHL header field of the mandatory IPv4 header, and must be in 142 the range from 20 octets (the minimum IPv4 header size) to 60 octets 143 (accommodating at most 40 octets of options). The upper-layer 144 protocol type is specified via the "Protocol" field of the mandatory 145 IPv4 header. 147 Protocol, IHL 148 +--------+ 149 | | 150 | v 151 +------//-----+------------------------+ 152 | | | 153 | IPv4 | Upper-Layer | 154 | Header | Protocol | 155 | | | 156 +-----//------+------------------------+ 158 variable length 159 <-------------> 161 Figure 1: IPv4 Packet Structure 163 IPv6 took a different approach to the IPv6 packet structure. Rather 164 than employing a variable-length header as IPv4 does, IPv6 employs a 165 linked-list-like packet structure, where a mandatory fixed-length 166 IPv6 header is followed by an arbitrary number of optional extension 167 headers, with the upper-layer header being the last header in the 168 IPv6 header chain. Each extension header typically specifies its 169 length (unless it is implicit from the extension header type), and 170 the "next header" type that follows in the IPv6 IPv6 header chain. 172 NH NH, EH-length NH, EH-length 173 +-------+ +------+ +-------+ 174 | | | | | | 175 | v | v | v 176 +-------------+-------------+-//-+---------------+--------------+ 177 | | | | | | 178 | IPv6 | Ext. | | Ext. | Upper-Layer | 179 | header | Header | | Header | Protocol | 180 | | | | | | 181 +-------------+-------------+-//-+---------------+--------------+ 183 fixed length variable number of EHs & length 184 <------------> <--------------------------------> 186 Figure 2: IPv6 Packet Structure 188 This packet structure has the following implications: 190 o [RFC8200] requires the entire IPv6 header chain to be contained in 191 the first fragment of a packet, therefore limiting the IPv6 192 extension header chain to the size of the Path-MTU. 194 o Other than the Path-MTU constraints, there are no other limits to 195 the number of IPv6 EHs that may be present in a packet. 196 Therefore, there is no upper-limit regarding "how deep into the 197 IPv6 packet" the upper-layer may be found. 199 o The only way for a node to obtain the upper-layer protocol type or 200 find the upper-layer protocol header is to parse and process the 201 entire IPv6 header chain, in sequence, starting from the mandatory 202 IPv6 header, until the last header in the IPv6 header chain is 203 found. 205 4. Previous Work on IPv6 Extension Headers 207 Some of the operational implications of IPv6 Extension Headers have 208 been discussed in IETF circles: 210 o [I-D.taylor-v6ops-fragdrop] discusses a rationale for which 211 operators drop IPv6 fragments. 213 o [I-D.wkumari-long-headers] discusses possible issues arising from 214 "long" IPv6 header chains. 216 o [I-D.kampanakis-6man-ipv6-eh-parsing] describes how 217 inconsistencies in the way IPv6 packets with extension headers are 218 parsed by different implementations may result in evasion of 219 security controls, and presents guidelines for parsing IPv6 220 extension headers with the goal of providing a common and 221 consistent parsing methodology for IPv6 implementations. 223 o [I-D.ietf-opsec-ipv6-eh-filtering] analyzes the security 224 implications of IPv6 EHs, and the operational implications of 225 dropping packets that employ IPv6 EHs and associated options. 227 o [RFC7113] discusses how some popular RA-Guard implementations are 228 subject to evasion by means of IPv6 extension headers. 230 o [RFC8900] analyzes the fragility introduced by IP fragmentation. 