idnits 2.17.1 draft-ietf-6man-rfc4941bis-02.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 : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) == The 'Obsoletes: ' line in the draft header should list only the _numbers_ of the RFCs which will be obsoleted by this document (if approved); it should not include the word 'RFC' in the list. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (July 8, 2019) is 1748 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) ** Obsolete normative reference: RFC 4941 (Obsoleted by RFC 8981) -- Obsolete informational reference (is this intentional?): RFC 3315 (Obsoleted by RFC 8415) Summary: 2 errors (**), 0 flaws (~~), 2 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 IPv6 Maintenance (6man) Working Group F. Gont 3 Internet-Draft SI6 Networks / UTN-FRH 4 Obsoletes: rfc4941 (if approved) S. Krishnan 5 Intended status: Standards Track Ericsson Research 6 Expires: January 9, 2020 T. Narten 7 IBM Corporation 8 R. Draves 9 Microsoft Research 10 July 8, 2019 12 Privacy Extensions for Stateless Address Autoconfiguration in IPv6 13 draft-ietf-6man-rfc4941bis-02 15 Abstract 17 Nodes use IPv6 stateless address autoconfiguration to generate 18 addresses using a combination of locally available information and 19 information advertised by routers. Addresses are formed by combining 20 network prefixes with an interface identifier. This document 21 describes an extension that causes nodes to generate global scope 22 addresses from interface identifiers that change over time. Changing 23 the interface identifier (and the global scope addresses generated 24 from it) over time makes it more difficult for eavesdroppers and 25 other information collectors to identify when different addresses 26 used in different transactions actually correspond to the same node. 28 Status of This Memo 30 This Internet-Draft is submitted in full conformance with the 31 provisions of BCP 78 and BCP 79. 33 Internet-Drafts are working documents of the Internet Engineering 34 Task Force (IETF). Note that other groups may also distribute 35 working documents as Internet-Drafts. The list of current Internet- 36 Drafts is at https://datatracker.ietf.org/drafts/current/. 38 Internet-Drafts are draft documents valid for a maximum of six months 39 and may be updated, replaced, or obsoleted by other documents at any 40 time. It is inappropriate to use Internet-Drafts as reference 41 material or to cite them other than as "work in progress." 43 This Internet-Draft will expire on January 9, 2020. 45 Copyright Notice 47 Copyright (c) 2019 IETF Trust and the persons identified as the 48 document authors. All rights reserved. 50 This document is subject to BCP 78 and the IETF Trust's Legal 51 Provisions Relating to IETF Documents 52 (https://trustee.ietf.org/license-info) in effect on the date of 53 publication of this document. Please review these documents 54 carefully, as they describe your rights and restrictions with respect 55 to this document. Code Components extracted from this document must 56 include Simplified BSD License text as described in Section 4.e of 57 the Trust Legal Provisions and are provided without warranty as 58 described in the Simplified BSD License. 60 Table of Contents 62 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 63 1.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 3 64 1.2. Problem Statement . . . . . . . . . . . . . . . . . . . . 3 65 2. Background . . . . . . . . . . . . . . . . . . . . . . . . . 4 66 2.1. Extended Use of the Same Identifier . . . . . . . . . . . 4 67 2.2. Possible Approaches . . . . . . . . . . . . . . . . . . . 5 68 3. Protocol Description . . . . . . . . . . . . . . . . . . . . 6 69 3.1. Assumptions . . . . . . . . . . . . . . . . . . . . . . . 7 70 3.2. Generation of Randomized Interface Identifiers . . . . . 7 71 3.2.1. Simple Randomized Interface Identifiers . . . . . . . 8 72 3.2.2. Hash-based Generation of Randomized Interface 73 Identifiers . . . . . . . . . . . . . . . . . . . . . 8 74 3.3. Generating Temporary Addresses . . . . . . . . . . . . . 10 75 3.4. Expiration of Temporary Addresses . . . . . . . . . . . . 11 76 3.5. Regeneration of Randomized Interface Identifiers . . . . 12 77 3.6. Deployment Considerations . . . . . . . . . . . . . . . . 13 78 4. Implications of Changing Interface Identifiers . . . . . . . 14 79 5. Defined Constants . . . . . . . . . . . . . . . . . . . . . . 14 80 6. Future Work . . . . . . . . . . . . . . . . . . . . . . . . . 15 81 7. Security Considerations . . . . . . . . . . . . . . . . . . . 15 82 8. Significant Changes from RFC4941 . . . . . . . . . . . . . . 16 83 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 16 84 10. References . . . . . . . . . . . . . . . . . . . . . . . . . 17 85 10.1. Normative References . . . . . . . . . . . . . . . . . . 17 86 10.2. Informative References . . . . . . . . . . . . . . . . . 18 87 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 19 89 1. Introduction 91 Stateless address autoconfiguration [RFC4862] defines how an IPv6 92 node generates addresses without the need for a Dynamic Host 93 Configuration Protocol for IPv6 (DHCPv6) server. The security and 94 privacy implications of such addresses have been discussed in great 95 detail in [RFC7721],[RFC7217], and RFC7707. This document specifies 96 an extension for SLAAC to generate temporary addresses, such that the 97 aforementioned issues are mitigated. 