idnits 2.17.1 draft-bellovin-keyroll2385-04.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 15. -- Found old boilerplate from RFC 3978, Section 5.5 on line 300. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 311. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 318. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 324. ** 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 issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the 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 (January 7, 2007) is 6317 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 1323 (Obsoleted by RFC 7323) ** Obsolete normative reference: RFC 2385 (Obsoleted by RFC 5925) -- Obsolete informational reference (is this intentional?): RFC 3036 (Obsoleted by RFC 5036) -- Obsolete informational reference (is this intentional?): RFC 3682 (Obsoleted by RFC 5082) -- Obsolete informational reference (is this intentional?): RFC 4634 (Obsoleted by RFC 6234) Summary: 5 errors (**), 0 flaws (~~), 1 warning (==), 10 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group S. Bellovin 3 Internet-Draft Columbia University 4 Intended status: Informational January 7, 2007 5 Expires: July 11, 2007 7 Key Change Strategies for TCP-MD5 8 draft-bellovin-keyroll2385-04.txt 10 Status of this Memo 12 By submitting this Internet-Draft, each author represents that any 13 applicable patent or other IPR claims of which he or she is aware 14 have been or will be disclosed, and any of which he or she becomes 15 aware will be disclosed, in accordance with Section 6 of BCP 79. 17 Internet-Drafts are working documents of the Internet Engineering 18 Task Force (IETF), its areas, and its working groups. Note that 19 other groups may also distribute working documents as Internet- 20 Drafts. 22 Internet-Drafts are draft documents valid for a maximum of six months 23 and may be updated, replaced, or obsoleted by other documents at any 24 time. It is inappropriate to use Internet-Drafts as reference 25 material or to cite them other than as "work in progress." 27 The list of current Internet-Drafts can be accessed at 28 http://www.ietf.org/ietf/1id-abstracts.txt. 30 The list of Internet-Draft Shadow Directories can be accessed at 31 http://www.ietf.org/shadow.html. 33 This Internet-Draft will expire on July 11, 2007. 35 Copyright Notice 37 Copyright (C) The Internet Society (2007). 39 Abstract 41 The TCP-MD5 option is most commonly used to secure BGP sessions 42 between routers. However, changing the long-term key is difficult, 43 since the change needs to be synchronized between different 44 organizations. We describe single-ended strategies that will permit 45 (mostly) unsynchronized key changes. 47 1. Introduction 49 The TCP-MD5 option [RFC2385] is most commonly used to secure BGP 50 sessions between routers. However, changing the long-term key is 51 difficult, since the change needs to be synchronized between 52 different organizations. Worse yet, if the keys are out of sync, it 53 may break the connection between the two routers, rendering repair 54 attempts difficult. 56 The proper solution involves some sort of key management protocol. 57 Apart from the complexity of such things, RFC 2385 was not written 58 with key changes in mind. In particular, there is no KeyID field in 59 the option, which means that even a key management protocol would run 60 into the same problem. 62 Fortunately, a heuristic permits key change despite this protocol 63 deficiency. The change can be installed unilaterally at one end of a 64 connection; it is fully compatible with the existing protocol. 66 1.1. Terminology 68 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 69 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 70 document are to be interpreted as described in [RFC2119]. 72 2. The Algorithm 74 Separate algorithms are necessary for transmission and reception. 75 Reception is easier; we explain it first. 77 2.1. Reception 79 A receiver has a list of valid keys. Each key has a (conceptual) 80 timestamp associated with it. When a segment arrives, each key is 81 tried in turn. The segment is discarded if and only if it cannot be 82 validated by any key in the list. 84 In principle, there is no need to test keys in any particular order. 85 For performance reasons, though, a simple MRU strategy -- try the 86 last valid key first -- should work well. More complex mechanisms, 87 such as examining the TCP sequence number of an arriving segment to 88 see whether it fits in a hole, are almost certainly unnecessary. On 89 the other hand, validating that a received segment is putatively 90 legal, by checking its sequence number against the advertised window, 91 can help avoid denial of service attacks. 93 The newest key that has successfully validated a segment is marked as 94 the "preferred" key; see below. 96 Implicit in this scheme is the assumption that older keys will 97 eventually be unneeded and can be removed. Accordingly, 98 implementations SHOULD provide an indication of when a key was last 99 used successfully. 