idnits 2.17.1 draft-iab-strint-report-03.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 Introduction section. ** The document seems to lack a both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. RFC 2119 keyword, line 793: '...s are labelled as MUST in the standard...' RFC 2119 keyword, line 794: '... rather than MAY. This is sometimes...' Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (September 4, 2015) is 3119 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Looks like a reference, but probably isn't: '1' on line 1073 -- Looks like a reference, but probably isn't: '2' on line 1075 -- Looks like a reference, but probably isn't: '3' on line 1077 -- Looks like a reference, but probably isn't: '4' on line 1079 -- Looks like a reference, but probably isn't: '5' on line 1081 -- Looks like a reference, but probably isn't: '6' on line 1083 -- Looks like a reference, but probably isn't: '7' on line 1085 -- Looks like a reference, but probably isn't: '8' on line 1087 -- Looks like a reference, but probably isn't: '9' on line 1089 -- Looks like a reference, but probably isn't: '10' on line 1091 -- Looks like a reference, but probably isn't: '11' on line 1093 -- Looks like a reference, but probably isn't: '12' on line 1095 -- Looks like a reference, but probably isn't: '13' on line 1097 -- Looks like a reference, but probably isn't: '14' on line 1146 -- Looks like a reference, but probably isn't: '15' on line 1148 -- Looks like a reference, but probably isn't: '16' on line 1150 -- Looks like a reference, but probably isn't: '17' on line 1154 -- Looks like a reference, but probably isn't: '18' on line 1171 -- Looks like a reference, but probably isn't: '19' on line 1183 -- Looks like a reference, but probably isn't: '20' on line 1186 -- Looks like a reference, but probably isn't: '21' on line 1198 -- Looks like a reference, but probably isn't: '22' on line 1201 -- Looks like a reference, but probably isn't: '23' on line 1210 -- Looks like a reference, but probably isn't: '24' on line 1222 -- Looks like a reference, but probably isn't: '25' on line 1235 -- Looks like a reference, but probably isn't: '26' on line 1247 -- Looks like a reference, but probably isn't: '27' on line 1251 -- Looks like a reference, but probably isn't: '28' on line 1259 -- Looks like a reference, but probably isn't: '29' on line 1260 -- Looks like a reference, but probably isn't: '30' on line 1260 -- Looks like a reference, but probably isn't: '31' on line 1270 -- Looks like a reference, but probably isn't: '32' on line 1305 -- Looks like a reference, but probably isn't: '33' on line 1312 -- Looks like a reference, but probably isn't: '34' on line 1313 -- Looks like a reference, but probably isn't: '35' on line 1314 == Unused Reference: 'Arkko1' is defined on line 971, but no explicit reference was found in the text -- Obsolete informational reference (is this intentional?): RFC 6962 (Obsoleted by RFC 9162) Summary: 2 errors (**), 0 flaws (~~), 2 warnings (==), 37 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group S. Farrell 3 Internet-Draft Trinity College, Dublin 4 Intended status: Informational R. Wenning 5 Expires: March 7, 2016 B. Bos 6 W3C 7 M. Blanchet 8 Viagenie 9 H. Tschofenig 10 ARM Ltd. 11 September 4, 2015 13 Report from the Strengthening the Internet (STRINT) workshop 14 draft-iab-strint-report-03 16 Abstract 18 The Strengthening the Internet (STRINT) workshop assembled one 19 hundred participants in London for two days in early 2014 to discuss 20 how the technical community, and in particular the IETF and the W3C, 21 should react to Pervasive Monitoring and more generally how to 22 strengthen the Internet in the face of such attacks. The discussions 23 covered issues of terminology, the role of user interfaces, classes 24 of mitigation, some specific use cases, transition strategies 25 (including opportunistic encryption), and more. The workshop ended 26 with a few high-level recommendations, which it is believed could be 27 implemented and which could help strengthen the Internet. This is 28 the report of that workshop. 30 Note that this document is a report on the proceedings of the 31 workshop. The views and positions documented in this report are 32 those of the workshop participants and do not necessarily reflect IAB 33 views and positions. 35 Status of This Memo 37 This Internet-Draft is submitted in full conformance with the 38 provisions of BCP 78 and BCP 79. 40 Internet-Drafts are working documents of the Internet Engineering 41 Task Force (IETF). Note that other groups may also distribute 42 working documents as Internet-Drafts. The list of current Internet- 43 Drafts is at http://datatracker.ietf.org/drafts/current/. 45 Internet-Drafts are draft documents valid for a maximum of six months 46 and may be updated, replaced, or obsoleted by other documents at any 47 time. It is inappropriate to use Internet-Drafts as reference 48 material or to cite them other than as "work in progress." 49 This Internet-Draft will expire on March 7, 2016. 51 Copyright Notice 53 Copyright (c) 2015 IETF Trust and the persons identified as the 54 document authors. All rights reserved. 56 This document is subject to BCP 78 and the IETF Trust's Legal 57 Provisions Relating to IETF Documents 58 (http://trustee.ietf.org/license-info) in effect on the date of 59 publication of this document. Please review these documents 60 carefully, as they describe your rights and restrictions with respect 61 to this document. 63 Table of Contents 65 1. Context . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 66 2. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 67 3. Workshop goals . . . . . . . . . . . . . . . . . . . . . . . 4 68 4. Workshop structure . . . . . . . . . . . . . . . . . . . . . 5 69 5. Topics . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 70 6. After the workshop . . . . . . . . . . . . . . . . . . . . . 20 71 7. IANA considerations . . . . . . . . . . . . . . . . . . . . . 21 72 8. Security considerations . . . . . . . . . . . . . . . . . . . 21 73 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 21 74 Appendix A. Logistics . . . . . . . . . . . . . . . . . . . . . 25 75 Appendix B. Agenda . . . . . . . . . . . . . . . . . . . . . . . 26 76 Appendix C. Workshop chairs & program committee . . . . . . . . 28 77 Appendix D. Participants . . . . . . . . . . . . . . . . . . . . 28 78 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 33 80 1. Context 82 The Vancouver IETF plenary[vancouverplenary] concluded that Pervasive 83 Monitoring (PM) represents an attack on the Internet, and the IETF 84 has begun to carry out the more obvious actions required to try to 85 handle this attack. However, there are additional much more complex 86 questions arising that need further consideration before any 87 additional concrete plans can be made. 89 The W3C [1] and IAB [2] therefore decided to host a workshop [3] on 90 the topic of "Strengthening the Internet Against Pervasive 91 Monitoring" before IETF 89 [4] in London in March 2014. The 92 FP7-funded STREWS [5] project organised the STRINT workshop on behalf 93 of the IAB and W3C. 95 The main workshop goal was to discuss what can be done, especially by 96 the two standards organisations IETF and W3C, against PM, both for 97 existing Internet protocols (HTTP/1, SMTP, etc.) and for new ones 98 (WebRTC, HTTP/2, etc.). 100 The starting point for the workshop was the existing IETF consensus 101 that PM is an attack[RFC7258] (the text of which had achieved IEFF 102 consensus at the time of the workshop, even though the RFC had yet to 103 be published). 105 2. Summary 107 The workshop was well attended (registration closed when the maximum 108 capacity of 100 was reached, but more than 150 expressed a desire to 109 register) and several people (about 165 at the maximum) listened to 110 the streaming audio. The submitted papers (67 in total) were 111 generally of good quality and all were published, except for a few 112 where authors who couldn't take part in the workshop preferred not to 113 publish. 115 The chairs of the workshop summarised the workshop in the final 116 session in the form of the following recommendations: 118 1. Well-implemented cryptography can be effective against PM and 119 will benefit the Internet if used more, despite its cost, which 120 is steadily decreasing anyway. 122 2. Traffic analysis also needs to be considered, but is less well 123 understood in the Internet community: relevant research and 124 protocol mitigations such as data minimisation need to be better 125 understood. 