99 The default address selection for IPv6 has been specified in 100 [RFC6724]. We note that the determination as to whether to use 101 stable versus temporary addresses can in some cases only be made by 102 an application. For example, some applications may always want to 103 use temporary addresses, while others may want to use them only in 104 some circumstances or not at all. An API such as that specified in 105 [RFC5014] can enable individual applications to indicate with 106 sufficient granularity their needs with regards to the use of 107 temporary addresses. 109 Section 2 provides background information on the issue. Section 3 110 describes a procedure for generating temporary interface identifiers 111 and global scope addresses. Section 4 discusses implications of 112 changing interface identifiers. 114 1.1. Terminology 116 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 117 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 118 document are to be interpreted as described in [RFC2119]. 120 The terms "public address", "stable address", "temporary address", 121 "constant IID", "stable IID", and "temporary IID" are to be 122 interpreted as specified in [RFC7721]. 124 The term "global scope addresses" is used in this document to 125 collectively refer to "Global unicast addresses" as defined in 126 [RFC4291] and "Unique local addresses" as defined in [RFC4193]. 128 1.2. Problem Statement 130 Addresses generated using stateless address autoconfiguration 131 [RFC4862] contain an embedded interface identifier, which remains 132 stable over time. Anytime a fixed identifier is used in multiple 133 contexts, it becomes possible to correlate seemingly unrelated 134 activity using this identifier. 136 The correlation can be performed by 137 o An attacker who is in the path between the node in question and 138 the peer(s) to which it is communicating, and who can view the 139 IPv6 addresses present in the datagrams. 141 o An attacker who can access the communication logs of the peers 142 with which the node has communicated. 144 Since the identifier is embedded within the IPv6 address, which is a 145 fundamental requirement of communication, it cannot be easily hidden. 146 This document proposes a solution to this issue by generating 147 interface identifiers that vary over time. 149 Note that an attacker, who is on path, may be able to perform 150 significant correlation based on 152 o The payload contents of the packets on the wire 154 o The characteristics of the packets such as packet size and timing 156 Use of temporary addresses will not prevent such payload-based 157 correlation. 159 2. Background 161 This section discusses the problem in more detail, provides context 162 for evaluating the significance of the concerns in specific 163 environments and makes comparisons with existing practices. 165 2.1. Extended Use of the Same Identifier 167 The use of a non-changing interface identifier to form addresses is a 168 specific instance of the more general case where a constant 169 identifier is reused over an extended period of time and in multiple 170 independent activities. Any time the same identifier is used in 171 multiple contexts, it becomes possible for that identifier to be used 172 to correlate seemingly unrelated activity. For example, a network 173 sniffer placed strategically on a link across which all traffic to/ 174 from a particular host crosses could keep track of which destinations 175 a node communicated with and at what times. Such information can in 176 some cases be used to infer things, such as what hours an employee 177 was active, when someone is at home, etc. Although it might appear 178 that changing an address regularly in such environments would be 179 desirable to lessen privacy concerns, it should be noted that the 180 network prefix portion of an address also serves as a constant 181 identifier. All nodes at, say, a home, would have the same network 182 prefix, which identifies the topological location of those nodes. 183 This has implications for privacy, though not at the same granularity 184 as the concern that this document addresses. Specifically, all nodes 185 within a home could be grouped together for the purposes of 186 collecting information. If the network contains a very small number 187 of nodes, say, just one, changing just the interface identifier will 188 not enhance privacy at all, since the prefix serves as a constant 189 identifier. 191 One of the requirements for correlating seemingly unrelated 192 activities is the use (and reuse) of an identifier that is 193 recognizable over time within different contexts. IP addresses 194 provide one obvious example, but there are more. Many nodes also 195 have DNS names associated with their addresses, in which case the DNS 196 name serves as a similar identifier. Although the DNS name 197 associated with an address is more work to obtain (it may require a 198 DNS query), the information is often readily available. In such 199 cases, changing the address on a machine over time would do little to 200 address the concerns raised in this document, unless the DNS name is 201 changed as well (see Section 4). 