232 A number of recent RFCs have discussed issues related to IPv6 233 extension headers, specifying updates to a previous revision of the 234 IPv6 standard ([RFC2460]), many of which have now been incorporated 235 into the current IPv6 core standard ([RFC8200]) or the IPv6 Node 236 Requirements ([RFC8504]). Namely, 237 o [RFC5095] discusses the security implications of Routing Header 238 Type 0 (RTH0), and deprecates it. 240 o [RFC5722] analyzes the security implications of overlapping 241 fragments, and provides recommendations in this area. 243 o [RFC7045] clarifies how intermediate nodes should deal with IPv6 244 extension headers. 246 o [RFC7112] discusses the issues arising in a specific fragmentation 247 case where the IPv6 header chain is fragmented into two or more 248 fragments (and formally forbids such fragmentation case). 250 o [RFC6946] discusses a flawed (but common) processing of the so- 251 called IPv6 "atomic fragments", and specified improved processing 252 of such packets. 254 o [RFC8021] deprecates the generation of IPv6 atomic fragments. 256 o [RFC8504] clarifies processing rules for packets with extension 257 headers, and also allows hosts to enforce limits on the number of 258 options included in IPv6 EHs. 260 o [RFC7739] discusses the security implications of predictable 261 fragment Identification values, and provides recommendations for 262 the generation of these values. 264 o [RFC6980] analyzes the security implications of employing IPv6 265 fragmentation with Neighbor Discovery for IPv6, and formally 266 recommends against such usage. 268 Additionally, [RFC8200] has relaxed the requirement that "all nodes 269 examine and process the Hop-by-Hop Options header" from [RFC2460], by 270 specifying that only nodes that have been explicitly configured to 271 process the Hop-by-Hop Options header are required to do so. 273 A number of studies have measured the extent to which packets 274 employing IPv6 extension headers are dropped in the public Internet: 276 o [PMTUD-Blackholes], [Gont-IEPG88], [Gont-Chown-IEPG89], and 277 [Linkova-Gont-IEPG90] presented some preliminary measurements 278 regarding the extent to which packet containing IPv6 EHs are 279 dropped in the public Internet. 281 o [RFC7872] presents more comprehensive results and documents the 282 methodology for obtaining the presented results. 284 o [Huston-2017] and [Huston-2020] measured packet drops resulting 285 from IPv6 fragmentation when communicating with DNS servers. 287 5. Packet Forwarding Engine Constraints 289 Most contemporary routers use dedicated hardware (e.g. ASICs or 290 NPUs) to determine how to forward packets across their internal 291 fabrics (see [IEPG94-Scudder] and [APNIC-Scudder] for details). One 292 of the common methods of handling next-hop lookup is to send a small 293 portion of the ingress packet to a lookup engine with specialised 294 hardware (e.g. ternary CAM or RLDRAM) to determine the packet's next- 295 hop. Technical constraints mean that there is a trade-off between 296 the amount of data sent to the lookup engine and the overall 297 performance of the lookup engine. If more data is sent, the lookup 298 engine can inspect further into the packet, but the overall 299 performance of the system will be reduced. If less data is sent, the 300 overall performance of the router will be increased but the packet 301 lookup engine may not be able to inspect far enough into a packet to 302 determine how it should be handled. 304 NOTE: 305 For example, contemporary high-end routers can use up to 192 bytes 306 of header (Cisco ASR9000 Typhoon) or 384 bytes of header (Juniper 307 MX Trio). 309 If a hardware forwarding engine on a contemporary router cannot make 310 a forwarding decision about a packet because critical information is 311 not sent to the look-up engine, then the router will normally drop 312 the packet. 314 NOTE: 315 Section 6 discusses some of the reasons for which a contemporary 316 router might need to access layer-4 information to make a 317 forwarding decision. 