127 3. Work should continue on progressing the PM threat model 128 draft[I-D.barnes-pervasive-problem] discussed in the workshop. 130 4. Later, the IETF may be in a position to start to develop an 131 update to BCP 72 [RFC3552], most likely as a new RFC enhancing 132 that BCP and dealing with recommendations on how to mitigate PM 133 and how to reflect that in IETF work. 135 5. The term "Opportunistic" has been widely used to refer to a 136 possible mitigation strategy for PM. The community need to 137 document definition(s) for this term, as it is being used 138 differently by different people and in different contexts. We 139 may also be able to develop a cookbook-like set of related 140 protocol techniques for developers. Since the workshop, the 141 IETF's security area has taken up this work, most recently 142 favouring the generic term "Opportunistic Security" (OS) 143 [I-D.kent-opportunistic-security]. Subsequent work on this 144 topic resulted in the publication of a definition of OS in 145 [RFC7435]. 147 6. The technical community could do better in explaining the real 148 technical downsides related to PM in terms that policy makers 149 can understand. 151 7. Many User Interfaces (UIs) could be better in terms of how they 152 present security state, though this is a significantly hard 153 problem. There may be benefits if certain dangerous choices 154 were simply not offered anymore. But that could require 155 significant co-ordination among competing software makers, 156 otherwise some will be considered "broken" by users. 158 8. Further discussion is needed on ways to better integrate UI 159 issues into the processes of IETF and W3C. 161 9. Examples of good software configurations that can be cut-and- 162 paste'd for popular software, etc., can help. This is not 163 necessarily standards work, but maybe the standards 164 organisations can help and can work with those developing such 165 package-specific documentation. 167 10. The IETF and W3C can do more so that default ("out-of-the-box") 168 settings for protocols better protect security and privacy. 170 11. Captive portals [6] (and some firewalls, too) can and should be 171 distinguished from real man-in-the-middle attacks. This might 172 mean establishing common conventions with makers of such 173 middleboxes, but might also need new protocols. However, the 174 incentives for deploying such new middlebox features might not 175 align. 177 3. Workshop goals 179 As stated, the STRINT workshop started from the position [RFC7258] 180 that PM is an attack. While some dissenting voices are expected and 181 need to be heard, that was the baseline assumption for the workshop, 182 and the high-level goal was to provide more consideration of that and 183 how it ought to affect future work within the IETF and W3C. 185 At the next level down the goals of the STRINT workshop were to: 187 o Discuss and hopefully come to agreement among the participants on 188 concepts in PM for both threats and mitigation, e.g., 189 "opportunistic" as the term applies to cryptography. 191 o Discuss the PM threat model, and how that might be usefully 192 documented for the IETF at least, e.g., via an update to BCP72. 193 [7] 195 o Discuss and progress common understanding in the trade-offs 196 between mitigating and suffering PM. 198 o Identify weak links in the chain of Web security architecture with 199 respect to PM. 201 o Identify potential work items for the IETF, IAB, IRTF and W3C that 202 would help mitigate PM. 204 o Discuss the kinds of action outside the IETF/W3C context might 205 help those done within the IETF/W3C. 207 4. Workshop structure 209 The workshop structure was designed to maximise discussion time. 210 There were no direct presentations of submitted papers. Instead, the 211 moderators of each session summarised topics that the Technical 212 Programme Committee (TPC) had agreed based on the submitted papers. 213 These summary presentations took at most 50% of the session and 214 usually less. 216 Because the papers would not be presented during the workshop, 217 participants were asked to read and discuss the papers beforehand, at 218 least those relevant to their fields of interest. (To help people 219 choose papers to read, authors were asked to provide short 220 abstracts.) 222 Most of the sessions had two moderators, one to lead the discussion, 223 while the other managed the queue of people who wanted to speak. 224 This worked well: everybody got a chance to speak and each session 225 still ended on time. 227 The penultimate session consisted of break-outs (which turned out to 228 be the most productive sessions of all, most likely simply due to the 229 smaller numbers of people involved). The subjects for the break-outs 230 were agreed during the earlier sessions and just before the break-out 231 session the participants collectively determined who would attend 232 which. 234 5. Topics 236 The following sections contain summaries of the various sessions. 237 See the minutes (see Appendix B) for more details. 239 5.1. Opening session 241 The first session discussed the goals of the workshop. Possible 242 approaches to improving security in the light of pervasive monitoring 243 include a critical look at what metadata is actually required, 244 whether old (less secure) devices can be replaced with new ones, what 245 are "low-hanging fruit" (issues that can be handled quickly and 246 easily), and what level of security is "good enough": a good solution 247 may be one that is good for 90% of people or 90% of organisations. 249 Some participants felt that standards are needed so that people can 250 see if their systems conform to a certain level of security, and easy 251 to remember names are needed for those standards, so that a buyer can 252 immediately see that a product "conforms to the named intended 253 standard." 255 5.2. Threats 257 One difference between "traditional" attacks and pervasive monitoring 258 is modus-operandi of the attacker: typically, one determines what 259 resources an attacker might want to target and at what cost and then 260 one defends against that threat. But a pervasive attacker has no 261 specific targets, other than to collect everything he can. The 262 calculation of the cost of losing resources vs. the cost of 263 protecting them is thus different. And unlike someone motivated to 264 make money, a PM attacker may not be concerned at the cost of the 265 attack (or may even prefer a higher cost, for "empire building" 266 reasons"). 268 The terminology used to talk about threats has to be chosen carefully 269 (this was a common theme in several sessions), because we need to 270 explain to people outside the technical community what they need to 271 do or not do. For example, authentication of endpoints doesn't so 272 much "protect against" man-in-the-middle (MITM) attacks as make them 273 visible. The attacker can still attack, but it does not remain 274 invisible while he does so. Somebody on either end of the 275 conversation needs to react to the alert from the system: stop the 276 conversation or find a different channel. 278 Paradoxically, while larger sites such as Facebook, Yahoo, and Google 279 supervise the security of their respective services more than other 280 smaller sites, such large sites offer a much more attractive target 281 to attack. Avoiding overuse of such repositories for private or 282 sensitive information may be a useful measure that increases the cost 283 of collecting for a pervasive attacker. This is sometimes called the 284 target-dispersal approach. 286 Lack of interoperability between systems can lead to poorly thought 287 out work-arounds and compromises that may themselves introduce 288 vulnerabilities. And thus improving interoperability needs to be 289 high on the list of priorities of standards makers and even more for 290 implementers. Of course, testing, such as interop testing, is at 291 some level, part of the process of IETF and W3C; and W3C is currently 292 increasing its testing efforts. 294 5.3. Increase usage of security tools 296 The first session on Communication Security (COMSEC) tools looked at 297 the question why existing security tools aren't used more. 299 The example of the public key infrastructure used to secure HTTP is 300 informative. One problem is that certificate authorities (CAs) may 301 issue a certificate for any domain. Thus a single compromised CA can 302 be used in combination with a MITM to impersonate any server. 