203 Web browsers and servers typically exchange "cookies" with each other 204 [RFC6265]. Cookies allow web servers to correlate a current activity 205 with a previous activity. One common usage is to send back targeted 206 advertising to a user by using the cookie supplied by the browser to 207 identify what earlier queries had been made (e.g., for what type of 208 information). Based on the earlier queries, advertisements can be 209 targeted to match the (assumed) interests of the end-user. 211 The use of a constant identifier within an address is of special 212 concern because addresses are a fundamental requirement of 213 communication and cannot easily be hidden from eavesdroppers and 214 other parties. Even when higher layers encrypt their payloads, 215 addresses in packet headers appear in the clear. Consequently, if a 216 mobile host (e.g., laptop) accessed the network from several 217 different locations, an eavesdropper might be able to track the 218 movement of that mobile host from place to place, even if the upper 219 layer payloads were encrypted. 221 The security and privacy implications of IPv6 addresses are discussed 222 in detail in [RFC7721], [RFC7707], and [RFC7217]. 224 2.2. Possible Approaches 226 One way to avoid having a stable non-changing address is to use 227 DHCPv6 [RFC3315] for obtaining addresses. Section 12 of [RFC3315] 228 discusses the use of DHCPv6 for the assignment and management of 229 "temporary addresses", which are never renewed and provide the same 230 property of temporary addresses described in this document with 231 regards to the privacy concern. 233 Another approach, compatible with the stateless address 234 autoconfiguration architecture, would be to change the interface 235 identifier portion of an address over time. Changing the interface 236 identifier can make it more difficult to look at the IP addresses in 237 independent transactions and identify which ones actually correspond 238 to the same node, both in the case where the routing prefix portion 239 of an address changes and when it does not. 241 Many machines function as both clients and servers. In such cases, 242 the machine would need a DNS name for its use as a server. Whether 243 the address stays fixed or changes has little privacy implication 244 since the DNS name remains constant and serves as a constant 245 identifier. When acting as a client (e.g., initiating 246 communication), however, such a machine may want to vary the 247 addresses it uses. In such environments, one may need multiple 248 addresses: a stable address registered in the DNS, that is used to 249 accept incoming connection requests from other machines, and a 250 temporary address used to shield the identity of the client when it 251 initiates communication. These two cases are roughly analogous to 252 telephone numbers and caller ID, where a user may list their 253 telephone number in the public phone book, but disable the display of 254 its number via caller ID when initiating calls. 256 On the other hand, a machine that functions only as a client may want 257 to employ only temporary addresses for public communication. 259 To make it difficult to make educated guesses as to whether two 260 different interface identifiers belong to the same node, the 261 algorithm for generating alternate identifiers must include input 262 that has an unpredictable component from the perspective of the 263 outside entities that are collecting information. 265 [I-D.gont-6man-non-stable-iids] specifies requirements for temporary 266 addresses. This document specifies a number of algorithms for 267 generating temporary addresses that comply with the aforementioned 268 requirements. 270 3. Protocol Description 272 The goal of this section is to define procedures that: 274 1. Do not result in any changes to the basic behavior of addresses 275 generated via stateless address autoconfiguration [RFC4862]. 277 2. Create temporary addresses based on an unpredictable interface 278 identifier for the purpose of initiating outgoing sessions. 279 These temporary addresses would be used for a short period of 280 time (hours to days) and would then be deprecated. Deprecated 281 addresses can continue to be used for already established 282 connections, but are not used to initiate new connections. New 283 temporary addresses are generated periodically to replace 284 temporary addresses that expire, with the exact time between 285 address generation a matter of local policy. 287 3. Produce a sequence of temporary global scope addresses from a 288 sequence of interface identifiers that appear to be random in the 289 sense that it is difficult for an outside observer to predict a 290 future address (or identifier) based on a current one and it is 291 difficult to determine previous addresses (or identifiers) 292 knowing only the present one. 294 4. By default, generate one address for each prefix advertised for 295 stateless address autoconfiguration. The interface identifier 296 generated for each of those prefixes should be (statistically) 297 different. That is, a new interface identifier should be 298 computed for each temporary address that is to be generated. 300 3.1. Assumptions 302 The following algorithm assumes that for a given temporary address, 303 an implementation can determine the prefix from which it was 304 generated. When a temporary address is deprecated, a new temporary 305 address is generated. The specific valid and preferred lifetimes for 306 the new address are dependent on the corresponding lifetime values 307 set for the prefix from which it was generated. 