idnits 2.17.1 draft-ietf-ipv6-privacy-addrs-v2-05.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 18. -- Found old boilerplate from RFC 3978, Section 5.5 on line 1030. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 1007. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 1014. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 1020. ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** This document has an original RFC 3978 Section 5.5 Disclaimer, instead of the newer disclaimer which includes the IETF Trust according to RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == No 'Intended status' indicated for this document; assuming Proposed Standard 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 draft header indicates that this document obsoletes RFC3041, but the abstract doesn't seem to mention this, which it should. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (August 2006) is 6458 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) == Missing Reference: 'RFC' is mentioned on line 865, but not defined ** Obsolete normative reference: RFC 3513 (ref. 'ADDRARCH') (Obsoleted by RFC 4291) == Outdated reference: A later version (-08) exists of draft-ietf-ipv6-rfc2462bis-07 == Outdated reference: A later version (-11) exists of draft-ietf-ipv6-2461bis-02 ** Downref: Normative reference to an Informational RFC: RFC 1321 (ref. 'MD5') -- Obsolete informational reference (is this intentional?): RFC 2965 (ref. 'COOKIES') (Obsoleted by RFC 6265) -- Obsolete informational reference (is this intentional?): RFC 3315 (ref. 'DHCPV6') (Obsoleted by RFC 8415) == Outdated reference: A later version (-12) exists of draft-ietf-dnsop-ipv6-dns-issues-10 -- Obsolete informational reference (is this intentional?): RFC 1750 (ref. 'RANDOM') (Obsoleted by RFC 4086) Summary: 6 errors (**), 0 flaws (~~), 6 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 IPv6 Working Group T. Narten 3 Internet-Draft IBM Corporation 4 Obsoletes: 3041 (if approved) R. Draves 5 Expires: February 2, 2007 Microsoft Research 6 S. Krishnan 7 Ericsson Research 8 August 2006 10 Privacy Extensions for Stateless Address Autoconfiguration in IPv6 11 draft-ietf-ipv6-privacy-addrs-v2-05 13 Status of this Memo 15 By submitting this Internet-Draft, each author represents that any 16 applicable patent or other IPR claims of which he or she is aware 17 have been or will be disclosed, and any of which he or she becomes 18 aware will be disclosed, in accordance with Section 6 of BCP 79. 20 Internet-Drafts are working documents of the Internet Engineering 21 Task Force (IETF), its areas, and its working groups. Note that 22 other groups may also distribute working documents as Internet- 23 Drafts. 25 Internet-Drafts are draft documents valid for a maximum of six months 26 and may be updated, replaced, or obsoleted by other documents at any 27 time. It is inappropriate to use Internet-Drafts as reference 28 material or to cite them other than as "work in progress." 30 The list of current Internet-Drafts can be accessed at 31 http://www.ietf.org/ietf/1id-abstracts.txt. 33 The list of Internet-Draft Shadow Directories can be accessed at 34 http://www.ietf.org/shadow.html. 36 This Internet-Draft will expire on February 2, 2007. 38 Copyright Notice 40 Copyright (C) The Internet Society (2006). 42 Abstract 44 Nodes use IPv6 stateless address autoconfiguration to generate 45 addresses using a combination of locally available information and 46 information advertised by routers. Addresses are formed by combining 47 network prefixes with an interface identifier. On interfaces that 48 contain embedded IEEE Identifiers, the interface identifier is 49 typically derived from it. On other interface types, the interface 50 identifier is generated through other means, for example, via random 51 number generation. This document describes an extension to IPv6 52 stateless address autoconfiguration for interfaces whose interface 53 identifier is derived from an IEEE identifier. Use of the extension 54 causes nodes to generate global scope addresses from interface 55 identifiers that change over time, even in cases where the interface 56 contains an embedded IEEE identifier. Changing the interface 57 identifier (and the global scope addresses generated from it) over 58 time makes it more difficult for eavesdroppers and other information 59 collectors to identify when different addresses used in different 60 transactions actually correspond to the same node. 62 Table of Contents 64 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 65 1.1. Conventions used in this document . . . . . . . . . . . . 4 66 1.2. Problem Statement . . . . . . . . . . . . . . . . . . . . 4 67 2. Background . . . . . . . . . . . . . . . . . . . . . . . . . . 5 68 2.1. Extended Use of the Same Identifier . . . . . . . . . . . 5 69 2.2. Address Usage in IPv4 Today . . . . . . . . . . . . . . . 6 70 2.3. The Concern With IPv6 Addresses . . . . . . . . . . . . . 7 71 2.4. Possible Approaches . . . . . . . . . . . . . . . . . . . 8 72 3. Protocol Description . . . . . . . . . . . . . . . . . . . . . 10 73 3.1. Assumptions . . . . . . . . . . . . . . . . . . . . . . . 10 74 3.2. Generation Of Randomized Interface Identifiers . . . . . . 12 75 3.2.1. When Stable Storage Is Present . . . . . . . . . . . . 12 76 3.2.2. In The Absence of Stable Storage . . . . . . . . . . . 13 77 3.2.3. Alternate approaches . . . . . . . . . . . . . . . . . 14 78 3.3. Generating Temporary Addresses . . . . . . . . . . . . . . 14 79 3.4. Expiration of Temporary Addresses . . . . . . . . . . . . 15 80 3.5. Regeneration of Randomized Interface Identifiers . . . . . 16 81 3.6. Deployment Considerations . . . . . . . . . . . . . . . . 17 82 4. Implications of Changing Interface Identifiers . . . . . . . . 19 83 5. Defined Constants . . . . . . . . . . . . . . . . . . . . . . 20 84 6. Future Work . . . . . . . . . . . . . . . . . . . . . . . . . 21 85 7. Security Considerations . . . . . . . . . . . . . . . . . . . 22 86 8. Significant Changes from RFC 3041 . . . . . . . . . . . . . . 23 87 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 24 88 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 25 89 10.1. Normative References . . . . . . . . . . . . . . . . . . . 25 90 10.2. Informative References . . . . . . . . . . . . . . . . . . 25 91 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 27 92 Intellectual Property and Copyright Statements . . . . . . . . . . 