303 Moreover, ongoing administration, including requesting, paying for 304 and installing new certificates, has proven over time to be an 305 insurmountable barrier for many web site administrators, leading them 306 not to bother to secure their systems. 308 Some ideas were discussed for improving the CA system, e.g., via 309 cross-certification of CAs and by means of "certificate 310 transparency": a public, permanent log of who issued which 311 certificate. [RFC6962] 313 Using other models than the hierarchical certificate model (as 314 alternative or in combination) may also help. PGP demonstrates a 315 model known as a "web of trust" where people verify the public key of 316 the people they meet. Because there is no innate transitive trust in 317 PGP, it is appropriate only for small scale uses; an example being a 318 team of people working on a project. 320 Yet another model is "trust on first use" (TOFU). This is used quite 321 effectively by SSH [RFC4252]. On the first connection, one has no 322 way to verify that the received public key belongs to the server one 323 is contacting, therefore, the key is accepted without further 324 verification. But on the subsequent connections, one can verify that 325 the received key is the same key as the first time. So a MITM has to 326 be there on all connections, including the first, otherwise it will 327 be detected by a key mismatch. 329 This works well for SSH, because people typically use SSH to 330 communicate with a small number of servers over and over again. And, 331 if they want, they may find a separate channel to get the public key 332 (or its fingerprint). It may also work for Web servers used by small 333 groups (the server of a sports club, a department of a company, 334 etc.), but probably works less well for public servers that are 335 visited once or a few times or for large services where many servers 336 may be used. 338 A similar proposal [draft-ietf-websec-key-pinning] for an HTTP header 339 introduces an aspect of TOFU into HTTP: Key pinning tells HTTP 340 clients that for a certain time after receiving this certificate, 341 they should not expect the certificate to change. If it does, even 342 if the new certificate looks valid, the client should assume a 343 security breach. 345 Session Initiation Protocol (SIP) [RFC3261] can require several 346 different intermediaries in different stages of the communication to 347 deal with NAT traversal and to handle policy. While both hop-by-hop 348 and end-to-end encryption are specified, in practice many SIP 349 providers disable these functions. The reasons for disabling end-to- 350 end security here are understandable: to overcome lack of 351 interoperability they often need to change protocol headers and 352 modify protocol data. Some workshop participants argued that SIP 353 would never have taken off if it hadn't been possible for providers 354 to monitor and interfere in communications in this way. Of course, 355 that means an attacker can listen in just as easily. 357 A new protocol for peer-to-peer communication of video and audio (and 358 potentially other data) is WebRTC. WebRTC re-uses many of the same 359 architectural concepts as SIP, but there is a reasonable chance that 360 it can do better in terms of protecting users: The people 361 implementing the protocols and offering the service have different 362 goals and interests. In particular, the first implementers are 363 browser makers, who may have different business models from other 364 more traditional Voice over IP providers. 366 XMPP[RFC6120] suffers from yet another problem. It has encryption 367 and authentication, and the OTR ("off the record") extension even 368 provides what is called Perfect Forward Secrecy (PFS, compromising 369 the current communication never gives an attacker enough information 370 to decrypt past communications that he may have recorded). But, in 371 practice, many people don't use XMPP at all, but rather Skype, 372 WhatsApp or other instant-messaging tools with unknown or no 373 security. The problem here seems to be one of user awareness. And 374 though OTR does provide security, it is not well integrated with XMPP 375 and nor is it available as a core feature of XMPP clients. 377 To increase usage of existing solutions, some tasks can be 378 identified, though how those map to actions for e.g. IETF/W3C is not 379 clear: 381 o Improvements to the certificate system, such as certificate 382 transparency (CT). 384 o Making it easier (cheaper, quicker) for system administrators to 385 deploy secure solutions. 387 o Improve awareness of the risks. Identify which communities 388 influence which decisions and what is the appropriate message for 389 each. 391 o Provide an upgrade path that doesn't break existing systems or 392 require that everybody upgrade at the same time. Opportunistic 393 Security may be one model for that. 395 5.4. Policy issues and non-technical actions 397 Previous sessions already concluded that the problem isn't just 398 technical, such as getting the right algorithms in the standards, 399 fixing interoperability, or educating implementers and systems 400 administrators. There are user interface issues and education issues 401 too. And there are also legal issues and policy issues for 402 governments. 404 It appears that the public in general demand more privacy and 405 security (e.g., for their children) but are also pessimistic about 406 getting that. They trust that somebody assures that nothing bad 407 happens to them, but they also expect to be spied on all the time. 409 (Perceived) threats of terrorism gave governments a reason to allow 410 widespread surveillance, far beyond what may previously have been 411 considered dangerous for freedom. 413 In this environment, the technical community will have a hard time 414 developing and deploying technologies that fully counter PM, which 415 means there has to be action in the social and political spheres, 416 too. 418 Technology isn't the only thing that can make life harder for 419 attackers. Government-sponsored PM is indirectly affected by trade 420 agreements and treaties and thus it makes sense to lobby for those to 421 be as privacy-friendly as possible. 423 Court cases on the grounds of human rights can also influence policy, 424 especially if they reach, for example, the European Court of Human 425 Rights. 427 In medicine and law, it is common to have ethics committees, not so 428 in software. Should standards bodies such as IETF and W3C have an 429 ethics committee? Standards such as the Geolocation API 430 [w3c-geo-api] have gotten scrutiny from privacy experts, but only in 431 an ad-hoc manner. (W3C has permanent groups to review standards for 432 accessibility and internationalisation. It also has a Privacy group, 433 but that currently doesn't do the same kind of systematic reviews.) 435 As the Internet Draft draft-barnes-pervasive-problem-00 (included as 436 paper 44 [8]) explains, PM doesn't just monitor the networks, but 437 also attacks at the endpoints, turning organisations or people into 438 (willing, unwilling, or unwitting) collaborators. One technical 439 means of protection is thus to design protocols such that there are 440 fewer potential collaborators, e.g., a provider of cloud storage 441 cannot hand over plaintext for content that is encrypted with a key 442 he doesn't have, and cannot hand over names if his client is 443 anonymous. 445 It is important to distinguish between PM and fighting crime. PM is 446 an attack, but a judge ordering the surveillance of a suspected 447 criminal is not. The latter, often abbreviated in this context as LI 448 (for Lawful Intercept), is outside the scope of this workshop. 450 5.5. Improving the tools 452 An earlier session discussed why existing COMSEC tools weren't used 453 more. This second session on COMSEC therefore discussed what 454 improvements and/or new tools were needed. 456 Discussion at the workshop indicated that an important meta-tool for 457 improving existing security technology could be Opportunistic 458 Security (OS) [I-D.kent-opportunistic-security]. The idea is that 459 software is enhanced with a module that tries to encrypt 460 communications when it detects that the other end also has the same 461 capability but otherwise lets the communication continue in the old 462 way. The detailed definition of OS is being discussed by the IETF 463 security area at the time of this workshop [saag]. 