309 Finally, this document assumes that when a node initiates outgoing 310 communication, temporary addresses can be given preference over 311 stable addresses (if available), when the device is configured to do 312 so. [RFC6724] mandates implementations to provide a mechanism, which 313 allows an application to configure its preference for temporary 314 addresses over stable addresses. It also allows for an 315 implementation to prefer temporary addresses by default, so that the 316 connections initiated by the node can use temporary addresses without 317 requiring application-specific enablement. This document also 318 assumes that an API will exist that allows individual applications to 319 indicate whether they prefer to use temporary or stable addresses and 320 override the system defaults. 322 3.2. Generation of Randomized Interface Identifiers 324 The following subsections specify some possible algorithms for 325 generating temporary interface identifiers that comply with the 326 requirements in [I-D.gont-6man-non-stable-iids]. The algorithm 327 specified in Section 3.2.1 benefits from a Pseudo-Random Number 328 Generator (PRNG) available on the system. On the other hand, the 329 algorithm specified in Section 3.2.2 allows for code reuse by nodes 330 that implement [RFC7217]. 332 3.2.1. Simple Randomized Interface Identifiers 334 One possible approach would be to select a pseudorandom number of the 335 appropriate length. A node employing this algorithm should generate 336 IIDs as follows: 338 1. Obtain a random number (see [RFC4086] for randomness requirements 339 for security) 341 2. The Interface Identifier is obtained by taking as many bits from 342 the aforementioned random number (obtained in the previous step) 343 as necessary. 345 We note that [RFC4291] requires that the Interface IDs of all 346 unicast addresses (except those that start with the binary 347 value 000) be 64 bits long. However, the method discussed in 348 this document could be employed for generating Interface IDs 349 of any arbitrary length, albeit at the expense of reduced 350 entropy (when employing Interface IDs smaller than 64 bits). 352 3. The resulting Interface Identifier SHOULD be compared against the 353 reserved IPv6 Interface Identifiers [RFC5453] [IANA-RESERVED-IID] 354 and against those Interface Identifiers already employed in an 355 address of the same network interface and the same network 356 prefix. In the event that an unacceptable identifier has been 357 generated, a new interface identifier should be generated, by 358 repeating the algorithm from the first step. 360 3.2.2. Hash-based Generation of Randomized Interface Identifiers 362 The algorithm in [RFC7217] can be augmented for the generation of 363 temporary addresses. The benefit of this would be that a node could 364 employ a single algorithm for generating stable and temporary 365 addresses, by employing appropriate parameters. 367 Nodes would employ the following algorithm for generating the 368 temporary IID: 370 1. Compute a random identifier with the expression: 372 RID = F(Prefix, MAC_Address, Network_ID, Time, DAD_Counter, 373 secret_key) 375 Where: 377 RID: 378 Random Identifier 380 F(): 381 A pseudorandom function (PRF) that MUST NOT be computable from 382 the outside (without knowledge of the secret key). F() MUST 383 also be difficult to reverse, such that it resists attempts to 384 obtain the secret_key, even when given samples of the output 385 of F() and knowledge or control of the other input parameters. 386 F() SHOULD produce an output of at least 64 bits. F() could 387 be implemented as a cryptographic hash of the concatenation of 388 each of the function parameters. SHA-1 [FIPS-SHS] and SHA-256 389 are two possible options for F(). Note: MD5 [RFC1321] is 390 considered unacceptable for F() [RFC6151]. 392 Prefix: 393 The prefix to be used for SLAAC, as learned from an ICMPv6 394 Router Advertisement message. 396 MAC_Address: 397 The MAC address corresponding to the underlying network 398 interface card. Employing the MAC address in this expression 399 (in replacement of the Net_Iface parameter of the expression 400 in RFC7217) means that the re-generation of a randomized MAC 401 address will result in a different temporary address. 403 Network_ID: 404 Some network-specific data that identifies the subnet to which 405 this interface is attached -- for example, the IEEE 802.11 406 Service Set Identifier (SSID) corresponding to the network to 407 which this interface is associated. Additionally, Simple DNA 408 [RFC6059] describes ideas that could be leveraged to generate 409 a Network_ID parameter. This parameter is SHOULD be employed 410 if some form of "Network_ID" is available. 412 Time: 413 An implementation-dependent representation of time. One 414 possible example is the representation in UNIX-like systems 415 [OPEN-GROUP], that measure time in terms of the number of 416 seconds elapsed since the Epoch (00:00:00 Coordinated 417 Universal Time (UTC), 1 January 1970). 419 DAD_Counter: 420 A counter that is employed to resolve Duplicate Address 421 Detection (DAD) conflicts. 423 secret_key: 425 A secret key that is not known by the attacker. The secret 426 key SHOULD be of at least 128 bits. It MUST be initialized to 427 a pseudo-random number (see [RFC4086] for randomness 428 requirements for security) when the operating system is 429 "bootstrapped". 