101 2.2. Transmission 103 Transmission is more complex, because the sender does not know which 104 keys can be accepted at the far end. Accordingly, the conservative 105 strategy is to delay using any new keys for a considerable amount of 106 time, probably measured in days. This time interval is the amount of 107 asynchronicity the parties wish to permit; it is agreed-upon out of 108 band and configured manually. 110 Some automation is possible, however. If a key has been used 111 successfully to validate an incoming segment, clearly the other side 112 knows it. Accordingly, any key marked as "preferred" by the 113 receiving part of a stack SHOULD be used for transmissions. 115 A sophisticated implementation could try alternate keys if the TCP 116 retransmission counter gets too high. (This is analogous to dead 117 gateway detection.) In particular, if a key change has just been 118 attempted but such segments are not acknowledged, it is reasonable to 119 fall back to the previous key and issue an alert of some sort. 120 Similarly, an implementation with a new but unused key could 121 occasionally try to use it, much in the way that TCP implementations 122 probe closed windows. Doing this avoids the "silent host" problem 123 discusssed in Section 3.1. This should be done at a moderately slow 124 rate. 126 Note that there is an ambiguity when an acknowledgment is received 127 for a segment transmitted with two different keys. The TCP Timestamp 128 option [RFC1323] can be used for disambiguation. 130 3. Operations 132 3.1. Single-Ended Operations 134 Suppose only one end of the connection has this algorithm 135 implemented. The new key is provisioned on that system, with a start 136 time far in the future -- sufficiently far, in fact, that it will not 137 be used spontaneously. After the key is ready, the other end is 138 notified, out-of-band, that a key change can commence. 140 At some point, the other end is upgraded. Because it does not have 141 multiple keys available, it will start using the new key immediately 142 for its transmission, and will drop all segments that use the old 143 key. As soon as it tries to transmit, the upgraded side will 144 designate the new key as preferred, and will use it for all of its 145 transmissions. Note specifically that this will include 146 retransmissions of any segments rejected because they used the old 147 key. 149 There is a problem if the unchanged machine is a "silent host" -- a 150 host that has nothing to say, and hence does not transmit. The best 151 way to avoid this is for an upgraded machine to try a variety of keys 152 in event of repeated unacknowledged packets, and to probe for new 153 unused keys durign silent periods, as discussed in Section 2.2. 154 Alternatively, application-level KeepAlive messages may be used, to 155 ensure that neither end of the connection is completely silent. See, 156 for example, Section 4.4 of [RFC4271] or Section 3.5.4 of [RFC3036]. 158 3.2. Double-Ended Operations 160 Double-ended operations are similar, save that both sides deploy the 161 new key at about the same time. One should be configured to start 162 using the new key at a point where it is reasonably certain that the 163 other side would have it installed, too. Assuming that that has in 164 fact happened, the new key will be marked "preferred" on both sides. 166 3.3. Monitoring 168 As noted, implementations should monitor when a key was last used for 169 transmission or reception. Any monitoring mechanism can be used; 170 most likely, it will be one or both of a MIB object or objects and 171 the vendor's usual command-line mechanism for displaying data of this 172 type. Regardless, the network operations center should keep track of 173 this. When a new key has been used successfully for both 174 transmission and reception for a reasonable amount of time -- the 175 exact value isn't crucial, but it should probably be longer than 176 twice the maximum segment lifetime -- the old key can be marked for 177 deletion. There is an implicit assumption here that there will not 178 be substantial overlap in the usage period of such keys; monitoring 179 systems should look for any such anomalies, of course. 181 4. Moving Forward 183 As implied in Section 1, this is an interim strategy, intended to 184 make TCP-MD5 operationally usable today. We do not suggest or 185 recommend it as a long-term solution. In this section, we make some 186 suggestions about the design of a future TCP authentication option. 188 The first and most obvious change is to replace keyed MD5 with a 189 stronger MAC [RFC4278]. Today, HMAC-SHA1 [RFC4634] is the preferred 190 choice, though others such as UMAC [RFC4418] should be considered as 191 well. 193 A new authentication option should contain some form of Key ID field. 194 Such an option would permit uambiguous identification of which key 195 was used to create the MAC for a given segment, sparing the receiver 196 the need to engage in the sort of heuristics described here. A Key 197 ID is useful with both manual and automatic key management. (Note 198 carefully that we do not prescribe any particular Key ID mechanism 199 here. Rather, we are stating a requirement: there must be a simple, 200 low-cost way to select a particular key, and it must be possible to 201 rekey without tearing down long-lived connections.) 