28 94 1. Introduction 96 Stateless address autoconfiguration [ADDRCONF] defines how an IPv6 97 node generates addresses without the need for a DHCPv6 server. Some 98 types of network interfaces come with an embedded IEEE Identifier 99 (i.e., a link-layer MAC address), and in those cases stateless 100 address autoconfiguration uses the IEEE identifier to generate a 64- 101 bit interface identifier [ADDRARCH]. By design, the interface 102 identifier is likely to be globally unique when generated in this 103 fashion. The interface identifier is in turn appended to a prefix to 104 form a 128-bit IPv6 address. Note that an IPv6 identifier does not 105 necessarily have to be 64 bits in length, but the algorithm specified 106 in this document is targeted towards 64-bit interface identifiers. 108 All nodes combine interface identifiers (whether derived from an IEEE 109 identifier or generated through some other technique) with the 110 reserved link-local prefix to generate link-local addresses for their 111 attached interfaces. Additional addresses can then be created by 112 combining prefixes advertised in Router Advertisements via Neighbor 113 Discovery [DISCOVERY] with the interface identifier. 115 Not all nodes and interfaces contain IEEE identifiers. In such 116 cases, an interface identifier is generated through some other means 117 (e.g., at random), and the resultant interface identifier may not be 118 globally unique and may also change over time. The focus of this 119 document is on addresses derived from IEEE identifiers, because 120 tracking of individual devices, the concern being addressed here, is 121 possible only in those cases where the interface identifier is 122 globally unique and non-changing. The rest of this document assumes 123 that IEEE identifiers are being used, but the techniques described 124 may also apply to interfaces with other types of globally unique 125 and/or persistent identifiers. 127 This document discusses concerns associated with the embedding of 128 non-changing interface identifiers within IPv6 addresses and 129 describes extensions to stateless address autoconfiguration that can 130 help mitigate those concerns for individual users and in environments 131 where such concerns are significant. Section 2 provides background 132 information on the issue. Section 3 describes a procedure for 133 generating alternate interface identifiers and global scope 134 addresses. Section 4 discusses implications of changing interface 135 identifiers. The term "global scope addresses" is used in this 136 document to collectively refer to "Global unicast addresses" as 137 defined in [ADDRARCH] and "Unique local addresses" as defined in 138 [ULA] 140 1.1. Conventions used in this document 142 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 143 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 144 document are to be interpreted as described in [RFC2119]. 146 1.2. Problem Statement 148 Addresses generated using Stateless address autoconfiguration 149 [ADDRCONF]contain an embedded interface identifier, which remains 150 constant over time. Anytime a fixed identifier is used in multiple 151 contexts, it becomes possible to correlate seemingly unrelated 152 activity using this identifier. 154 The correlation can be performed by 156 o An attacker who is in the path between the node in question and 157 the peer(s) it is communicating to, and can view the IPv6 158 addresses present in the datagrams. 160 o An attacker who can access the communication logs of the peers 161 with which the node has communicated. 163 Since the identifier is embedded within the IPv6 address, which is a 164 fundamental requirement of communication, it cannot be easily hidden. 165 This document proposes a solution to this issue by generating 166 interface identifiers which vary over time. 168 Note that an attacker, who is on path, may be able to perform 169 significant correlation based on 171 o The payload contents of the packets on the wire 173 o The characteristics of the packets such as packet size and timing 175 Use of temporary addresses will not prevent such payload based 176 correlation. 178 2. Background 180 This section discusses the problem in more detail, provides context 181 for evaluating the significance of the concerns in specific 182 environments and makes comparisons with existing practices. 184 2.1. Extended Use of the Same Identifier 186 The use of a non-changing interface identifier to form addresses is a 187 specific instance of the more general case where a constant 188 identifier is reused over an extended period of time and in multiple 189 independent activities. Anytime the same identifier is used in 190 multiple contexts, it becomes possible for that identifier to be used 191 to correlate seemingly unrelated activity. For example, a network 192 sniffer placed strategically on a link across which all traffic to/ 193 from a particular host crosses could keep track of which destinations 194 a node communicated with and at what times. Such information can in 195 some cases be used to infer things, such as what hours an employee 196 was active, when someone is at home, etc. Although it might appear 197 that changing an address regularly in such environments would be 198 desirable to lessen privacy concerns, it should be noted that the 199 network prefix portion of an address also serves as a constant 200 identifier. All nodes at (say) a home, would have the same network 201 prefix, which identifies the topological location of those nodes. 202 This has implications for privacy, though not at the same granularity 203 as the concern that this document addresses. Specifically, all nodes 204 within a home could be grouped together for the purposes of 205 collecting information. If the network contains a very small number 206 of nodes, say just one, changing just the interface identifier will 207 not enhance privacy at all, since the prefix serves as a constant 208 identifier. 210 One of the requirements for correlating seemingly unrelated 211 activities is the use (and reuse) of an identifier that is 212 recognizable over time within different contexts. IP addresses 213 provide one obvious example, but there are more. Many nodes also 214 have DNS names associated with their addresses, in which case the DNS 215 name serves as a similar identifier. Although the DNS name 216 associated with an address is more work to obtain (it may require a 217 DNS query) the information is often readily available. In such 218 cases, changing the address on a machine over time would do little to 219 address the concerns raised in this document, unless the DNS name is 220 changed as well (see Section 4). 