431 2. The Interface Identifier is finally obtained by taking as many 432 bits from the RID value (computed in the previous step) as 433 necessary, starting from the least significant bit. The 434 resulting Interface Identifier SHOULD be compared against the 435 reserved IPv6 Interface Identifiers [RFC5453] [IANA-RESERVED-IID] 436 and against those Interface Identifiers already employed in an 437 address of the same network interface and the same network 438 prefix. In the event that an unacceptable identifier has been 439 generated, the value DAD_Counter should be incremented by 1, and 440 the algorithm should be restarted from the first step. 442 3.3. Generating Temporary Addresses 444 [RFC4862] describes the steps for generating a link-local address 445 when an interface becomes enabled as well as the steps for generating 446 addresses for other scopes. This document extends [RFC4862] as 447 follows. When processing a Router Advertisement with a Prefix 448 Information option carrying a global scope prefix for the purposes of 449 address autoconfiguration (i.e., the A bit is set), the node MUST 450 perform the following steps: 452 1. Process the Prefix Information Option as defined in [RFC4862], 453 adjusting the lifetimes of existing temporary addresses. If a 454 received option may extend the lifetimes of temporary addresses, 455 with the overall constraint that no temporary addresses should 456 ever remain "valid" or "preferred" for a time longer than 457 (TEMP_VALID_LIFETIME) or (TEMP_PREFERRED_LIFETIME - 458 DESYNC_FACTOR) respectively. The configuration variables 459 TEMP_VALID_LIFETIME and TEMP_PREFERRED_LIFETIME correspond to 460 approximate target lifetimes for temporary addresses. 462 2. One way an implementation can satisfy the above constraints is to 463 associate with each temporary address a creation time (called 464 CREATION_TIME) that indicates the time at which the address was 465 created. When updating the preferred lifetime of an existing 466 temporary address, it would be set to expire at whichever time is 467 earlier: the time indicated by the received lifetime or 468 (CREATION_TIME + TEMP_PREFERRED_LIFETIME - DESYNC_FACTOR). A 469 similar approach can be used with the valid lifetime. 471 3. If the node has not configured any temporary address for the 472 corresponding prefix, the node SHOULD create a new temporary 473 address for such prefix. 475 4. When creating a temporary address, the lifetime values MUST be 476 derived from the corresponding prefix as follows: 478 * Its Valid Lifetime is the lower of the Valid Lifetime of the 479 prefix and TEMP_VALID_LIFETIME 481 * Its Preferred Lifetime is the lower of the Preferred Lifetime 482 of prefix and TEMP_PREFERRED_LIFETIME - DESYNC_FACTOR. 484 5. A temporary address is created only if this calculated Preferred 485 Lifetime is greater than REGEN_ADVANCE time units. In 486 particular, an implementation MUST NOT create a temporary address 487 with a zero Preferred Lifetime. 489 6. New temporary addresses MUST be created by appending a randomized 490 interface identifier (generates as described in Section 3.2 of 491 this document) to the prefix that was received. 493 7. The node MUST perform duplicate address detection (DAD) on the 494 generated temporary address. If DAD indicates the address is 495 already in use, the node MUST generate a new randomized interface 496 identifier, and repeat the previous steps as appropriate up to 497 TEMP_IDGEN_RETRIES times. If after TEMP_IDGEN_RETRIES 498 consecutive attempts no non-unique address was generated, the 499 node MUST log a system error and MUST NOT attempt to generate 500 temporary addresses for that interface. Note that DAD MUST be 501 performed on every unicast address generated from this randomized 502 interface identifier. 504 3.4. Expiration of Temporary Addresses 506 When a temporary address becomes deprecated, a new one MUST be 507 generated. This is done by repeating the actions described in 508 Section 3.3, starting at step 4). Note that, except for the 509 transient period when a temporary address is being regenerated, in 510 normal operation at most one temporary address per prefix should be 511 in a non-deprecated state at any given time on a given interface. 512 Note that if a temporary address becomes deprecated as result of 513 processing a Prefix Information Option with a zero Preferred 514 Lifetime, then a new temporary address MUST NOT be generated. To 515 ensure that a preferred temporary address is always available, a new 516 temporary address SHOULD be regenerated slightly before its 517 predecessor is deprecated. This is to allow sufficient time to avoid 518 race conditions in the case where generating a new temporary address 519 is not instantaneous, such as when duplicate address detection must 520 be run. The node SHOULD start the address regeneration process 521 REGEN_ADVANCE time units before a temporary address would actually be 522 deprecated. 524 As an optional optimization, an implementation MAY remove a 525 deprecated temporary address that is not in use by applications or 526 upper layers as detailed in Section 6. 