319 Historically, some packet forwarding engines punted packets of this 320 form to the control plane for more in-depth analysis, but this is 321 unfeasible on most current router architectures as a result of the 322 vast difference between the hardware forwarding capacity of the 323 router and processing capacity of the control plane and the size of 324 the management link which connects the control plane to the 325 forwarding plane. 327 If an IPv6 header chain is sufficiently long that its header exceeds 328 the packet look-up capacity of the router, then it may be dropped due 329 to hardware inability to determine how it should be handled. 331 5.1. Recirculation 333 Although TLV chains are amenable to iterative processing on 334 architectures which have packet look-up engines with deep inspection 335 capabilities, some packet forwarding engines manage IPv6 Extension 336 Header chains using recirculation. This approach processes Extension 337 Headers one at a time: when processing on one Extension Header is 338 completed, the packet is looped back through the processing engine 339 again. This recirculation process continues repeatedly until there 340 are no more Extension Headers left to be processed. 342 Recirculation is typically used on packet forwarding engines with 343 limited look-up capability, as it allows arbitrarily long header 344 chains to be processed without the complexity and cost associated 345 with packet forwarding engines which have deep look-up capabilities. 346 However, recirculation can impact the forwarding capacity of 347 hardware, as each packet will pass through the processing engine 348 multiple times. Depending on configuration, the type of packets 349 being processed, and the hardware capabilities of the packet 350 forwarding engine, this may impact data-plane throughput performance 351 on the router. 353 6. Requirement to Process Layer-3/layer-4 information in Intermediate 354 Systems 356 The following subsections discuss some of reasons for which 357 contemporary routers and middle-boxes may need to process Layer-3/ 358 layer-4 information to make a forwarding decision. 360 6.1. ECMP and Hash-based Load-Sharing 362 In the case of ECMP (equal cost multi path) load sharing, the router 363 on the sending side of the link needs to make a decision regarding 364 which of the links to use for a given packet. Since round-robin 365 usage of the links is usually avoided in order to prevent packet 366 reordering, forwarding engines need to use a mechanism which will 367 consistently forward the same data streams down the same forwarding 368 paths. Most forwarding engines achieve this by calculating a simple 369 hash using an n-tuple gleaned from a combination of layer-2 through 370 to layer-4 packet header information. This n-tuple will typically 371 use the src/dst MAC address, src/dst IP address, and if possible 372 further layer-4 src/dst port information. As layer-4 port 373 information increases the entropy of the hash, it is normally highly 374 desirable to use it where possible. 376 We note that in the IPv6 world, flows are expected to be identified 377 by means of the IPv6 Flow Label [RFC6437]. Thus, ECMP and Hash-based 378 Load-Sharing would be possible without the need to process the entire 379 IPv6 header chain to obtain upper-layer information to identify 380 flows. However, we note that for a long time many IPv6 381 implementations failed to set the Flow Label, and ECMP and Hash-based 382 Load-Sharing devices also did not employ the Flow Label for 383 performing their task. 385 Clearly, widespread support of [RFC6437] would relieve middle-boxes 386 from having to process the entire IPv6 header chain, making Flow 387 Label-based ECMP and Hash-based Load-Sharing [RFC6438] feasible. 