222 Web browsers and servers typically exchange "cookies" with each other 223 [COOKIES]. Cookies allow web servers to correlate a current activity 224 with a previous activity. One common usage is to send back targeted 225 advertising to a user by using the cookie supplied by the browser to 226 identify what earlier queries had been made (e.g., for what type of 227 information). Based on the earlier queries, advertisements can be 228 targeted to match the (assumed) interests of the end-user. 230 The use of a constant identifier within an address is of special 231 concern because addresses are a fundamental requirement of 232 communication and cannot easily be hidden from eavesdroppers and 233 other parties. Even when higher layers encrypt their payloads, 234 addresses in packet headers appear in the clear. Consequently, if a 235 mobile host (e.g., laptop) accessed the network from several 236 different locations, an eavesdropper might be able to track the 237 movement of that mobile host from place to place, even if the upper 238 layer payloads were encrypted. 240 2.2. Address Usage in IPv4 Today 242 Addresses used in today's Internet are often non-changing in practice 243 for extended periods of time. In an increasing number of sites, 244 addresses are assigned statically and typically change infrequently. 245 Over the last few years, sites have begun moving away from static 246 allocation to dynamic allocation via DHCP [DHCP]. In theory, the 247 address a client gets via DHCP can change over time, but in practice 248 servers often return the same address to the same client (unless 249 addresses are in such short supply that they are reused immediately 250 by a different node when they become free). Thus, even within sites 251 using DHCP, clients frequently end up using the same address for 252 weeks to months at a time. 254 For home users accessing the Internet over dialup lines, the 255 situation is generally different. Such users do not have permanent 256 connections and are often assigned temporary addresses each time they 257 connect to their ISP. Consequently, the addresses they use change 258 frequently over time and are shared among a number of different 259 users. Thus, an address does not reliably identify a particular 260 device over time spans of more than a few minutes. 262 A more interesting case concerns always-on connections (e.g., cable 263 modems, ISDN, DSL, etc.) that result in a home site using the same 264 address for extended periods of time. This is a scenario that is 265 just starting to become common in IPv4 and promises to become more of 266 a concern as always-on internet connectivity becomes widely 267 available. 269 Finally, it should be noted that nodes that need a (non-changing) DNS 270 name generally have static addresses assigned to them to simplify the 271 configuration of DNS servers. Although Dynamic DNS [DDNS] can be 272 used to update the DNS dynamically, it may not always be available 273 depending on the administrative policy. In addition, changing an 274 address but keeping the same DNS name does not really address the 275 underlying concern, since the DNS name becomes a non-changing 276 identifier. Servers generally require a DNS name (so clients can 277 connect to them), and clients often do as well (e.g., some servers 278 refuse to speak to a client whose address cannot be mapped into a DNS 279 name that also maps back into the same address). Section 4 describes 280 one approach to this issue. 282 2.3. The Concern With IPv6 Addresses 284 The division of IPv6 addresses into distinct topology and interface 285 identifier portions raises an issue new to IPv6 in that a fixed 286 portion of an IPv6 address (i.e., the interface identifier) can 287 contain an identifier that remains constant even when the topology 288 portion of an address changes (e.g., as the result of connecting to a 289 different part of the Internet). In IPv4, when an address changes, 290 the entire address (including the local part of the address) usually 291 changes. It is this new issue that this document addresses. 293 If addresses are generated from an interface identifier, a home 294 user's address could contain an interface identifier that remains the 295 same from one dialup session to the next, even if the rest of the 296 address changes. The way PPP is used today, however, PPP servers 297 typically unilaterally inform the client what address they are to use 298 (i.e., the client doesn't generate one on its own). This practice, 299 if continued in IPv6, would avoid the concerns that are the focus of 300 this document. 302 A more troubling case concerns mobile devices (e.g., laptops, PDAs, 303 etc.) that move topologically within the Internet. Whenever they 304 move they form new addresses for their current topological point of 305 attachment. This is typified today by the "road warrior" who has 306 Internet connectivity both at home and at the office. While the 307 node's address changes as it moves, however, the interface identifier 308 contained within the address remains the same (when derived from an 309 IEEE Identifier). In such cases, the interface identifier can be 310 used to track the movement and usage of a particular machine. For 311 example, a server that logs usage information together with a source 312 addresses, is also recording the interface identifier since it is 313 embedded within an address. Consequently, any data-mining technique 314 that correlates activity based on addresses could easily be extended 315 to do the same using the interface identifier. This is of particular 316 concern with the expected proliferation of next-generation network- 317 connected devices (e.g., PDAs, cell phones, etc.) in which large 318 numbers of devices are in practice associated with individual users 319 (i.e., not shared). Thus, the interface identifier embedded within 320 an address could be used to track activities of an individual, even 321 as they move topologically within the internet. 323 In summary, IPv6 addresses on a given interface generated via 324 Stateless Autoconfiguration contain the same interface identifier, 325 regardless of where within the Internet the device connects. This 326 facilitates the tracking of individual devices (and thus potentially 327 users). The purpose of this document is to define mechanisms that 328 eliminate this issue, in those situations where it is a concern. 330 2.4. Possible Approaches 332 One way to avoid having a static non-changing address is to use 333 DHCPv6[DHCPV6] for obtaining addresses. Section 12 of [DHCPV6] 334 discusses the use of DHCPv6 for the assignment and management of 335 "temporary addresses", which are never renewed and provide the same 336 property of temporary addresses described in this document with 337 regards to the privacy concern. 