528 3.5. Regeneration of Randomized Interface Identifiers 530 The frequency at which temporary addresses change depends on how a 531 device is being used (e.g., how frequently it initiates new 532 communication) and the concerns of the end user. The most egregious 533 privacy concerns appear to involve addresses used for long periods of 534 time (weeks to months to years). The more frequently an address 535 changes, the less feasible collecting or coordinating information 536 keyed on interface identifiers becomes. Moreover, the cost of 537 collecting information and attempting to correlate it based on 538 interface identifiers will only be justified if enough addresses 539 contain non-changing identifiers to make it worthwhile. Thus, having 540 large numbers of clients change their address on a daily or weekly 541 basis is likely to be sufficient to alleviate most privacy concerns. 543 There are also client costs associated with having a large number of 544 addresses associated with a node (e.g., in doing address lookups, the 545 need to join many multicast groups, etc.). Thus, changing addresses 546 frequently (e.g., every few minutes) may have performance 547 implications. 549 Nodes following this specification SHOULD generate new temporary 550 addresses on a periodic basis. This can be achieved automatically by 551 generating a new randomized interface identifier at least once every 552 (TEMP_PREFERRED_LIFETIME - REGEN_ADVANCE - DESYNC_FACTOR) time units. 553 As described above, generating a new temporary address REGEN_ADVANCE 554 time units before a temporary address becomes deprecated produces 555 addresses with a preferred lifetime no larger than 556 TEMP_PREFERRED_LIFETIME. The value DESYNC_FACTOR is a random value 557 (different for each client) that ensures that clients don't 558 synchronize with each other and generate new addresses at exactly the 559 same time. When the preferred lifetime expires, a new temporary 560 address MUST be generated using the new randomized interface 561 identifier. 563 Because the precise frequency at which it is appropriate to generate 564 new addresses varies from one environment to another, implementations 565 SHOULD provide end users with the ability to change the frequency at 566 which addresses are regenerated. The default value is given in 567 TEMP_PREFERRED_LIFETIME and is one day. In addition, the exact time 568 at which to invalidate a temporary address depends on how 569 applications are used by end users. Thus, the suggested default 570 value of one week (TEMP_VALID_LIFETIME) may not be appropriate in all 571 environments. Implementations SHOULD provide end users with the 572 ability to override both of these default values. 574 Finally, when an interface connects to a new (different) link, a new 575 set of temporary addresses MUST be generated immediately. If a 576 device moves from one ethernet to another, generating a new set of 577 temporary addresses ensures that the device uses different randomized 578 interface identifiers for the temporary addresses associated with the 579 two links, making it more difficult to correlate addresses from the 580 two different links as being from the same node. The node MAY follow 581 any process available to it, to determine that the link change has 582 occurred. One such process is described by "Simple Procedures for 583 Detecting Network Attachment in IPv6" [RFC6059]. Detecting link 584 changes would prevent link down/up events from causing temporary 585 addresses to be (unnecessarily) regenerated. 587 3.6. Deployment Considerations 589 Devices implementing this specification MUST provide a way for the 590 end user to explicitly enable or disable the use of temporary 591 addresses. In addition, a site might wish to disable the use of 592 temporary addresses in order to simplify network debugging and 593 operations. Consequently, implementations SHOULD provide a way for 594 trusted system administrators to enable or disable the use of 595 temporary addresses. 597 Additionally, sites might wish to selectively enable or disable the 598 use of temporary addresses for some prefixes. For example, a site 599 might wish to disable temporary address generation for "Unique local" 600 [RFC4193] prefixes while still generating temporary addresses for all 601 other global prefixes. Another site might wish to enable temporary 602 address generation only for the prefixes 2001::/16 and 2002::/16 603 while disabling it for all other prefixes. To support this behavior, 604 implementations SHOULD provide a way to enable and disable generation 605 of temporary addresses for specific prefix subranges. This per- 606 prefix setting SHOULD override the global settings on the node with 607 respect to the specified prefix subranges. Note that the per-prefix 608 setting can be applied at any granularity, and not necessarily on a 609 per subnet basis. 611 The use of temporary addresses may cause unexpected difficulties with 612 some applications. As described below, some servers refuse to accept 613 communications from clients for which they cannot map the IP address 614 into a DNS name. In addition, some applications may not behave 615 robustly if temporary addresses are used and an address expires 616 before the application has terminated, or if it opens multiple 617 sessions, but expects them to all use the same addresses. 619 If a very small number of nodes (say, only one) use a given prefix 620 for extended periods of time, just changing the interface identifier 621 part of the address may not be sufficient to ensure privacy, since 622 the prefix acts as a constant identifier. The procedures described 623 in this document are most effective when the prefix is reasonably non 624 static or is used by a fairly large number of nodes. 