389 While support of [RFC6437] is currently widespread for current 390 versions of all popular host implementations, there is still only 391 marginal usage of the IPv6 Flow Label for ECMP and load balancing 392 [Cunha-2020] -- possibly as a result of issues that have been found 393 in host implementations and middle-boxes [Jaeggli-2018]. 395 6.2. Enforcing infrastructure ACLs 397 Generally speaking, infrastructure ACLs (iACLs) drop unwanted packets 398 destined to parts of a provider's infrastructure, because they are 399 not operationally needed and can be used for attacks of different 400 sorts against router control planes. Some traffic needs to be 401 differentiated depending on layer-3 or layer-4 criteria to achieve a 402 useful balance of protection and functionality, for example: 404 o Permit some amount of ICMP echo (ping) traffic towards a router's 405 addresses for troubleshooting. 407 o Permit BGP sessions on the shared network of an exchange point 408 (potentially differentiating between the amount of packets/seconds 409 permitted for established sessions and connection establishment), 410 but do not permit other traffic from the same peer IP addresses. 412 6.3. DDoS Management and Customer Requests for Filtering 414 The case of customer DDoS protection and edge-to-core customer 415 protection filters is similar in nature to the infrastructure ACL 416 protection. Similar to infrastructure ACL protection, layer-4 ACLs 417 generally need to be applied as close to the edge of the network as 418 possible, even though the intent is usually to protect the customer 419 edge rather than the provider core. Application of layer-4 DDoS 420 protection to a network edge is often automated using Flowspec 421 [RFC5575]. 423 For example, a web site which normally only handled traffic on TCP 424 ports 80 and 443 could be subject to a volumetric DDoS attack using 425 NTP and DNS packets with randomised source IP address, thereby 426 rendering traditional [RFC5635] source-based real-time black hole 427 mechanisms useless. In this situation, DDoS protection ACLs could be 428 configured to block all UDP traffic at the network edge without 429 impairing the web server functionality in any way. Thus, being able 430 to block arbitrary protocols at the network edge can avoid DDoS- 431 related problems both in the provider network and on the customer 432 edge link. 434 6.4. Network Intrusion Detection and Prevention 436 Network Intrusion Detection Systems (NIDS) examine network traffic 437 and try to identify traffic patterns that can be correlated to 438 network-based attacks. These systems generally inspect application- 439 layer traffic (if possible), but at the bare minimum inspect layer-4 440 flows. When attack activity is inferred, the operator is signaled of 441 the potential intrusion attempt. 443 Network Intrusion Prevention Systems (IPS) operate similarly to 444 NIDS's, but they may also prevent intrusions by reacting to detected 445 attack attempts by e.g. triggering packet filtering policies at 446 firewalls and other devices. 448 Use of extension headers may result problematic for NIDS/IPS, since: 450 o Extension headers increase the complexity of resulting traffic, 451 and the associated work and system requirements to process it. 453 o Use of unknown extension headers may prevent an NIDS/IPS to 454 process layer-4 information 456 o Use of IPv6 fragmentation requires a stateful fragment-reassembly 457 operation, even for decoy traffic employing forged source 458 addresses (see e.g. [nmap]). 460 As a result, in order to increase the efficiency or effectiveness of 461 these systems, packets employing IPv6 extension headers may be 462 dropped at the network ingress point(s) of networks that deploy these 463 systems. 