339 Another approach, compatible with the stateless address 340 autoconfiguration architecture, would be to change the interface 341 identifier portion of an address over time and generate new addresses 342 from the interface identifier for some address scopes. Changing the 343 interface identifier can make it more difficult to look at the IP 344 addresses in independent transactions and identify which ones 345 actually correspond to the same node, both in the case where the 346 routing prefix portion of an address changes and when it does not. 348 Many machines function as both clients and servers. In such cases, 349 the machine would need a DNS name for its use as a server. Whether 350 the address stays fixed or changes has little privacy implication 351 since the DNS name remains constant and serves as a constant 352 identifier. When acting as a client (e.g., initiating 353 communication), however, such a machine may want to vary the 354 addresses it uses. In such environments, one may need multiple 355 addresses: a "public" (i.e., non-secret) server address, registered 356 in the DNS, that is used to accept incoming connection requests from 357 other machines, and a "temporary" address used to shield the identity 358 of the client when it initiates communication. These two cases are 359 roughly analogous to telephone numbers and caller ID, where a user 360 may list their telephone number in the public phone book, but disable 361 the display of its number via caller ID when initiating calls. 363 To make it difficult to make educated guesses as to whether two 364 different interface identifiers belong to the same node, the 365 algorithm for generating alternate identifiers must include input 366 that has an unpredictable component from the perspective of the 367 outside entities that are collecting information. Picking 368 identifiers from a pseudo-random sequence suffices, so long as the 369 specific sequence cannot be determined by an outsider examining 370 information that is readily available or easily determinable (e.g., 371 by examining packet contents). This document proposes the generation 372 of a pseudo-random sequence of interface identifiers via an MD5 hash. 373 Periodically, the next interface identifier in the sequence is 374 generated, a new set of temporary addresses is created, and the 375 previous temporary addresses are deprecated to discourage their 376 further use. The precise pseudo-random sequence depends on both a 377 random component and the globally unique interface identifier (when 378 available), to increase the likelihood that different nodes generate 379 different sequences. 381 3. Protocol Description 383 The goal of this section is to define procedures that: 385 1. Do not result in any changes to the basic behavior of addresses 386 generated via stateless address autoconfiguration [ADDRCONF]. 388 2. Create additional addresses based on a random interface 389 identifier for the purpose of initiating outgoing sessions These 390 "random" or temporary addresses would be used for a short period 391 of time (hours to days) and would then be deprecated. Deprecated 392 address can continue to be used for already established 393 connections, but are not used to initiate new connections. New 394 temporary addresses are generated periodically to replace 395 temporary addresses that expire, with the exact time between 396 address generation a matter of local policy. 398 3. Produce a sequence of temporary global scope addresses from a 399 sequence of interface identifiers that appear to be random in the 400 sense that it is difficult for an outside observer to predict a 401 future address (or identifier) based on a current one and it is 402 difficult to determine previous addresses (or identifiers) 403 knowing only the present one. 405 4. By default, generate a set of addresses from the same 406 (randomized) interface identifier, one address for each prefix 407 for which a global address has been generated via stateless 408 address autoconfiguration. Using the same interface identifier 409 to generate a set of temporary addresses reduces the number of IP 410 multicast groups a host must join. Nodes join the solicited-node 411 multicast address for each unicast address they support, and 412 solicited-node addresses are dependent only on the low-order bits 413 of the corresponding address. This default behaviour was made to 414 address the concern that a node that joins a large number of 415 multicast groups may be required to put its interface into 416 promiscuous mode, resulting in possible reduced performance. 418 A node highly concerned about privacy MAY use different interface 419 identifiers on different prefixes, resulting in a set of global 420 addresses that cannot be easily tied to each other. For example 421 a node MAY create different interface identifiers I1,I2, and I3 422 for use with different prefixes P1,P2, and P3 on the same 423 interface. 425 3.1. Assumptions 427 The following algorithm assumes that each interface maintains an 428 associated randomized interface identifier. When temporary addresses 429 are generated, the current value of the associated randomized 430 interface identifier is used. While the same identifier can be used 431 to create more than one temporary address, the value SHOULD change 432 over time as described in Section 3.5. 434 The algorithm also assumes that for a given temporary address, an 435 implementation can determine the prefix from which it was generated. 436 When a temporary address is deprecated, a new temporary address is 437 generated. The specific valid and preferred lifetimes for the new 438 address are dependent on the corresponding lifetime values set for 439 the prefix from which it was generated. 441 Finally, this document assumes that when a node initiates outgoing 442 communication, temporary addresses can be given preference over 443 public addresses, when the device is configured to do so. 444 [ADDR_SELECT] mandates implementations to provide a mechanism, which 445 allows an application to configure its preference for temporary 446 addresses over public addresses. It also allows for an 447 implementation to prefer temporary addresses by default, so that the 448 connections initiated by the node can use temporary addresses without 449 requiring application-specific enablement. This document also 450 assumes that an API will exist that allows individual applications to 451 indicate whether they prefer to use temporary or public addresses and 452 override the system defaults. 454 3.2. Generation Of Randomized Interface Identifiers 456 We describe two approaches for the generation and maintenance of the 457 randomized interface identifier. The first assumes the presence of 458 stable storage that can be used to record state history for use as 459 input into the next iteration of the algorithm across system 460 restarts. A second approach addresses the case where stable storage 461 is unavailable and there is a need to generate randomized interface 462 identifiers without previous state. 464 The random interface identifier generation algorithm, as described in 465 this document, uses MD5 as the hash algorithm. The node MAY use 466 another algorithm instead of MD5 to produce the random interface 467 identifier. 