203 Finally, an automated key management mechanism should be defined. 204 The general reasoning for that is set forth in [RFC4107]; specific 205 issues pertaining to BGP and TCP are given in [RFC3562]. 207 5. Security Considerations 209 In theory, accepting multiple keys simultaneously makes life easier 210 for an attacker. In practice, if the recommendations in [RFC3562] 211 are followed, this should not be a problem. 213 New keys must be communicated securely. Specifically, new key 214 messages must be kept confidential and must be properly 215 authenticated. 217 Having multiple keys makes CPU denial of service attacks easier. 218 This suggests that keeping the overlap period reasonably short is a 219 good idea. In addition, the Generalized TTL Security Mechanism 220 [RFC3682], if applicable to the local topology, can help. Note that 221 most of the time, only one key will exist; virtually all of the 222 remaining time there will be only two keys in existence. 224 6. IANA Considerations 226 There are no IANA actions required. The TCP-MD5 option number is 227 defined in [RFC2385], and is currently listed by IANA. 229 7. Acknowledgments 231 I'd like to thank Ron Bonica, Randy Bush, Ross Callon, Rob Evans, 232 Eric Rescorla, and Sam Weiler for their comments and inspiration. 234 8. References 236 8.1. Normative 238 [RFC1323] Jacobson, V., Braden, B., and D. Borman, "TCP Extensions 239 for High Performance", RFC 1323, May 1992. 241 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 242 Requirement Levels", BCP 14, RFC 2119, March 1997. 244 [RFC2385] Heffernan, A., "Protection of BGP Sessions via the TCP MD5 245 Signature Option", RFC 2385, August 1998. 247 8.2. Informative 249 [RFC3036] Andersson, L., Doolan, P., Feldman, N., Fredette, A., and 250 B. Thomas, "LDP Specification", RFC 3036, January 2001. 252 [RFC3562] Leech, M., "Key Management Considerations for the TCP MD5 253 Signature Option", RFC 3562, July 2003. 255 [RFC3682] Gill, V., Heasley, J., and D. Meyer, "The Generalized TTL 256 Security Mechanism (GTSM)", RFC 3682, February 2004. 258 [RFC4107] Bellovin, S. and R. Housley, "Guidelines for Cryptographic 259 Key Management", BCP 107, RFC 4107, June 2005. 261 [RFC4271] Rekhter, Y., Li, T., and S. Hares, "A Border Gateway 262 Protocol 4 (BGP-4)", RFC 4271, January 2006. 264 [RFC4278] Bellovin, S. and A. Zinin, "Standards Maturity Variance 265 Regarding the TCP MD5 Signature Option (RFC 2385) and the 266 BGP-4 Specification", RFC 4278, January 2006. 268 [RFC4418] Krovetz, T., "UMAC: Message Authentication Code using 269 Universal Hashing", RFC 4418, March 2006. 271 [RFC4634] Eastlake, D. and T. Hansen, "US Secure Hash Algorithms 272 (SHA and HMAC-SHA)", RFC 4634, August 2006. 274 Author's Address 276 Steven M. Bellovin 277 Columbia University 278 1214 Amsterdam Avenue 279 MC 0401 280 New York, NY 10027 281 US 283 Phone: +1 212 939 7149 284 Email: bellovin@acm.org 286 Full Copyright Statement 288 Copyright (C) The Internet Society (2007). 290 This document is subject to the rights, licenses and restrictions 291 contained in BCP 78, and except as set forth therein, the authors 292 retain all their rights. 294 This document and the information contained herein are provided on an 295 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 296 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 297 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 298 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE 299 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 300 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 302 Intellectual Property 304 The IETF takes no position regarding the validity or scope of any 305 Intellectual Property Rights or other rights that might be claimed to 306 pertain to the implementation or use of the technology described in 307 this document or the extent to which any license under such rights 308 might or might not be available; nor does it represent that it has 309 made any independent effort to identify any such rights. Information 310 on the procedures with respect to rights in RFC documents can be 311 found in BCP 78 and BCP 79. 313 Copies of IPR disclosures made to the IETF Secretariat and any 314 assurances of licenses to be made available, or the result of an 315 attempt made to obtain a general license or permission for the use of 316 such proprietary rights by implementers or users of this 317 specification can be obtained from the IETF on-line IPR repository at 318 http://www.ietf.org/ipr. 320 The IETF invites any interested party to bring to its attention any 321 copyrights, patents or patent applications, or other proprietary 322 rights that may cover technology that may be required to implement 323 this standard. Please address the information to the IETF at 324 ietf-ipr@ietf.org. 326 Acknowledgment 328 Funding for the RFC Editor function is provided by the IETF 329 Administrative Support Activity (IASA).