465 6.5. Firewalling 467 Firewalls enforce security policies by means of packet filtering. 468 These systems generally inspect layer-3 and layer-4 traffic, and may 469 also examine application-layer traffic flows. 471 As with NIDS/IPS (Section 6.4), use of IPv6 extension headers may 472 represent a challenge to network firewalls, since: 474 o Extension headers increase the complexity of resulting traffic, 475 and the associated work and system requirements to process it (see 476 e.g. [Zack-FW-Benchmark]). 478 o Use of unknown extension headers may prevent an NIDS/IPS to 479 process layer-4 information 481 o Use of IPv6 fragmentation requires a stateful fragment-reassembly 482 operation, even for decoy traffic employing forged source 483 addresses (see e.g. [nmap]). 485 Additionally, a common firewall filtering policy is the so-called 486 "default deny", where all traffic is blocked (by default), and only 487 expected traffic is added to an "allow/accept list". 489 As a result, whether because of the challenges represented by 490 extension headers or because the use of IPv6 extension headers has 491 not been explicitly allowed, packets employing IPv6 extension headers 492 may be dropped by network firewalls. 494 7. Operational Implications 496 7.1. Inability to Find Layer-4 Information 498 As discussed in Section 6, contemporary routers and middle-boxes that 499 need to find the layer-4 header must process the entire IPv6 500 extension header chain. When such devices are unable to obtain the 501 required information, they may simply resort to dropping the 502 corresponding packets. 504 7.2. Route-Processor Protection 506 Most contemporary routers have a fast hardware-assisted forwarding 507 plane and a loosely coupled control plane, connected together with a 508 link that has much less capacity than the forwarding plane could 509 handle. Traffic differentiation cannot be done by the control plane 510 side, because this would overload the internal link connecting the 511 forwarding plane to the control plane. 513 The Hop-by-Hop Options header has been particularly challenging since 514 in most circumstances, the corresponding packet is punted to the 515 control plane for processing. As a result, operators usually drop 516 IPv6 packets containing this extension header. Please see [RFC6192] 517 for advice regarding protection of the router control plane. 519 7.3. Inability to Perform Fine-grained Filtering 521 Some router implementations lack fine-grained filtering of IPv6 522 extension headers. For example, an operator may want to drop packets 523 containing Routing Header Type 0 (RHT0) but may only be able to 524 filter on the extension header type (Routing Header). As a result, 525 the operator may end up enforcing a more coarse filtering policy 526 (e.g. "drop all packets containing a Routing Header" vs. "only drop 527 packets that contain a Routing Header Type 0"). 529 7.4. Security Concerns Associated with IPv6 Extension Headers 531 The security implications of IPv6 Extension Headers generally fall 532 into one or more of these categories: 534 o Evasion of security controls 536 o DoS due to processing requirements 538 o DoS due to implementation errors 540 o Extension Header-specific issues 542 Unlike IPv4 packets where the upper-layer protocol can be trivially 543 found by means of the "IHL" ("Internet Header Length") IPv4 header 544 field, the structure of IPv6 packets is more flexible and complex, 545 and may represent a challenge for devices that need to find this 546 information, since locating upper-layer protocol information requires 547 that all IPv6 extension headers be examined. This has presented 548 implementation difficulties, and packet filtering mechanisms that 549 require upper-layer information (even if just the upper layer 550 protocol type) can be trivially circumvented by inserting IPv6 551 Extension Headers between the main IPv6 header and the upper layer 552 protocol. [RFC7113] describes this issue for the RA-Guard case, but 553 the same techniques can be employed to circumvent other IPv6 firewall 554 and packet filtering mechanisms. Additionally, implementation 555 inconsistencies in packet forwarding engines may result in evasion of 556 security controls [I-D.kampanakis-6man-ipv6-eh-parsing] [Atlasis2014] 557 [BH-EU-2014]. 