626 4. Implications of Changing Interface Identifiers 628 The desires of protecting individual privacy versus the desire to 629 effectively maintain and debug a network can conflict with each 630 other. Having clients use addresses that change over time will make 631 it more difficult to track down and isolate operational problems. 632 For example, when looking at packet traces, it could become more 633 difficult to determine whether one is seeing behavior caused by a 634 single errant machine, or by a number of them. 636 Some servers refuse to grant access to clients for which no DNS name 637 exists. That is, they perform a DNS PTR query to determine the DNS 638 name, and may then also perform an AAAA query on the returned name to 639 verify that the returned DNS name maps back into the address being 640 used. Consequently, clients not properly registered in the DNS may 641 be unable to access some services. As noted earlier, however, a 642 node's DNS name (if non-changing) serves as a constant identifier. 643 The wide deployment of the extension described in this document could 644 challenge the practice of inverse-DNS-based "authentication," which 645 has little validity, though it is widely implemented. In order to 646 meet server challenges, nodes could register temporary addresses in 647 the DNS using random names (for example, a string version of the 648 random address itself). 650 Use of the extensions defined in this document may complicate 651 debugging and other operational troubleshooting activities. 652 Consequently, it may be site policy that temporary addresses should 653 not be used. Consequently, implementations MUST provide a method for 654 the end user or trusted administrator to override the use of 655 temporary addresses. 657 5. Defined Constants 659 Constants defined in this document include: 661 TEMP_VALID_LIFETIME -- Default value: 1 week. Users should be able 662 to override the default value. 664 TEMP_PREFERRED_LIFETIME -- Default value: 1 day. Users should be 665 able to override the default value. 667 REGEN_ADVANCE -- 5 seconds 669 MAX_DESYNC_FACTOR -- 10 minutes. Upper bound on DESYNC_FACTOR. 671 DESYNC_FACTOR -- A random value within the range 0 - 672 MAX_DESYNC_FACTOR. It is computed once at system start (rather than 673 each time it is used) and must never be greater than 674 (TEMP_VALID_LIFETIME - REGEN_ADVANCE). 676 TEMP_IDGEN_RETRIES -- Default value: 3 678 6. Future Work 680 An implementation might want to keep track of which addresses are 681 being used by upper layers so as to be able to remove a deprecated 682 temporary address from internal data structures once no upper layer 683 protocols are using it (but not before). This is in contrast to 684 current approaches where addresses are removed from an interface when 685 they become invalid [RFC4862], independent of whether or not upper 686 layer protocols are still using them. For TCP connections, such 687 information is available in control blocks. For UDP-based 688 applications, it may be the case that only the applications have 689 knowledge about what addresses are actually in use. Consequently, an 690 implementation generally will need to use heuristics in deciding when 691 an address is no longer in use. 693 Recommendations on DNS practices to avoid the problem described in 694 Section 4 when reverse DNS lookups fail may be needed. [RFC4472] 695 contains a more detailed discussion of the DNS-related issues. 697 While this document discusses ways of obscuring a user's IP address, 698 the method described is believed to be ineffective against 699 sophisticated forms of traffic analysis. To increase effectiveness, 700 one may need to consider use of more advanced techniques, such as 701 Onion Routing [ONION]. 703 7. Security Considerations 705 Ingress filtering has been and is being deployed as a means of 706 preventing the use of spoofed source addresses in Distributed Denial 707 of Service (DDoS) attacks. In a network with a large number of 708 nodes, new temporary addresses are created at a fairly high rate. 709 This might make it difficult for ingress filtering mechanisms to 710 distinguish between legitimately changing temporary addresses and 711 spoofed source addresses, which are "in-prefix" (using a 712 topologically correct prefix and non-existent interface ID). This 713 can be addressed by using access control mechanisms on a per-address 714 basis on the network egress point. 716 8. Significant Changes from RFC4941 718 This section summarizes the changes in this document relative to RFC 719 4941 that an implementer of RFC 4941 should be aware of. 721 1. Discussion of IEEE-based IIDs has been removed, since the current 722 recommendation ([RFC8064]) is to employ [RFC7217]). 724 2. The document employs the terminology from [RFC7721]. 726 3. Sections 2.2 and 2.3 of [RFC4941] have been removed since the 727 topic has been discussed in more detail in e.g. [RFC7721]. 729 4. The algorithm to generate randomized interface identifiers was 730 replaced by two possible alternative algorithms. 732 5. Generation of stable addresses is not implied or required by this 733 document. 735 6. Temporary addresses are *not* disabled by default. 737 7. Section 3.2.1 and 3.2.2 from [RFC4941] were replaced with 738 alternative algorithms. 740 8. Section 3.2.3 from [RFC4941] was removed, based on the 741 explanation of that very section of RFC4941. 743 9. All the verified errata for [RFC4941] has been incorporated. 