465 OS would protect against a passive eavesdropper but should also allow 466 for endpoint authentication to protect against an active attacker (a 467 MITM). As OS spreads, more and more communications would be 468 encrypted (and hopefully authenticated) and thus there is less and 469 less for an eavesdropper to collect. 471 Of course, an implementation of OS could give a false sense of 472 security as well: some connections are encrypted, some are not. A 473 user might see something like a padlock icon in browsers, but there 474 was agreement at the workshop that such user interface features ought 475 not be changed because OS is being used. 477 There is also the possibility that a MITM intercepts the reply from a 478 server that says "yes, I can do encryption" and removes it, causing 479 the client to fall back to an unencrypted protocol. Mitigations 480 against this can be to have other channels of finding out a server's 481 capabilities and remembering that a server could do encryption 482 previously. 484 There is also, again, a terminology problem. The technical 485 descriptions of OS talk about "silent fail" when a connection 486 couldn't be encrypted and has to fall back to the old, unencrypted 487 protocol. Actually, it's not a fail; it's no worse than it was 488 before. A successful encryption would rather be a "silent 489 improvement." 491 That raises the question of the UI: How do you explain to a user what 492 their security options are, and, in case an error occurs, how do you 493 explain the implications of the various responses? 495 The people working on encryption are mathematicians and engineers, 496 and typically not the same people who know about UI. We need to 497 involve the experts. We also need to distinguish between usability 498 of the UI, user understanding, and user experience. For an 499 e-commerce site, e.g., it is not just important that the user's data 500 is technically safe, but also that he feels secure. Otherwise he 501 still won't buy anything. 503 When talking about users, we also need to distinguish the end user 504 (who we typically think about when we talk about UI) from the server 505 administrators and other technical people involved in enabling a 506 connection. When something goes wrong (e.g., the user's software 507 detects an invalid certificate), the message usually goes to the end 508 user. But he isn't necessarily the person who can do something about 509 it. E.g., if the problem is a certificate that expired yesterday, 510 the options for the user are to break the connection (the safe 511 choice, but it means he can't get his work done) or continue anyway 512 (there could be a MITM...). The server administrator, on the other 513 hand, could actually solve the problem. 515 Encryption and authentication have a cost, in terms of setting them 516 up, but also in terms of the time it takes for software to do the 517 calculations. The set-up cost can be reduced with sensible defaults, 518 predefined profiles and cut-and-paste configurations. And for some 519 connections, authentication without encryption could be enough, in 520 the case that the data doesn't need to be kept secret, but it is 521 important to know that it is the real data. Most mail user agents 522 (UA) already provide independent options for encryption and signing, 523 but Web servers only support authentication if the connection is also 524 encrypted. 526 On the other hand, as e-mail also shows, it is difficult for users to 527 understand what encryption and authentication do separately. 529 And it also has to be kept in mind that encrypting only the 530 "sensitive" data and not the rest decreases the cost for an attacker, 531 too: It becomes easy to know which connections are worth attacking. 532 Selective field confidentiality is also more prone to lead to 533 developer error, as not all developers will know the provenance of 534 values to be processed. 536 One problem with the TOFU model as used by SSH (see explanation 537 above) is that it lacks a solution for key continuity: When a key is 538 changed (which can happen, e.g., when a server is replaced or the 539 software upgraded), there is no way to inform the client. (In 540 practice, people use other means, such as calling people on the phone 541 or asking their colleagues in the office, but that doesn't scale and 542 doesn't always happen either.) An improvement in the SSH protocol 543 could thus be a way to transfer a new key to a client in a safe way. 545 5.6. Hiding metadata 547 Encryption and authentication help protect the content of messages. 548 Correctly implemented encryption is very hard to crack. (To get the 549 content, an attacker would rather attempt to steal the keys, corrupt 550 the encoding software, or get the content via a collaborator.) But 551 encrypting the content doesn't hide the fact that you are 552 communicating. This metadata (who talks to whom, when and for how 553 long) is often as interesting as the content itself, and in some 554 cases the size and timing of messages is even an accurate predictor 555 of the content. So how to stop an attacker from collecting metadata, 556 given that much of that data is actually needed by routers and other 557 services to deliver the message to the right place? 559 It is useful to distinguish different kinds of metadata: explicit (or 560 metadata proper) and implicit (sometimes called traffic data). 561 Implicit metadata is things that can be derived from a message or are 562 necessary for its delivery, such as the destination address, the 563 size, the time, or the frequency with which messages pass. Explicit 564 metadata is things like quality ratings, provenance or copyright 565 data: data about the data, useful for an application, but not 566 required to deliver the data to its endpoint. 568 A system such as Tor hides much of the metadata by passing through 569 several servers, encrypting all the data except that which a 570 particular server needs to see. Each server thus knows which server 571 a message came from and where he has to send it to, but cannot know 572 where the previous server got it from or where the next server is 573 instructed to send it. However, deliberately passing through 574 multiple servers makes the communication slower than taking the most 575 direct route and increases the amount of traffic the network as a 576 whole has to process. 578 There are three kinds of measures that can be taken to make metadata 579 harder to get: aggregation, contraflow and multipath (see paper 4 580 [9]). New protocols should be designed such that these measures are 581 not inadvertently disallowed, e.g., because the design assumes that 582 the whole of a conversation passes through the same route. 584 "Aggregation" means collecting conversations from multiple sources 585 into one stream. E.g., if HTTP connections pass through a proxy, all 586 the conversations appear to come from the proxy instead of from their 587 original sources. (This assumes that telltale information in the 588 headers is stripped by the proxy, or that the connection is 589 encrypted.) It also works in the other direction: if multiple Web 590 sites are hosted on the same server, an attacker cannot see which of 591 those Web sites a user is reading. (This assumes that the name of 592 the site is in the path info of the URL and not in the domain name, 593 otherwise watching DNS queries can still reveal the name.) 595 "Contraflow" means routing a conversation via one or more other 596 servers than the normal route, e.g., by using a tunnel (e.g., with 597 SSH or a VPN) to another server. Tor is an example of this. An 598 attacker must watch more routes and do more effort to correlate 599 conversations. (Again, this assumes that there is no telltale 600 information left in the messages that leave the tunnel.) 602 "Multipath" splits up a single conversation (or a set of related 603 conversations) and routes the parts in different ways. E.g., sending 604 a request via a satellite link and receiving the response via a land 605 line; or starting a conversation on a cellular link and continuing it 606 via Wi-Fi. This again increases the cost for an attacker, who has to 607 monitor and correlate multiple networks. 