559 Packets with attached IPv6 Extension Headers may impact performance 560 on routers that forward them. Unless appropriate mitigations are put 561 in place (e.g., packet dropping and/or rate-limiting), an attacker 562 could simply send a large amount of IPv6 traffic employing IPv6 563 Extension Headers with the purpose of performing a Denial of Service 564 (DoS) attack (see Section 7 for further details). 566 NOTE: 568 In the most trivial case, a packet that includes a Hop-by-Hop 569 Options header might go through the slow forwarding path, and be 570 processed by the router's CPU. Another possible case might be 571 where a router that has been configured to enforce an ACL based on 572 upper-layer information (e.g., upper layer protocol or TCP 573 Destination Port), needs to process the entire IPv6 header chain 574 (in order to find the required information), causing the packet to 575 be processed in the slow path [Cisco-EH-Cons]. We note that, for 576 obvious reasons, the aforementioned performance issues may affect 577 other devices such as firewalls, Network Intrusion Detection 578 Systems (NIDS), etc. [Zack-FW-Benchmark]. The extent to which 579 these devices are affected is typically implementation-dependent. 581 IPv6 implementations, like all other software, tend to mature with 582 time and wide-scale deployment. While the IPv6 protocol itself has 583 existed for over 20 years, serious bugs related to IPv6 Extension 584 Header processing continue to be discovered (see e.g. [Cisco-Frag1], 585 [Cisco-Frag2], and [FreeBSD-SA]). Because there is currently little 586 operational reliance on IPv6 Extension headers, the corresponding 587 code paths are rarely exercised, and there is the potential for bugs 588 that still remain to be discovered in some implementations. 590 IPv6 Fragment Headers are employed to allow fragmentation of IPv6 591 packets. While many of the security implications of the 592 fragmentation / reassembly mechanism are known from the IPv4 world, 593 several related issues have crept into IPv6 implementations. These 594 range from denial of service attacks to information leakage, as 595 discussed in [RFC7739], [Bonica-NANOG58] and [Atlasis2012]). 597 8. IANA Considerations 599 There are no IANA registries within this document. The RFC-Editor 600 can remove this section before publication of this document as an 601 RFC. 603 9. Security Considerations 605 The security implications of IPv6 extension headers are discussed in 606 Section 7.4. This document does not introduce any new security 607 issues. 609 10. Acknowledgements 611 The authors would like to thank (in alphabetical order) Mikael 612 Abrahamsson, Fred Baker, Brian Carpenter, Tim Chown, Owen DeLong, Tom 613 Herbert, Lee Howard, Tom Petch, Sander Steffann, Eduard Vasilenko, 614 Eric Vyncke, Jingrong Xie, and Andrew Yourtchenko, for providing 615 valuable comments on earlier versions of this document. 617 Fernando Gont would like to thank Jan Zorz / Go6 Lab 618 , Jared Mauch, and Sander Steffann 619 , for providing access to systems and networks 620 that were employed to perform experiments and measurements involving 621 packets with IPv6 Extension Headers. 