469 3.2.1. When Stable Storage Is Present 471 The following algorithm assumes the presence of a 64-bit "history 472 value" that is used as input in generating a randomized interface 473 identifier. The very first time the system boots (i.e., out-of-the- 474 box), a random value SHOULD be generated using techniques that help 475 ensure the initial value is hard to guess [RANDOM]. Whenever a new 476 interface identifier is generated, a value generated by the 477 computation is saved in the history value for the next iteration of 478 the algorithm. 480 A randomized interface identifier is created as follows: 482 1. Take the history value from the previous iteration of this 483 algorithm (or a random value if there is no previous value) and 484 append to it the interface identifier generated as described in 485 [ADDRARCH]. 487 2. Compute the MD5 message digest [MD5] over the quantity created in 488 the previous step. 490 3. Take the left-most 64-bits of the MD5 digest and set bit 6 (the 491 left-most bit is numbered 0) to zero. This creates an interface 492 identifier with the universal/local bit indicating local 493 significance only. 495 4. Compare the generated identifier against a list of reserved 496 interface identifiers and to those already assigned to an address 497 on the local device. In the event that an unacceptable 498 identifier has been generated, the node MUST restart the process 499 at step 1 above, using the right-most 64 bits of the MD5 digest 500 obtained in step 2 in place of the history value in step 1. 502 5. Save the generated identifier as the associated randomized 503 interface identifier. 505 6. Take the rightmost 64-bits of the MD5 digest computed in step 2) 506 and save them in stable storage as the history value to be used 507 in the next iteration of the algorithm. 509 MD5 was chosen for convenience, and because its particular properties 510 were adequate to produce the desired level of randomization.The node 511 MAY use another algorithm instead of MD5 to produce the random 512 interface identifier 514 In theory, generating successive randomized interface identifiers 515 using a history scheme as above has no advantages over generating 516 them at random. In practice, however, generating truly random 517 numbers can be tricky. Use of a history value is intended to avoid 518 the particular scenario where two nodes generate the same randomized 519 interface identifier, both detect the situation via DAD, but then 520 proceed to generate identical randomized interface identifiers via 521 the same (flawed) random number generation algorithm. The above 522 algorithm avoids this problem by having the interface identifier 523 (which will often be globally unique) used in the calculation that 524 generates subsequent randomized interface identifiers. Thus, if two 525 nodes happen to generate the same randomized interface identifier, 526 they should generate different ones on the followup attempt. 528 3.2.2. In The Absence of Stable Storage 530 In the absence of stable storage, no history value will be available 531 across system restarts to generate a pseudo-random sequence of 532 interface identifiers. Consequently, the initial history value used 533 above SHOULD be generated at random. A number of techniques might be 534 appropriate. Consult [RANDOM] for suggestions on good sources for 535 obtaining random numbers. Note that even though machines may not 536 have stable storage for storing a history value, they will in many 537 cases have configuration information that differs from one machine to 538 another (e.g., user identity, security keys, serial numbers, etc.). 539 One approach to generating a random initial history value in such 540 cases is to use the configuration information to generate some data 541 bits (which may remain constant for the life of the machine, but will 542 vary from one machine to another), append some random data and 543 compute the MD5 digest as before. 545 3.2.3. Alternate approaches 547 Note that there are other approaches to generate random interface 548 identifiers, albeit with different goals and applicability. One such 549 approach is CGA [CGA], which generates a random interface identifier 550 based on the public key of the node. The goal of CGAs is to prove 551 ownership of an address and to prevent spoofing and stealing of 552 existing IPv6 addresses. They are used for securing neighbor 553 discovery using [SEND]. The CGA random interface identifier 554 generation algorithm may not be suitable for privacy addresses 555 because of the following properties 557 o It requires the node to have a public key. This means that the 558 node can still be identified by its public key 560 o The random interface identifier process is computationally 561 intensive and hence discourages frequent regeneration 563 3.3. Generating Temporary Addresses 565 [ADDRCONF] describes the steps for generating a link-local address 566 when an interface becomes enabled as well as the steps for generating 567 addresses for other scopes. This document extends [ADDRCONF] as 568 follows. When processing a Router Advertisement with a Prefix 569 Information option carrying a global scope prefix for the purposes of 570 address autoconfiguration (i.e., the A bit is set), the node MUST 571 perform the following steps: 573 1. Process the Prefix Information Option as defined in [ADDRCONF], 574 either creating a new public address or adjusting the lifetimes 575 of existing addresses, both public and temporary. If a received 576 option will extend the lifetime of a public address, the 577 lifetimes of temporary addresses should be extended, subject to 578 the overall constraint that no temporary addresses should ever 579 remain "valid" or "preferred" for a time longer than 580 (TEMP_VALID_LIFETIME - DESYNC_FACTOR) or (TEMP_PREFERRED_LIFETIME 581 - DESYNC_FACTOR) respectively. The configuration variables 582 TEMP_VALID_LIFETIME and TEMP_PREFERRED_LIFETIME correspond to 583 approximate target lifetimes for temporary addresses. 585 2. One way an implementation can satisfy the above constraints is to 586 associate with each temporary address a creation time (called 587 CREATION_TIME) that indicates the time at which the address was 588 created. When updating the preferred lifetime of an existing 589 temporary address, it would be set to expire at whichever time is 590 earlier: the time indicated by the received lifetime or 591 (CREATION_TIME + TEMP_PREFERRED_LIFETIME - DESYNC_FACTOR). A 592 similar approach can be used with the valid lifetime. 594 3. When a new public address is created as described in [ADDRCONF], 595 the node SHOULD also create a new temporary address. 597 4. When creating a temporary address, the lifetime values MUST be 598 derived from the corresponding prefix as follows: 600 * Its Valid Lifetime is the lower of the Valid Lifetime of the 601 public address or TEMP_VALID_LIFETIME 603 * Its Preferred Lifetime is the lower of the Preferred Lifetime 604 of the public address or TEMP_PREFERRED_LIFETIME - 605 DESYNC_FACTOR. 