609 Protecting metadata automatically with technology at a lower layer 610 than the application layer is difficult. The applications themselves 611 need to pass less data, e.g., use anonymous temporary handles instead 612 of permanent identifiers. There is often no real need for people to 613 use the same identifier on different computers (smartphone, desktop, 614 etc.) other than that the application they use was designed that way. 616 One thing that can be done relatively easily in the short term is to 617 go through existing protocols to check what data they send that isn't 618 really necessary. One candidate mentioned for such a study was XMPP. 620 "Fingerprinting" is the process of distinguishing different senders 621 of messages based on metadata: Clients can be recognised (or at least 622 grouped) because their messages always have a combination of features 623 that other clients do not have. Reducing redundant metadata and 624 reducing the number of optional features in a protocol reduces the 625 variation between clients and thus makes fingerprinting harder. 627 Traffic analysis is a research discipline that produces sometimes 628 surprising findings, which are little known among protocol 629 developers. Some collections of results are 631 o A selected bibliography on anonymity [10] by the Free Haven 632 Project, 634 o The yearly Symposium on Privacy Enhancing Technologies (PETS) 635 [11], and 637 o The yearly Workshop on Privacy in the Electronic Society (WPES) 638 [12]. 640 Techniques that deliberately change the timing or size of messages, 641 such as padding, can also help reduce fingerprinting. Obviously, 642 they make conversations slower and/or use more bandwidth, but in some 643 cases that is not an issue, e.g., if the conversation is limited by 644 the speed of a human user anyway. HTTP/2 has a built-in padding 645 mechanism. However, it is not so easy to use these techniques well, 646 and not actually make messages easier to recognise rather than 647 harder. 649 Different users in different contexts may have different security 650 needs, so maybe the priority can be a user choice (if that can be 651 done without making high-security users stand out from other users). 652 Although many people would not understand what their choices are, 653 some do, such as political activists or journalists. 655 5.7. Deployment, intermediaries and middleboxes 657 Secure protocols have often been designed in the past for end-to-end 658 security: Intermediaries cannot read or modify the messages. This is 659 the model behind TLS for example. 661 In practice, however, people have more or less valid reasons to 662 insist on intermediaries: companies filtering incoming and outgoing 663 traffic for viruses or other reasons, giving priority to certain 664 communications or caching to reduce bandwidth. 666 In the presence of end-to-end encryption and authentication, these 667 intermediaries have two choices: use fake certificates to impersonate 668 the endpoints or have access to the private keys of the endpoints. 669 The former is a MITM attack that is difficult to distinguish from a 670 more malicious one, and the latter obviously decreases the security 671 of the endpoints by copying supposedly protected data and 672 concentrating such data in a single place. 674 As mentioned in Section 5.2 above, aggregation of data in a single 675 place makes that place an attractive target. And in the case of PM 676 even if the data is not concentrated physically in one place, it is 677 under control of a single legal entity that can be made into a 678 collaborator. 680 The way Web communication with TLS typically works is that the client 681 authenticates the server, but the server does not authenticate the 682 client at the TLS layer. (If the client needs to be identified, that 683 is mainly done at the application layer via passwords or cookies.) 684 Thus the presence of a MITM (middlebox) could be detected by the 685 client (because of the incorrect certificate), but not by the server. 686 If the client doesn't immediately close the connection (which they do 687 not in many cases), the server may thus disclose information that the 688 user would rather not have disclosed. 690 One widespread example of middleboxes is captive portals, as found on 691 the Wi-Fi hotspots in hotels, airports, etc. Even the hotspots 692 offering free access often intercept communications to redirect the 693 user to a login or policy page. 695 When the communication they intercept is, e.g., the automatic update 696 of your calendar program or a chat session, the redirect obviously 697 doesn't work: these applications don't know how to display a Web 698 page. With the increasing use of applications, it may be a while 699 before the user actually opens a browser. The flood of error 700 messages may also have as a result that the user no longer reads the 701 errors, allowing an actual malicious attack to go unnoticed. 703 Some operating systems now come with heuristics that try to recognise 704 captive portals and either automatically login or show their login 705 page in a separate application. (But some hotspot providers 706 apparently don't want automatic logins and actually reverse- 707 engineered the heuristics to try and fool them.) 709 It seems some protocol is missing in this case. Captive portals 710 shouldn't have to do MITM attacks to be noticed. Something like an 711 extension to DHCP that tells a connecting device about the login page 712 may help, although that still doesn't solve the problem for devices 713 that do not have a Web browser, such as game consoles or SIP phones. 714 HTTP response code 511 (defined in [RFC6585]) is another attempt at a 715 partial solution (Partial, because it can only work at the moment the 716 user uses a browser to connect to a Web site and doesn't use HTTPS). 718 A practical problem with deployment of such a protocol may be that 719 many such captive portals are very old and never updated. The hotel 720 staff only knows how to reboot the system and as long as it works, 721 the hotel has no incentive to buy a new one. As evidence of this: 722 how many such systems require you to get a password and the ticket 723 shows the price as zero? This is typically because the owner doesn't 724 know how to reconfigure the hotspot, but he does know how to change 725 the price in his cash register. 727 5.8. Break-out 1 - research 729 Despite some requests earlier in the workshop, the research break-out 730 did not discuss clean-slate approaches. The challenge was rather 731 that the relationship between security research and standardisation 732 needs improvement. Research on linkability is not yet well known in 733 the IETF. But the other side of the coin needs improvement too: 734 While doing protocol design, standardisation should indicate what 735 specific problems are in need of more research. 737 The break-out then made a non-exclusive list of topics that are in 738 need of further research: 740 o The interaction of compression and encryption as demonstrated by 741 the CRIME SSL/TLS vulnerability [13] 743 o A more proactive deprecation of algorithms based on research 744 results 746 o Mitigation for return-oriented programming attacks 748 o How to better obfuscate so called "metadata" 750 o How to make the existence of traffic and their endpoints stealthy 752 5.9. Break-out 2 - clients 754 Browsers are the first clients one thinks of when talking about 755 encrypted connections, authentication and certificates, but there are 756 many others. 758 Other common case of "false" alarms for MITM (after captive portals) 759 include expired and mis-configured certificates. This is quite 760 common in intranets, when the sysadmin hasn't bothered updating a 761 certificate and rather tells his handful of users to just "click 762 continue." The problem is on the one hand that users may not 763 understand the difference between this case and the same error 764 message when they connect to a server outside the company, and on the 765 other hand that the incorrect certificate installed by the sysadmin 766 is not easily distinguishable from an incorrect certificate from a 767 MITM. The error message is almost the same and the user may just 768 click continue again. 770 One way to get rid of such certificates is if client software no 771 longer offers the option to continue after a certificate error. That 772 requires that all major clients (such as browsers) change their 773 behaviour at the same time, otherwise the first one to do so will be 774 considered broken by users, because the others still work. Also it 775 requires a period in which that software gives increasingly strong 776 warnings about the cut-off date after which the connection will fail 777 with this certificate. 779 Yet another source of error messages is self-signed certificates. 