745 9. Acknowledgments 747 The authors would like to thank (in alphabetical order) Brian 748 Carpenter, Tim Chown, Lorenzo Colitti, David Farmer, Tom Herbert, Bob 749 Hinden, Christian Huitema, Dave Plonka, Michael Richardson, Mark 750 Smith, and Johanna Ullrich for providing valuable comments on earlier 751 versions of this document. 753 This document is based on [RFC4941] (a revision of RFC3041). Suresh 754 Krishnan was the sole author of RFC4941. He would like to 755 acknowledge the contributions of the ipv6 working group and, in 756 particular, Jari Arkko, Pekka Nikander, Pekka Savola, Francis Dupont, 757 Brian Haberman, Tatuya Jinmei, and Margaret Wasserman for their 758 detailed comments. 760 Rich Draves and Thomas Narten were the authors of RFC 3041. They 761 would like to acknowledge the contributions of the ipv6 working group 762 and, in particular, Ran Atkinson, Matt Crawford, Steve Deering, 763 Allison Mankin, and Peter Bieringer. 765 10. References 767 10.1. Normative References 769 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 770 Requirement Levels", BCP 14, RFC 2119, 771 DOI 10.17487/RFC2119, March 1997, 772 . 774 [RFC4086] Eastlake 3rd, D., Schiller, J., and S. Crocker, 775 "Randomness Requirements for Security", BCP 106, RFC 4086, 776 DOI 10.17487/RFC4086, June 2005, 777 . 779 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 780 Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005, 781 . 783 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 784 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 785 2006, . 787 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 788 Address Autoconfiguration", RFC 4862, 789 DOI 10.17487/RFC4862, September 2007, 790 . 792 [RFC4941] Narten, T., Draves, R., and S. Krishnan, "Privacy 793 Extensions for Stateless Address Autoconfiguration in 794 IPv6", RFC 4941, DOI 10.17487/RFC4941, September 2007, 795 . 797 [RFC5453] Krishnan, S., "Reserved IPv6 Interface Identifiers", 798 RFC 5453, DOI 10.17487/RFC5453, February 2009, 799 . 801 [RFC6724] Thaler, D., Ed., Draves, R., Matsumoto, A., and T. Chown, 802 "Default Address Selection for Internet Protocol Version 6 803 (IPv6)", RFC 6724, DOI 10.17487/RFC6724, September 2012, 804 . 806 [RFC7217] Gont, F., "A Method for Generating Semantically Opaque 807 Interface Identifiers with IPv6 Stateless Address 808 Autoconfiguration (SLAAC)", RFC 7217, 809 DOI 10.17487/RFC7217, April 2014, 810 . 812 [RFC8064] Gont, F., Cooper, A., Thaler, D., and W. Liu, 813 "Recommendation on Stable IPv6 Interface Identifiers", 814 RFC 8064, DOI 10.17487/RFC8064, February 2017, 815 . 817 10.2. Informative References 819 [FIPS-SHS] 820 NIST, "Secure Hash Standard (SHS)", FIPS 821 Publication 180-4, March 2012, 822 . 825 [I-D.gont-6man-non-stable-iids] 826 Gont, F., Huitema, C., Krishnan, S., Gont, G., and M. 827 Corbo, "Recommendation on Temporary IPv6 Interface 828 Identifiers", draft-gont-6man-non-stable-iids-04 (work in 829 progress), March 2018. 831 [IANA-RESERVED-IID] 832 IANA, "Reserved IPv6 Interface Identifiers", 833 . 835 [ONION] Reed, MGR., Syverson, PFS., and DMG. Goldschlag, "Proxies 836 for Anonymous Routing", Proceedings of the 12th Annual 837 Computer Security Applications Conference, San Diego, CA, 838 December 1996. 840 [OPEN-GROUP] 841 The Open Group, "The Open Group Base Specifications Issue 842 7 / IEEE Std 1003.1-2008, 2016 Edition", 843 Section 4.16 Seconds Since the Epoch, 2016, 844 . 847 [RFC1321] Rivest, R., "The MD5 Message-Digest Algorithm", RFC 1321, 848 DOI 10.17487/RFC1321, April 1992, 849 . 851 [RFC3315] Droms, R., Ed., Bound, J., Volz, B., Lemon, T., Perkins, 852 C., and M. Carney, "Dynamic Host Configuration Protocol 853 for IPv6 (DHCPv6)", RFC 3315, DOI 10.17487/RFC3315, July 854 2003, . 856 [RFC4472] Durand, A., Ihren, J., and P. Savola, "Operational 857 Considerations and Issues with IPv6 DNS", RFC 4472, 858 DOI 10.17487/RFC4472, April 2006, 859 . 861 [RFC5014] Nordmark, E., Chakrabarti, S., and J. Laganier, "IPv6 862 Socket API for Source Address Selection", RFC 5014, 863 DOI 10.17487/RFC5014, September 2007, 864 . 866 [RFC6059] Krishnan, S. and G. Daley, "Simple Procedures for 867 Detecting Network Attachment in IPv6", RFC 6059, 868 DOI 10.17487/RFC6059, November 2010, 869 . 871 [RFC6151] Turner, S. and L. Chen, "Updated Security Considerations 872 for the MD5 Message-Digest and the HMAC-MD5 Algorithms", 873 RFC 6151, DOI 10.17487/RFC6151, March 2011, 874 . 876 [RFC6265] Barth, A., "HTTP State Management Mechanism", RFC 6265, 877 DOI 10.17487/RFC6265, April 2011, 878 . 880 [RFC7707] Gont, F. and T. Chown, "Network Reconnaissance in IPv6 881 Networks", RFC 7707, DOI 10.17487/RFC7707, March 2016, 882 . 884 [RFC7721] Cooper, A., Gont, F., and D. Thaler, "Security and Privacy 885 Considerations for IPv6 Address Generation Mechanisms", 886 RFC 7721, DOI 10.17487/RFC7721, March 2016, 887 . 889 Authors' Addresses 890 Fernando Gont 891 SI6 Networks / UTN-FRH 892 Evaristo Carriego 2644 893 Haedo, Provincia de Buenos Aires 1706 894 Argentina 896 Phone: +54 11 4650 8472 897 Email: fgont@si6networks.com 898 URI: http://www.si6networks.com 900 Suresh Krishnan 901 Ericsson Research 902 8400 Decarie Blvd. 903 Town of Mount Royal, QC 904 Canada 906 Email: suresh.krishnan@ericsson.com 908 Thomas Narten 909 IBM Corporation 910 P.O. Box 12195 911 Research Triangle Park, NC 912 USA 914 Email: narten@us.ibm.com 916 Richard Draves 917 Microsoft Research 918 One Microsoft Way 919 Redmond, WA 920 USA 922 Email: richdr@microsoft.com