623 11. References 625 11.1. Normative References 627 [RFC5095] Abley, J., Savola, P., and G. Neville-Neil, "Deprecation 628 of Type 0 Routing Headers in IPv6", RFC 5095, 629 DOI 10.17487/RFC5095, December 2007, 630 . 632 [RFC5722] Krishnan, S., "Handling of Overlapping IPv6 Fragments", 633 RFC 5722, DOI 10.17487/RFC5722, December 2009, 634 . 636 [RFC6946] Gont, F., "Processing of IPv6 "Atomic" Fragments", 637 RFC 6946, DOI 10.17487/RFC6946, May 2013, 638 . 640 [RFC6980] Gont, F., "Security Implications of IPv6 Fragmentation 641 with IPv6 Neighbor Discovery", RFC 6980, 642 DOI 10.17487/RFC6980, August 2013, 643 . 645 [RFC7112] Gont, F., Manral, V., and R. Bonica, "Implications of 646 Oversized IPv6 Header Chains", RFC 7112, 647 DOI 10.17487/RFC7112, January 2014, 648 . 650 [RFC8021] Gont, F., Liu, W., and T. Anderson, "Generation of IPv6 651 Atomic Fragments Considered Harmful", RFC 8021, 652 DOI 10.17487/RFC8021, January 2017, 653 . 655 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 656 (IPv6) Specification", STD 86, RFC 8200, 657 DOI 10.17487/RFC8200, July 2017, 658 . 660 [RFC8504] Chown, T., Loughney, J., and T. Winters, "IPv6 Node 661 Requirements", BCP 220, RFC 8504, DOI 10.17487/RFC8504, 662 January 2019, . 664 11.2. Informative References 666 [APNIC-Scudder] 667 Scudder, J., "Modern router architecture and IPv6", APNIC 668 Blog, June 4, 2020, . 671 [Atlasis2012] 672 Atlasis, A., "Attacking IPv6 Implementation Using 673 Fragmentation", BlackHat Europe 2012. Amsterdam, 674 Netherlands. March 14-16, 2012, 675 . 678 [Atlasis2014] 679 Atlasis, A., "A Novel Way of Abusing IPv6 Extension 680 Headers to Evade IPv6 Security Devices", May 2014, 681 . 684 [BH-EU-2014] 685 Atlasis, A., Rey, E., and R. Schaefer, "Evasion of High- 686 End IDPS Devices at the IPv6 Era", BlackHat Europe 2014, 687 2014, . 690 [Bonica-NANOG58] 691 Bonica, R., "IPV6 FRAGMENTATION: The Case For 692 Deprecation", NANOG 58. New Orleans, Louisiana, USA. June 693 3-5, 2013, . 696 [Cisco-EH-Cons] 697 Cisco, "IPv6 Extension Headers Review and Considerations", 698 October 2006, 699 . 702 [Cisco-Frag1] 703 Cisco, "Cisco IOS Software IPv6 Virtual Fragmentation 704 Reassembly Denial of Service Vulnerability", September 705 2013, . 708 [Cisco-Frag2] 709 Cisco, "Cisco IOS XR Software Crafted IPv6 Packet Denial 710 of Service Vulnerability", June 2015, 711 . 714 [Cunha-2020] 715 Cunha, I., "IPv4 vs IPv6 load balancing in Internet 716 routes", NPS/CAIDA 2020 Virtual IPv6 Workshop, 2020, 717 . 720 [FreeBSD-SA] 721 FreeBSD, "FreeBSD Security Advisory FreeBSD-SA-20:24.ipv6: 722 IPv6 Hop-by-Hop options use-after-free bug", September 723 2020, . 726 [Gont-Chown-IEPG89] 727 Gont, F. and T. Chown, "A Small Update on the Use of IPv6 728 Extension Headers", IEPG 89. London, UK. March 2, 2014, 729 . 732 [Gont-IEPG88] 733 Gont, F., "Fragmentation and Extension header Support in 734 the IPv6 Internet", IEPG 88. Vancouver, BC, Canada. 735 November 13, 2013, . 738 [Huston-2017] 739 Huston, G., "Dealing with IPv6 fragmentation in the 740 DNS", APNIC Blog, 2017, 741 . 744 [Huston-2020] 745 Huston, G., "Measurement of IPv6 Extension Header 746 Support", NPS/CAIDA 2020 Virtual IPv6 Workshop, 2020, 747 . 750 [I-D.ietf-opsec-ipv6-eh-filtering] 751 Gont, F. and W. LIU, "Recommendations on the Filtering of 752 IPv6 Packets Containing IPv6 Extension Headers", draft- 753 ietf-opsec-ipv6-eh-filtering-06 (work in progress), July 754 2018. 756 [I-D.kampanakis-6man-ipv6-eh-parsing] 757 Kampanakis, P., "Implementation Guidelines for parsing 758 IPv6 Extension Headers", draft-kampanakis-6man-ipv6-eh- 759 parsing-01 (work in progress), August 2014. 761 [I-D.taylor-v6ops-fragdrop] 762 Jaeggli, J., Colitti, L., Kumari, W., Vyncke, E., Kaeo, 763 M., and T. Taylor, "Why Operators Filter Fragments and 764 What It Implies", draft-taylor-v6ops-fragdrop-02 (work in 765 progress), December 2013. 