607 5. A temporary address is created only if this calculated Preferred 608 Lifetime is greater than REGEN_ADVANCE time units. In 609 particular, an implementation MUST NOT create a temporary address 610 with a zero Preferred Lifetime. 612 6. New temporary addresses MUST be created by appending the 613 interface's current randomized interface identifier to the prefix 614 that was received. 616 7. The node MUST Perform duplicate address detection (DAD) on the 617 generated temporary address. If DAD indicates the address is 618 already in use, the node MUST generate a new randomized interface 619 identifier as described in Section 3.2 above, and repeat the 620 previous steps as appropriate up to TEMP_IDGEN_RETRIES times. If 621 after TEMP_IDGEN_RETRIES consecutive attempts no non-unique 622 address was generated, the node MUST log a system error and MUST 623 NOT attempt to generate temporary addresses for that interface. 624 Note that DAD MUST be performed on every unicast address 625 generated from this randomized interface identifier. 627 3.4. Expiration of Temporary Addresses 629 When a temporary address becomes deprecated, a new one MUST be 630 generated. This is done by repeating the actions described in 631 Section 3.3, starting at step 3). Note that, except for the 632 transient period when a temporary address is being regenerated, in 633 normal operation at most one temporary address per prefix should be 634 in a non-deprecated state at any given time on a given interface. 635 Note that if a temporary address becomes deprecated as result of 636 processing a Prefix Information Option with a zero Preferred 637 Lifetime, then a new temporary address MUST NOT be generated. To 638 ensure that a preferred temporary address is always available, a new 639 temporary address SHOULD be regenerated slightly before its 640 predecessor is deprecated. This is to allow sufficient time to avoid 641 race conditions in the case where generating a new temporary address 642 is not instantaneous, such as when duplicate address detection must 643 be run. The node SHOULD start the address regeneration process 644 REGEN_ADVANCE time units before a temporary address would actually be 645 deprecated. 647 As an optional optimization, an implementation MAY remove a 648 deprecated temporary address that is not in use by applications or 649 upper-layers as detailed in Section 6. 651 3.5. Regeneration of Randomized Interface Identifiers 653 The frequency at which temporary addresses changes depends on how a 654 device is being used (e.g., how frequently it initiates new 655 communication) and the concerns of the end user. The most egregious 656 privacy concerns appear to involve addresses used for long periods of 657 time (weeks to months to years). The more frequently an address 658 changes, the less feasible collecting or coordinating information 659 keyed on interface identifiers becomes. Moreover, the cost of 660 collecting information and attempting to correlate it based on 661 interface identifiers will only be justified if enough addresses 662 contain non-changing identifiers to make it worthwhile. Thus, having 663 large numbers of clients change their address on a daily or weekly 664 basis is likely to be sufficient to alleviate most privacy concerns. 666 There are also client costs associated with having a large number of 667 addresses associated with a node (e.g., in doing address lookups, the 668 need to join many multicast groups, etc.). Thus, changing addresses 669 frequently (e.g., every few minutes) may have performance 670 implications. 672 Nodes following this specification SHOULD generate new temporary 673 addresses on a periodic basis. This can be achieved automatically by 674 generating a new randomized interface identifier at least once every 675 (TEMP_PREFERRED_LIFETIME - REGEN_ADVANCE - DESYNC_FACTOR) time units. 676 As described above, generating a new temporary address REGEN_ADVANCE 677 time units before a temporary address becomes deprecated produces 678 addresses with a preferred lifetime no larger than 679 TEMP_PREFERRED_LIFETIME. The value DESYNC_FACTOR is a random value 680 (different for each client) that ensures that clients don't 681 synchronize with each other and generate new addresses at exactly the 682 same time. When the preferred lifetime expires, a new temporary 683 address MUST be generated using the new randomized interface 684 identifier. 686 Because the precise frequency at which it is appropriate to generate 687 new addresses varies from one environment to another, implementations 688 SHOULD provide end users with the ability to change the frequency at 689 which addresses are regenerated. The default value is given in 690 TEMP_PREFERRED_LIFETIME and is one day. In addition, the exact time 691 at which to invalidate a temporary address depends on how 692 applications are used by end users. Thus, the suggested default 693 value of one week (TEMP_VALID_LIFETIME) may not be appropriate in all 694 environments. Implementations SHOULD provide end users with the 695 ability to override both of these default values. 697 Finally, when an interface connects to a new link, a new randomized 698 interface identifier SHOULD be generated immediately together with a 699 new set of temporary addresses. If a device moves from one ethernet 700 to another, generating a new set of temporary addresses from a 701 different randomized interface identifier ensures that the device 702 uses different randomized interface identifiers for the temporary 703 addresses associated with the two links, making it more difficult to 704 correlate addresses from the two different links as being from the 705 same node. The node MAY follow any process available to it, to 706 determine that the link change has occurred. One such process is 707 described by Detecting Network Attachment [DNA]. 709 3.6. Deployment Considerations 711 Devices implementing this specification MUST provide a way for the 712 end user to explicitly enable or disable the use of temporary 713 addresses. In addition, a site might wish to disable the use of 714 temporary addresses in order to simplify network debugging and 715 operations. Consequently, implementations SHOULD provide a way for 716 trusted system administrators to enable or disable the use of 717 temporary addresses. 