780 Such certificates are actually only errors for sites that are not 781 expected to have them. If a message about a self-signed certificate 782 appears when connecting to Facebook or Google, you're clearly not 783 connected to the real Facebook or Google. But for a personal Website 784 it shouldn't cause such scary warnings. There may be ways to improve 785 the explanations in the error message and provide an easy way to 786 verify the certificate (by e-mail, over the phone or some other 787 channel) and trust it. 789 5.10. Break-out 3 - on by default 791 One step in improving security is to require the relevant features, 792 in particular encryption and authentication, to be implemented in 793 compliant products: The features are labelled as MUST in the standard 794 rather than MAY. This is sometimes referred to as Mandatory To 795 Implement (MTI) and is the current practice for IETF 796 protocols[RFC3365]. 798 But that may not be enough to counter PM. It may be that the 799 features are there, but not used, because only very knowledgeable 800 users or sysadmins turn them on. Or it may be that implementations 801 do not actually follow the MTI parts of specifications. Or it may be 802 that some security features are implemented but interoperability for 803 those doesn't really work. Or, even worse, it may be that protocol 804 designers have only followed the letter of the MTI best practice and 805 not its spirit, with the result that security features are hard to 806 use or make deployment harder. One can thus argue that such features 807 should be defined to be on by default. 809 Going further one might argue that these features should not even be 810 options, i.e., there should be no way to turn them off. This is 811 sometimes called Mandatory To Use (MTU). 813 The question raised at this session was for what protocols on-by- 814 default is appropriate, and how can one explain to the developers of 815 such protocols that it is needed? 817 There would of course be resistance to MTU security from implemeters 818 and deployments that practice deep packet inspection (DPI) and also 819 perhaps from some governments. On the other hand, there may also be 820 governments that outlaw protocols without proper encryption. 822 This break-out concluded that there could be value in attempting to 823 document a new Best Current Practice for the IETF that moves from the 824 current MTI position to one where security features are on-by- 825 default. Some of the workshop participants expressed interest in 826 authoring a draft for such a new BCP and progressing that through the 827 IETF consensus process (where it would no doubt be controversial). 829 5.11. Break-out 4 - measurement 831 There was a small break-out on the idea of measurement as a way to 832 encourage or gamify the increased use of security mechanisms. 834 5.12. Break-out 5 - opportunistic 836 This break out considered the use of the term "opportunistic" as it 837 applies to cryptographic security and attempted to progress the work 838 towards arriving at an agreed-upon definition for use of that term, 839 at it applies to IETF and W3C work. 841 While various terms had been used, with many people talking about 842 opportunistic encryption, that usage was felt to be problematic both 843 because it conflicted with the use of the same term in [RFC4322] and 844 because it was being used differently in different parts of the 845 community. 847 At the session it was felt that the term "opportunistic keying" was 848 better, but as explained above subsequent list discussion resulted in 849 a move to the term "Opportunistic Security" (OS). 851 Aside from terminology, disussion focused on the use of Diffie- 852 Hellman (D-H) key exchange as the preferred mechanism of OS, with 853 fall back to cleartext if D-H doesn't succeed as a counter for 854 passive attacks. 856 There was also of course the desire to be able to easily escalate 857 from countering passive attacks to also handling endpoint 858 authentication and thereby also countering MITM attacks. 860 Making OS visible to users was again considered to be undesirable, as 861 users could not be expected to distinguish between cleartext, OS and 862 (one-sided or mutual) endpoint authentication. 864 Finally, it was noted that it may take some effort to establish how 865 middleboxes might affect OS at different layers and that OS really is 866 not suitable as the only migitation to use for high-sensitivity 867 sessions such as financial transactions. 869 5.13. Unofficial Transport/Routing Break-out 871 Some routing and transport area directors felt a little left out by 872 all the application layer break-outs, so they had their own 873 brainstorm about what could be done at the Transport and Routing 874 layers from which these notes resulted. 876 The LEDBAT [RFC6817] protocol was targeted towards a bulk-transfer 877 service that is reordering and delay insensitive. Use of LDEBAT 878 could offer the following benefits for an application: 880 a. Because it is reordering-insensitive, traffic can be sprayed 881 across a large number of forwarding paths. Assuming such 882 different paths exist, this would make it more challenging to 883 capture and analyze a full interaction. 885 b. The application can vary the paths by indicating per packet a 886 different flow. In IPv6, this can be done via different IPv6 887 flow labels. For IPv4, this can be done by encapsulating the IP 888 packet into UDP and varying the UDP source port. 890 c. Since LEDBAT is delay-insensitive and applications using it would 891 need to be as well, it would be possible to obfuscate the 892 application signatures by varying the packet lengths and 893 frequency. 895 d. This can also hide the transport header (for IP in UDP). 897 e. If the Reverse Path Forwarding(RPF)[RFC3704] check problem can be 898 fixed, perhaps the source could be hidden, however it assumes the 899 trafic is within trusted perimeters. 901 f. The use of LEDBAT is orthogonal to the use of encryption and 902 provides different benefits (harder to intercept the whole 903 conversation, ability to obfuscate the traffic analysis), and 904 also has different costs (longer latency, new transport protocol 905 usage) to its users. 907 The idea of encrypting traffic from customer edge (CE) to CE as part 908 of an L3VPN or such was also discussed. This could allow hiding of 909 addresses, including source, and headers. From conversation with Ron 910 Bonica, some customers already do encryption (though not hiding the 911 source address) like this. So, it is unclear that this is very 912 practically useful as an enhancement except for encouraging 913 deployment and use. 915 Finally, it was discussed whether it would be useful to have a means 916 of communicating where and what layers are doing encryption on an 917 application's traffic path. The initial idea of augmenting ICMP has 918 some issues (not visible to application, ICMP packets frequently 919 filtered) as well as potential work (determining how to trust the 920 report of encryption). It would be interesting to understand if such 921 communication is actually needed and what the requirements would be. 923 6. After the workshop 925 Holding the workshop just before the IETF had the intended effect: a 926 number of people went to both the workshop and the IETF, and they 927 took the opportunity of being together at the IETF to continue the 928 discussions. 930 IETF Working groups meeting in London took the recommendations from 931 the workshop into account. It was even the first item in the report 932 about the IETF meeting by the IETF chair, Jari Arkko: 934 "Strengthening the security and privacy of the Internet continued 935 to draw a lot of attention. The STRINT workshop organised by the 936 IAB and W3C just before the IETF attracted 100 participants and 937 over 60 papers. Even more people would have joined us, but there 938 was no space. During the IETF meeting, we continued discussing 939 the topic at various working groups. A while ago we created the 940 first working group specifically aimed at addressing some of the 941 issues surrounding pervasive monitoring. The Using TLS for 942 Applications (UTA) working group had its first meeting in London. 943 But many other working groups also address these issues in their 944 own work. The TCPM working group discussed a proposal to add 945 opportunistic keying mechanisms directly onto the TCP protocol. 946 And the DNSE BOF considered the possibility of adding 947 confidentiality support to DNS queries. Finally, there is an 948 ongoing effort to review old specifications to search for areas 949 that might benefit from taking privacy and data minimisation 950 better into account."