767 [I-D.wkumari-long-headers] 768 Kumari, W., Jaeggli, J., Bonica, R., and J. Linkova, 769 "Operational Issues Associated With Long IPv6 Header 770 Chains", draft-wkumari-long-headers-03 (work in progress), 771 June 2015. 773 [IEPG94-Scudder] 774 Petersen, B. and J. Scudder, "Modern Router Architecture 775 for Protocol Designers", IEPG 94. Yokohama, Japan. 776 November 1, 2015, . 779 [Jaeggli-2018] 780 Jaeggli, G., "Dealing with IPv6 fragmentation in the 781 DNS", APNIC Blog, 2018, 782 . 785 [Linkova-Gont-IEPG90] 786 Linkova, J. and F. Gont, "IPv6 Extension Headers in the 787 Real World v2.0", IEPG 90. Toronto, ON, Canada. July 20, 788 2014, . 791 [nmap] Fyodor, "Dealing with IPv6 fragmentation in the 792 DNS", Firewall/IDS Evasion and Spoofing, 793 . 795 [PMTUD-Blackholes] 796 De Boer, M. and J. Bosma, "Discovering Path MTU black 797 holes on the Internet using RIPE Atlas", July 2012, 798 . 801 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 802 (IPv6) Specification", RFC 2460, DOI 10.17487/RFC2460, 803 December 1998, . 805 [RFC5575] Marques, P., Sheth, N., Raszuk, R., Greene, B., Mauch, J., 806 and D. McPherson, "Dissemination of Flow Specification 807 Rules", RFC 5575, DOI 10.17487/RFC5575, August 2009, 808 . 810 [RFC5635] Kumari, W. and D. McPherson, "Remote Triggered Black Hole 811 Filtering with Unicast Reverse Path Forwarding (uRPF)", 812 RFC 5635, DOI 10.17487/RFC5635, August 2009, 813 . 815 [RFC6192] Dugal, D., Pignataro, C., and R. Dunn, "Protecting the 816 Router Control Plane", RFC 6192, DOI 10.17487/RFC6192, 817 March 2011, . 819 [RFC6437] Amante, S., Carpenter, B., Jiang, S., and J. Rajahalme, 820 "IPv6 Flow Label Specification", RFC 6437, 821 DOI 10.17487/RFC6437, November 2011, 822 . 824 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 825 for Equal Cost Multipath Routing and Link Aggregation in 826 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 827 . 829 [RFC7045] Carpenter, B. and S. Jiang, "Transmission and Processing 830 of IPv6 Extension Headers", RFC 7045, 831 DOI 10.17487/RFC7045, December 2013, 832 . 834 [RFC7113] Gont, F., "Implementation Advice for IPv6 Router 835 Advertisement Guard (RA-Guard)", RFC 7113, 836 DOI 10.17487/RFC7113, February 2014, 837 . 839 [RFC7739] Gont, F., "Security Implications of Predictable Fragment 840 Identification Values", RFC 7739, DOI 10.17487/RFC7739, 841 February 2016, . 843 [RFC7872] Gont, F., Linkova, J., Chown, T., and W. Liu, 844 "Observations on the Dropping of Packets with IPv6 845 Extension Headers in the Real World", RFC 7872, 846 DOI 10.17487/RFC7872, June 2016, 847 . 849 [RFC8900] Bonica, R., Baker, F., Huston, G., Hinden, R., Troan, O., 850 and F. Gont, "IP Fragmentation Considered Fragile", 851 BCP 230, RFC 8900, DOI 10.17487/RFC8900, September 2020, 852 . 854 [Zack-FW-Benchmark] 855 Zack, E., "Firewall Security Assessment and Benchmarking 856 IPv6 Firewall Load Tests", IPv6 Hackers Meeting #1, 857 Berlin, Germany. June 30, 2013, 858 . 862 Authors' Addresses 864 Fernando Gont 865 SI6 Networks 866 Segurola y Habana 4310, 7mo Piso 867 Villa Devoto, Ciudad Autonoma de Buenos Aires 868 Argentina 870 Email: fgont@si6networks.com 871 URI: https://www.si6networks.com 873 Nick Hilliard 874 INEX 875 4027 Kingswood Road 876 Dublin 24 877 IE 879 Email: nick@inex.ie 881 Gert Doering 882 SpaceNet AG 883 Joseph-Dollinger-Bogen 14 884 Muenchen D-80807 885 Germany 887 Email: gert@space.net 889 Warren Kumari 890 Google 891 1600 Amphitheatre Parkway 892 Mountain View, CA 94043 893 US 895 Email: warren@kumari.net 896 Geoff Huston 898 Email: gih@apnic.net 899 URI: http://www.apnic.net 901 Will (Shucheng) Liu 902 Huawei Technologies 903 Bantian, Longgang District 904 Shenzhen 518129 905 P.R. China 907 Email: liushucheng@huawei.com