719 Additionally, sites might wish to selectively enable or disable the 720 use of temporary addresses for some prefixes. For example, a site 721 might wish to disable temporary address generation for "Unique local" 722 [ULA] prefixes while still generating temporary addresses for all 723 other global prefixes. Another site might wish to enable temporary 724 address generation only for the prefixes 2001::/16 and 2002::/16 725 while disabling it for all other prefixes. To support this behavior, 726 implementations SHOULD provide a way to enable and disable generation 727 of temporary addresses for specific prefix subranges. This per- 728 prefix setting SHOULD override the global settings on the node with 729 respect to the specified prefix subranges. Note that the pre-prefix 730 setting can be applied at any granularity, and not necessarily on a 731 per subnet basis. 733 The use of temporary addresses may cause unexpected difficulties with 734 some applications. As described below, some servers refuse to accept 735 communications from clients for which they cannot map the IP address 736 into a DNS name. In addition, some applications may not behave 737 robustly if temporary addresses are used and an address expires 738 before the application has terminated, or if it opens multiple 739 sessions, but expects them to all use the same addresses. 740 Consequently, the use of temporary addresses SHOULD be disabled by 741 default in order to minimize potential disruptions. Individual 742 applications, which have specific knowledge about the normal duration 743 of connections, MAY override this as appropriate. 745 If a very small number of nodes (say only one) use a given prefix for 746 extended periods of time, just changing the interface identifier part 747 of the address may not be sufficient to ensure privacy, since the 748 prefix acts as a constant identifier. The procedures described in 749 this document are most effective when the prefix is reasonably non 750 static or is used by a fairly large number of nodes. 752 4. Implications of Changing Interface Identifiers 754 The IPv6 addressing architecture goes to some lengths to ensure that 755 interface identifiers are likely to be globally unique where easy to 756 do so. The widespread use of temporary addresses may result in a 757 significant fraction of Internet traffic not using addresses in which 758 the interface identifier portion is globally unique. Consequently, 759 usage of the algorithms in this document may complicate providing 760 such a future flexibility, if global uniqueness is necessary. 762 The desires of protecting individual privacy versus the desire to 763 effectively maintain and debug a network can conflict with each 764 other. Having clients use addresses that change over time will make 765 it more difficult to track down and isolate operational problems. 766 For example, when looking at packet traces, it could become more 767 difficult to determine whether one is seeing behavior caused by a 768 single errant machine, or by a number of them. 770 Some servers refuse to grant access to clients for which no DNS name 771 exists. That is, they perform a DNS PTR query to determine the DNS 772 name, and may then also perform an AAAA query on the returned name to 773 verify that the returned DNS name maps back into the address being 774 used. Consequently, clients not properly registered in the DNS may 775 be unable to access some services. As noted earlier, however, a 776 node's DNS name (if non-changing) serves as a constant identifier. 777 The wide deployment of the extension described in this document could 778 challenge the practice of inverse-DNS-based "authentication," which 779 has little validity, though it is widely implemented. In order to 780 meet server challenges, nodes could register temporary addresses in 781 the DNS using random names (for example a string version of the 782 random address itself). 784 Use of the extensions defined in this document may complicate 785 debugging and other operational troubleshooting activities. 786 Consequently, it may be site policy that temporary addresses should 787 not be used. Consequently, implementations MUST provide a method for 788 the end user or trusted administrator to override the use of 789 temporary addresses. 791 5. Defined Constants 793 Constants defined in this document include: 795 TEMP_VALID_LIFETIME -- Default value: 1 week. Users should be able 796 to override the default value. 798 TEMP_PREFERRED_LIFETIME -- Default value: 1 day. Users should be 799 able to override the default value. 801 REGEN_ADVANCE -- 5 seconds 803 MAX_DESYNC_FACTOR -- 10 minutes. Upper bound on DESYNC_FACTOR. 805 DESYNC_FACTOR -- A random value within the range 0 - 806 MAX_DESYNC_FACTOR. It is computed once at system start (rather than 807 each time it is used) and must never be greater than 808 (TEMP_VALID_LIFETIME - REGEN_ADVANCE). 810 TEMP_IDGEN_RETRIES -- Default value: 3 812 6. Future Work 814 An implementation might want to keep track of which addresses are 815 being used by upper layers so as to be able to remove a deprecated 816 temporary address from internal data structures once no upper layer 817 protocols are using it (but not before). This is in contrast to 818 current approaches where addresses are removed from an interface when 819 they become invalid [ADDRCONF], independent of whether or not upper 820 layer protocols are still using them. For TCP connections, such 821 information is available in control blocks. For UDP-based 822 applications, it may be the case that only the applications have 823 knowledge about what addresses are actually in use. Consequently, an 824 implementation generally will need to use heuristics in deciding when 825 an address is no longer in use. 827 The determination as to whether to use public versus temporary 828 addresses can in some cases only be made by an application. For 829 example, some applications may always want to use temporary 830 addresses, while others may want to use them only in some 831 circumstances or not at all. Suitable API extensions will likely 832 need to be developed to enable individual applications to indicate 833 with sufficient granularity their needs with regards to the use of 834 temporary addresses. Recommendations on DNS practices to avoid the 835 problem described in Section 4 when reverse DNS lookups fail may be 836 needed. [DNSOP] contains a more detailed discussion of the DNS 837 related issues. 839 While this document discusses ways of obscuring a user's permanent IP 840 address, the method described is believed to be ineffective against 841 sophisticated forms of traffic analysis. To increase effectiveness, 842 one may need to consider use of more advanced techniques, such as 843 Onion Routing [ONION]. 845 7. Security Considerations 847 Ingress filtering has been and is being deployed as a means of 848 preventing the use of spoofed source addresses in Distributed Denial 849 of Service(DDoS) attacks. In a network with a large number of nodes, 850 new temporary addresses are created at a fairly high rate. This 851 might make it difficult for ingress filtering mechanisms to 852 distinguish between legitimately changing temporary addresses and 853 spoofed source addresses, which are "in-prefix"(They use a 854 topologically correct prefix and non-existent interface ID). This 855 can be addressed by using access control mechanisms on a per address 856 basis on the network egress point. 