[Arkko1] 952 Two papers that were written for the workshop, but not finished in 953 time, are worth mentioning, too: One by the same Jari Arkko, titled 954 "Privacy and Networking Functions" [Arkko2]; and one by Johan 955 Pouwelse, "The Shadow Internet: liberation from Surveillance, 956 Censorship and Servers" [draft-pouwelse-perpass-shadow-internet] 958 7. IANA considerations 960 There are none. We hope the RFC editor deletes this section. 962 8. Security considerations 964 This document does not define a technology but is all about security 965 and privacy. 967 9. References 969 9.1. Informative references 971 [Arkko1] Arkko, J., "IETF-89 Summary", March 2014, 972 . 974 [Arkko2] Arkko, J., "Privacy and Networking Functions", March 2014, 975 . 978 (Work in progress.) 980 [draft-ietf-websec-key-pinning] 981 Evans, C., Palmer, C., and R. Sleevi, "Public Key Pinning 982 Extension for HTTP", February 2014. 984 (Work in progress.) 986 [draft-pouwelse-perpass-shadow-internet] 987 Pouwelse, J., Ed., "The Shadow Internet: liberation from 988 Surveillance, Censorship and Servers", February 2014, 989 . 992 (Work in progress.) 994 [I-D.barnes-pervasive-problem] 995 Barnes, R., Schneier, B., Jennings, C., and T. Hardie, 996 "Pervasive Attack: A Threat Model and Problem Statement", 997 draft-barnes-pervasive-problem-01 (work in progress), July 998 2014. 1000 [I-D.kent-opportunistic-security] 1001 Kent, S., "Opportunistic Security as a Countermeasure to 1002 Pervasive Monitoring", draft-kent-opportunistic- 1003 security-01 (work in progress), April 2014. 1005 [RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, 1006 A., Peterson, J., Sparks, R., Handley, M., and E. 1007 Schooler, "SIP: Session Initiation Protocol", RFC 3261, 1008 DOI 10.17487/RFC3261, June 2002, 1009 . 1011 [RFC3365] Schiller, J., "Strong Security Requirements for Internet 1012 Engineering Task Force Standard Protocols", BCP 61, 1013 RFC 3365, DOI 10.17487/RFC3365, August 2002, 1014 . 1016 [RFC3552] Rescorla, E. and B. Korver, "Guidelines for Writing RFC 1017 Text on Security Considerations", BCP 72, RFC 3552, 1018 DOI 10.17487/RFC3552, July 2003, 1019 . 1021 [RFC3704] Baker, F. and P. Savola, "Ingress Filtering for Multihomed 1022 Networks", BCP 84, RFC 3704, DOI 10.17487/RFC3704, March 1023 2004, . 1025 [RFC4252] Ylonen, T. and C. Lonvick, Ed., "The Secure Shell (SSH) 1026 Authentication Protocol", RFC 4252, DOI 10.17487/RFC4252, 1027 January 2006, . 1029 [RFC4322] Richardson, M. and D. Redelmeier, "Opportunistic 1030 Encryption using the Internet Key Exchange (IKE)", 1031 RFC 4322, DOI 10.17487/RFC4322, December 2005, 1032 . 1034 [RFC6120] Saint-Andre, P., "Extensible Messaging and Presence 1035 Protocol (XMPP): Core", RFC 6120, DOI 10.17487/RFC6120, 1036 March 2011, . 1038 [RFC6585] Nottingham, M. and R. Fielding, "Additional HTTP Status 1039 Codes", RFC 6585, April 2012. 1041 [RFC6817] Shalunov, S., Hazel, G., Iyengar, J., and M. Kuehlewind, 1042 "Low Extra Delay Background Transport (LEDBAT)", RFC 6817, 1043 DOI 10.17487/RFC6817, December 2012, 1044 . 1046 [RFC6962] Laurie, B., Langley, A., and E. Kasper, "Certificate 1047 Transparency", RFC 6962, DOI 10.17487/RFC6962, June 2013, 1048 . 1050 [RFC7258] Farrell, S. and H. Tschofenig, "Pervasive Monitoring Is an 1051 Attack", BCP 188, RFC 7258, DOI 10.17487/RFC7258, May 1052 2014, . 1054 [RFC7435] Dukhovni, V., "Opportunistic Security: Some Protection 1055 Most of the Time", RFC 7435, DOI 10.17487/RFC7435, 1056 December 2014, . 1058 [saag] Area, S., "IETF Security Area mailing list", March 2014, 1059 . 1062 [vancouverplenary] 1063 IETF, , "IETF 88 Technical Plenary Minutes", 1064 . 1067 [w3c-geo-api] 1068 Popescu, A., "Geolocation API Specification", October 1069 2013, . 1071 9.2. URIs 1073 [1] http://www.w3.org/ 1075 [2] https://www.iab.org/ 1077 [3] https://www.w3.org/2014/strint/Overview.html 1079 [4] https://www.ietf.org/meeting/89/index.html 1081 [5] http://www.strews.eu/ 1083 [6] https://en.wikipedia.org/wiki/Captive_portal 1085 [7] http://tools.ietf.org/html/bcp72 1087 [8] https://www.w3.org/2014/strint/papers/44.pdf 1089 [9] https://www.w3.org/2014/strint/papers/04.pdf 1091 [10] http://freehaven.net/anonbib/ 1093 [11] http://www.informatik.uni-trier.de/~Ley/db/conf/pet/index.html 1095 [12] http://www.informatik.uni-trier.de/~Ley/db/conf/wpes/index.html 1097 [13] https://community.qualys.com/blogs/securitylabs/2012/09/14/crime 1098 -information-leakage-attack-against-ssltls 1100 [14] http://www.strews.eu/ 1102 [15] http://cordis.europa.eu/fp7/ict/ 1104 [16] http://blog.digital.telefonica.com/ 1106 [17] https://www.ietf.org/meeting/89/index.html 1108 [18] http://lists.i1b.org/pipermail/strint-attendees-i1b.org/ 1110 [19] https://www.w3.org/2014/strint/ 1112 [20] https://twitter.com/search?q=%23strint 1114 [21] http://www.w3.org/2014/02/28-strint-minutes.html 1116 [22] http://down.dsg.cs.tcd.ie/strint-slides/s0-welcome.pdf 1118 [23] http://down.dsg.cs.tcd.ie/strint-slides/s1-threat.pdf 1120 [24] http://down.dsg.cs.tcd.ie/strint-slides/s2-comsec.pdf 1122 [25] http://down.dsg.cs.tcd.ie/strint-slides/s3-policy.pdf 1124 [26] http://www.w3.org/2014/03/01-strint-minutes.html 1126 [27] http://down.dsg.cs.tcd.ie/strint-slides/s4-opportunistic.pdf 1128 [28] http://down.dsg.cs.tcd.ie/strint-slides/s5-1metadata-pironti.pdf 1130 [29] http://down.dsg.cs.tcd.ie/strint-slides/s5-2metadata-hardie.pdf 1132 [30] http://down.dsg.cs.tcd.ie/strint-slides/s5-3metadata-cooper.pdf 1134 [31] http://down.dsg.cs.tcd.ie/strint-slides/s6-deploy.pdf 1136 [32] https://www.w3.org/2014/strint/slides/summary.pdf 1138 [33] https://www.cs.tcd.ie/Stephen.Farrell/ 1140 [34] http://www.w3.org/People/Rigo/ 1142 [35] http://www.tschofenig.priv.at/wp/?page_id=5 1144 Appendix A. Logistics 1146 The workshop was organised by the STREWS [14] project (a research 1147 project funded under the European Union's 7th Framework Programme 1148 [15]), as the first of two workshops in its work plan. The 1149 organisers were supported by the IAB and W3C, and, for the local 1150 organisation, by Telefonica Digital. [16] 1152 One of the suggestions in the project description of the STREWS 1153 project was to attach the first workshop to an IETF meeting. The 1154 best opportunity was IETF 89 [17] in London, which would begin on 1155 Sunday March 2, 2014. Telefonica Digital offered meeting rooms at 1156 its offices in central London for the preceding Friday and Saturday, 1157 just minutes away from the IETF's location. 1159 The room held 100 people, which was thought to be sufficient. There 1160 turned out to be more interest than expected and we could have filled 1161 a larger room, but 100 people is probably an upper limit for good 1162 discussions anyway. 1164 Apart from the usual equipment in the room (projector, white boards, 1165 microphones, coffee...), we also set up some extra communication 1166 channels: 1168 o A mailing list where participants could discuss the agenda and the 1169 published papers about three weeks in advance of the workshop 1170 itself. (Only participants were allowed to write to the mailing 1171 list, but the archive [18] is public.) 1173 o Publicly advertised streaming audio (one-way only). At some 1174 point, no less than 165 people were listening. 1176 o An IRC channel for live minute taking, passing links and other 1177 information, and as a help for remote participants to follow the 1178 proceedings. 1180 o An Etherpad, where the authors of papers could provide an abstract 1181 of their submissions, to help participants who could not read all 1182 66 papers in full in advance of the workshop. The abstracts were 1183 also used on the workshop's Web site [19]. 1185 o A "Twitter hashtag" (#strint). Four weeks after the workshop, 1186 there were still a few new messages [20] about events related to 1187 workshop topics. 1189 Appendix B. Agenda 1191 This was the final agenda of the workshop, as determined by the TPC 1192 and participants on the mailing list prior to the workshop. The 1193 included links are to the slides that the moderators used to 1194 introduce each discussion topic and to the minutes. 