858 8. Significant Changes from RFC 3041 860 This section summarizes the changes in this document relative to RFC 861 3041 that an implementer of RFC 3041 should be aware of. 863 1. Excluded certain interface identifiers from the range of 864 acceptable interface identifiers. Interface IDs such as those 865 for reserved anycast addresses [RFC], etc. 867 2. Added a configuration knob that provides the end user with a way 868 to enable or disable the use of temporary addresses on a per- 869 prefix basis. 871 3. Added a check for denial of service attacks using low valid 872 lifetimes in router advertisements 874 4. DAD is now run on all temporary addresses, not just the first one 875 generated from an interface identifier. 877 5. Changed the default setting for usage of temporary addresses to 878 be disabled. 880 6. The node is now allowed to generate different interface 881 identifiers for different prefixes, if it so desires. 883 7. The algorithm used for generating random interface identifiers is 884 no longer restricted to just MD5 886 8. Reduced default number of retries to from and added a 887 configuration variable 889 9. RA processing algorithm is no longer included in the document, 890 and is replaced by a reference to [ADDRCONF]. 892 9. Acknowledgements 894 The authors would like to acknowledge the contributions of the ipv6 895 working group and, in particular, Ran Atkinson, Matt Crawford, Steve 896 Deering, Allison Mankin, Peter Bieringer, Jari Arkko, Pekka Nikander, 897 Pekka Savola, Francis Dupont, Brian Haberman, Tatuya Jinmei and 898 Margaret Wasserman for their detailed comments. 900 10. References 902 10.1. Normative References 904 [ADDRARCH] 905 Hinden, R. and S. Deering, "Internet Protocol Version 6 906 (IPv6) Addressing Architecture", RFC 3513, April 2003. 908 [ADDRCONF] 909 Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 910 Address Autoconfiguration", draft-ietf-ipv6-rfc2462bis-07 911 (work in progress), December 2004. 913 [DISCOVERY] 914 Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 915 "Neighbor Discovery for IP version 6 (IPv6)", 916 draft-ietf-ipv6-2461bis-02 (work in progress), 917 February 2005. 919 [MD5] Rivest, R., "The MD5 Message-Digest Algorithm", RFC 1321, 920 April 1992. 922 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 923 Requirement Levels", RFC 2119, March 1997. 925 10.2. Informative References 927 [ADDR_SELECT] 928 Draves, R., "Default Address Selection for Internet 929 Protocol version 6 (IPv6)", RFC 3484, February 2003. 931 [CGA] Aura, T., "Cryptographically Generated Addresses (CGA)", 932 RFC 3972, March 2005. 934 [COOKIES] Kristol, D. and L. Montulli, "HTTP State Management 935 Mechanism", RFC 2965, October 2000. 937 [DDNS] Vixie, P., Thomson, S., Rekhter, Y., and J. Bound, 938 "Dynamic Updates in the Domain Name System (DNS UPDATE)", 939 RFC 2136, April 1997. 941 [DHCP] Droms, R., "Dynamic Host Configuration Protocol", 942 RFC 2131, March 1997. 944 [DHCPV6] Droms, R., Bound, J., Volz, B., Lemon, T., Perkins, C., 945 and M. Carney, "Dynamic Host Configuration Protocol for 946 IPv6 (DHCPv6)", RFC 3315, July 2003. 948 [DNA] Choi, J. and G. Daley, "Detecting Network Attachment in 949 IPv6 Goals", draft-ietf-dna-goals-04 (work in progress), 950 December 2004. 952 [DNSOP] Durand, A., Ihren, J., and P. Savola, "Operational 953 Considerations and Issues with IPv6 DNS", 954 draft-ietf-dnsop-ipv6-dns-issues-10 (work in progress), 955 October 2004. 957 [ONION] Reed, MGR., Syverson, PFS., and DMG. Goldschlag, "Proxies 958 for Anonymous Routing", Proceedings of the 12th Annual 959 Computer Security Applications Conference, San Diego, CA, 960 December 1996. 962 [RANDOM] Eastlake, D., Crocker, S., and J. Schiller, "Randomness 963 Recommendations for Security", RFC 1750, December 1994. 965 [SEND] Arkko, J., Kempf, J., Zill, B., and P. Nikander, "SEcure 966 Neighbor Discovery (SEND)", RFC 3971, March 2005. 968 [ULA] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 969 Addresses", draft-ietf-ipv6-unique-local-addr-09 (work in 970 progress), January 2005. 972 Authors' Addresses 974 Thomas Narten 975 IBM Corporation 976 P.O. Box 12195 977 Research Triangle Park, NC 978 USA 980 Email: narten@raleigh.ibm.com 982 Richard Draves 983 Microsoft Research 984 One Microsoft Way 985 Redmond, WA 986 USA 988 Email: richdr@microsoft.com 990 Suresh Krishnan 991 Ericsson Research 992 8400 Decarie Blvd. 993 Town of Mount Royal, QC 994 Canada 996 Email: suresh.krishnan@ericsson.com 998 Intellectual Property Statement 1000 The IETF takes no position regarding the validity or scope of any 1001 Intellectual Property Rights or other rights that might be claimed to 1002 pertain to the implementation or use of the technology described in 1003 this document or the extent to which any license under such rights 1004 might or might not be available; nor does it represent that it has 1005 made any independent effort to identify any such rights. Information 1006 on the procedures with respect to rights in RFC documents can be 1007 found in BCP 78 and BCP 79. 1009 Copies of IPR disclosures made to the IETF Secretariat and any 1010 assurances of licenses to be made available, or the result of an 1011 attempt made to obtain a general license or permission for the use of 1012 such proprietary rights by implementers or users of this 1013 specification can be obtained from the IETF on-line IPR repository at 1014 http://www.ietf.org/ipr. 1016 The IETF invites any interested party to bring to its attention any 1017 copyrights, patents or patent applications, or other proprietary 1018 rights that may cover technology that may be required to implement 1019 this standard. Please address the information to the IETF at 1020 ietf-ipr@ietf.org. 1022 Disclaimer of Validity 1024 This document and the information contained herein are provided on an 1025 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 1026 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 1027 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 1028 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE 1029 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 1030 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 1032 Copyright Statement 1034 Copyright (C) The Internet Society (2006). This document is subject 1035 to the rights, licenses and restrictions contained in BCP 78, and 1036 except as set forth therein, the authors retain all their rights. 1038 Acknowledgment 1040 Funding for the RFC Editor function is currently provided by the 1041 Internet Society.