1196 B.1. Friday 28 February 1198 Minutes [21] 1200 Workshop starts, welcome, logistics, opening/overview [slides] 1201 [22] 1203 * Goal is to plan how we respond to PM threats 1205 * Specific questions to be discussed in sessions 1207 * Outcomes are actions for IETF, W3C, IRTF, etc. 1209 I. Threats - What problem are we trying to solve? (Presenter: 1210 Richard Barnes; Moderator: Cullen Jennings) [slides] [23] 1212 * What attacks have been described? (Attack taxonomy) 1214 * What should we assume the attackers' capabilities are? 1216 * When is it really "pervasive monitoring" and when is it not? 1218 * Scoping - what's in and what's out? (for IETF/W3C) 1220 II. COMSEC 1 - How can we increase usage of current COMSEC tools? 1221 (Presenter: Hannes Tschofenig; Moderator: Leif Johansson) [slides] 1222 [24] 1224 * Whirlwind catalog of current tools 1226 * Why aren't people using them? In what situations are / aren't 1227 they used? 1229 * Securing AAA and management protocols - why not? 1231 * How can we (IETF/W3C/community) encourage more/better use? 1233 III. Policy - What policy / legal/ other issues need to be taken 1234 into account? (Presenter: Christine Runnegar; Moderator: Rigo 1235 Wenning) [slides] [25] 1236 * What non-technical activities do we need to be aware of? 1238 * How might such non-technical activities impact on IETF/W3C? 1240 * How might IETF/W3C activities impact on those non-technical 1241 activities? 1243 Session IV - Saturday plan, open-mic, wrap up day 1245 B.2. Saturday 1 March 1247 Minutes [26] 1249 IV. COMSEC 2 - What improvements to COMSEC tools are 1250 needed?(Presenter: Mark Nottingham; Moderator: Steve Bellovin) 1251 [slides] [27] 1253 * Opportunistic encryption - what is it and where it might apply 1255 * Mitigations aiming to block PM vs. detect PM - when to try 1256 which? 1258 V. Metadata - How can we reduce the metadata that protocols 1259 expose? (Presenter: Alfredo Pironti [slides] [28] / Ted Hardie 1260 [slides] [29]; Moderator: Alissa Cooper [slides] [30]) 1262 * Meta-data, fingerprinting, minimisation 1264 * What's out there? 1266 * How can we do better? 1268 VI. Deployment - How can we address PM in deployment / 1269 operations? (Presenter: Eliot Lear; Moderator: Barry Leiba) 1270 [slides] [31] 1272 * "Mega"-commercial services (clouds, large scale email & SN, 1273 SIP, WebRTC...) 1275 * Target dispersal - good goal or wishful thinking? 1277 * Middleboxes: when a help and when a hindrance? 1279 VII. 3 x Break-out Sessions / Bar-Camp style (Hannes Tschofenig) 1281 * Content to be defined during meeting, as topics come up 1283 * Sum up at the end to gather conclusions for report 1284 Break-outs: 1286 1. Research Questions (Moderator: Kenny Paterson) 1288 + Do we need more/different crypto tools? 1290 + How can applications make better use of COMSEC tools? 1292 + What research topics could be handled in IRTF? 1294 + What other research would help? 1296 2. clients 1298 3. on by default 1300 4. measuring 1302 5. opportunistic 1304 VIII. Break-out reports, Open mic & Conclusions - What are we 1305 going to do to address PM? [slides] [32] 1307 * Gather conclusions / recommendations / goals from earlier 1308 sessions 1310 Appendix C. Workshop chairs & program committee 1312 The workshop chairs were three: Stephen Farrell [33] (TCD) and Rigo 1313 Wenning [34] (W3C) from the STREWS project, and Hannes Tschofenig 1314 [35] (ARM) from the STREWS Interest Group. 1316 A program committee (PC) was charged with evaluating the submitted 1317 papers. It was made up of the members of the STREWS project, the 1318 members of the STREWS Interest Group, plus invited experts: Bernard 1319 Aboba (Microsoft), Dan Appelquist (Telefonica & W3C TAG), Richard 1320 Barnes (Mozilla), Bert Bos (W3C), Lieven Desmet (KU Leuven), Karen 1321 O'Donoghue (ISOC), Russ Housley (Vigil Security), Martin Johns (SAP), 1322 Ben Laurie (Google), Eliot Lear (Cisco), Kenny Paterson (Royal 1323 Holloway), Eric Rescorla (RTFM), Wendy Seltzer (W3C), Dave Thaler 1324 (Microsoft) and Sean Turner (IECA). 1326 Appendix D. Participants 1328 The participants to the workshop were: 1330 o Bernard Aboba (Microsoft Corporation) 1331 o Thijs Alkemade (Adium) 1333 o Daniel Appelquist (Telefonica Digital) 1335 o Jari Arkko (Ericsson) 1337 o Alia Atlas (Juniper Networks) 1339 o Emmanuel Baccelli (INRIA) 1341 o Mary Barnes 1343 o Richard Barnes (Mozilla) 1345 o Steve Bellovin (Columbia University) 1347 o Andrea Bittau (Stanford University) 1349 o Marc Blanchet (Viagenie) 1351 o Carsten Bormann (Uni Bremen TZI) 1353 o Bert Bos (W3C) 1355 o Ian Brown (Oxford University) 1357 o Stewart Bryant (Cisco Systems) 1359 o Randy Bush (IIJ / Dragon Research Labs) 1361 o Kelsey Cairns (Washington State University) 1363 o Stuart Cheshire (Apple) 1365 o Vincent Cheval (University of Birmingham) 1367 o Benoit Claise (Cisco) 1369 o Alissa Cooper (Cisco) 1371 o Dave Crocker (Brandenburg InternetWorking) 1373 o Leslie Daigle (Internet Society) 1375 o George Danezis (University College London) 1377 o Spencer Dawkins (Huawei) 1378 o Mark Donnelly (Painless Security) 1380 o Nick Doty (W3C) 1382 o Dan Druta (AT&T) 1384 o Peter Eckersley (Electronic Frontier Foundation) 1386 o Lars Eggert (NetApp) 1388 o Kai Engert (Red Hat) 1390 o Monika Ermert 1392 o Stephen Farrell (Trinity College Dublin) 1394 o Barbara Fraser (Cisco) 1396 o Virginie Galindo (gemalto) 1398 o Stefanie Gerdes (Uni Bremen TZI) 1400 o Daniel Kahn Gillmor (ACLU) 1402 o Wendy M. Grossman 1404 o Christian Grothoff (The GNUnet Project) 1406 o Oliver Hahm (INRIA) 1408 o Joseph Lorenzo Hall (Center for Democracy & Technology) 1410 o Phillip Hallam-Baker 1412 o Harry Halpin (W3C/MIT and IRI) 1414 o Ted Hardie (Google) 1416 o Joe Hildebrand (Cisco Systems) 1418 o Russ Housley (Vigil Security, LLC) 1420 o Cullen Jennings (CISCO) 1422 o Leif Johansson (SUNET) 1424 o Harold Johnson (Irdeto) 1425 o Alan Johnston (Avaya) 1427 o L. Aaron Kaplan (CERT.at) 1429 o Steve Kent (BBN Technologies) 1431 o Achim Klabunde (European Data Protection Supervisor) 1433 o Hans Kuhn (NOC) 1435 o Christian de Larrinaga 1437 o Ben Laurie (Google) 1439 o Eliot Lear (Cisco Ssytems) 1441 o Barry Leiba (Huawei Technologies) 1443 o Sebastian Lekies (SAP AG) 1445 o Orit Levin (Microsoft Corporation) 1447 o Carlo Von LynX (#youbroketheinternet) 1449 o Xavier Marjou (Orange) 1451 o Larry Masinter (Adobe) 1453 o John Mattsson (Ericsson) 1455 o Patrick McManus (Mozilla) 1457 o Doug Montgomery (NIST) 1459 o Kathleen Moriarty (EMC) 1461 o Alec Muffett (Facebook) 1463 o Suhas Nandakumar (Cisco Systems) 1465 o Linh Nguyen (ERCIM/W3C) 1467 o Linus Nordberg (NORDUnet) 1469 o Mark Nottingham 1471 o Karen O'Donoghue (Internet Society) 1472 o Piers O'Hanlon (Oxford Internet Institute) 1474 o Kenny Paterson (Royal Holloway, University of London) 1476 o Jon Peterson (Neustar) 1478 o Joshua Phillips (University of Birmingham) 1480 o Alfredo Pironti (INRIA) 1482 o Dana Polatin-Reuben (University of Oxford) 1484 o Prof. Johan Pouwelse (Delft University of Technology) 1486 o Max Pritikin (Cisco) 1488 o Eric Rescorla (Mozilla) 1490 o Pete Resnick (Qualcomm Technologies, Inc.) 1492 o Tom Ristenpart (University of Wisconsin) 1494 o Andrei Robachevsky (Internet Society) 1496 o David Rogers (Copper Horse) 1498 o Scott Rose (NIST) 1500 o Christine Runnegar (Internet Society) 1502 o Philippe De Ryck (DistriNet - KU Leuven) 1504 o Peter Saint-Andre (&yet) 1506 o Runa A. Sandvik (Center for Democracy and Technology) 1508 o Jakob Schlyter 1510 o Dr. Jan Seedorf (NEC Laboratories Europe) 1512 o Wendy Seltzer (W3C) 1514 o Melinda Shore (No Mountain Software) 1516 o Dave Thaler (Microsoft) 1518 o Brian Trammell (ETH Zurich) 1519 o Hannes Tschofenig (ARM Limited) 1521 o Sean Turner (IECA, Inc.) 1523 o Matthias Waehlisch (Freie Universitaet Berlin) 1525 o Greg Walton (Oxford University) 1527 o Rigo Wenning (W3C) 1529 o Tara Whalen (Apple Inc.) 1531 o Greg Wood (Internet Society) 1533 o Jiangshan Yu (University of Birmingham) 1535 o Aaron Zauner 1537 o Dacheng Zhang (Huawei) 1539 o Phil Zimmermann (Silent Circle LLC) 1541 o Juan-Carlos Zuniga (InterDigital) 1543 Authors' Addresses 1545 Stephen Farrell 1546 Trinity College, Dublin 1548 Email: stephen.farrell@cs.tcd.ie 1550 Rigo Wenning 1551 World Wide Web Consortium 1552 2004, route des Lucioles 1553 B.P. 93 1554 Sophia-Antipolis 06902 1555 France 1557 Email: rigo@w3.org 1558 Bert Bos 1559 World Wide Web Consortium 1560 2004, route des Lucioles 1561 B.P. 93 1562 Sophia-Antipolis 06902 1563 France 1565 Email: bert@w3org 1567 Marc Blanchet 1568 Viagenie 1569 246 Aberdeen 1570 Quebec, QC G1R 2E1 1571 Canada 1573 Email: Marc.Blanchet@viagenie.ca 1574 URI: http://viagenie.ca 1576 Hannes Tschofenig 1577 ARM Ltd. 1578 110 Fulbourn Rd 1579 Cambridge CB1 9NJ 1580 Great Britain 1582 Email: Hannes.tschofenig@gmx.net 1583 URI: http://www.tschofenig.priv.at