idnits 2.17.1 draft-ietf-dnsop-nsec-aggressiveuse-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 has examples using IPv4 documentation addresses according to RFC6890, but does not use any IPv6 documentation addresses. Maybe there should be IPv6 examples, too? Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year (Using the creation date from RFC4035, updated by this document, for RFC5378 checks: 2002-10-29) -- 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 (October 4, 2016) is 2755 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) ** Downref: Normative reference to an Historic RFC: RFC 5074 ** Obsolete normative reference: RFC 7719 (Obsoleted by RFC 8499) == Outdated reference: A later version (-05) exists of draft-ietf-dnsop-nxdomain-cut-03 Summary: 2 errors (**), 0 flaws (~~), 2 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group K. Fujiwara 3 Internet-Draft JPRS 4 Updates: 4035 (if approved) A. Kato 5 Intended status: Standards Track Keio/WIDE 6 Expires: April 7, 2017 W. Kumari 7 Google 8 October 4, 2016 10 Aggressive use of NSEC/NSEC3 11 draft-ietf-dnsop-nsec-aggressiveuse-03 13 Abstract 15 The DNS relies upon caching to scale; however, the cache lookup 16 generally requires an exact match. This document specifies the use 17 of NSEC/NSEC3 resource records to allow DNSSEC validating resolvers 18 to generate negative answers within a range. This increases 19 performance / decreases latency, decreases resource utilization on 20 both authoritative and recursive servers, and also increases privacy. 21 It may also help increase resilience to certain DoS attacks in some 22 circumstances. 24 This document updates RFC4035 by allowing validating resolvers to 25 generate negative answers based upon NSEC/NSEC3 records. 27 [ Ed note: Text inside square brackets ([]) is additional background 28 information, answers to frequently asked questions, general musings, 29 etc. They will be removed before publication.This document is being 30 collaborated on in Github at: https://github.com/wkumari/draft-ietf- 31 dnsop-nsec-aggressiveuse. The most recent version of the document, 32 open issues, etc should all be available here. The authors 33 (gratefully) accept pull requests.] 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 April 7, 2017. 51 Copyright Notice 53 Copyright (c) 2016 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. Code Components extracted from this document must 62 include Simplified BSD License text as described in Section 4.e of 63 the Trust Legal Provisions and are provided without warranty as 64 described in the Simplified BSD License. 66 Table of Contents 68 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 69 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 70 3. Problem Statement . . . . . . . . . . . . . . . . . . . . . . 3 71 4. Background . . . . . . . . . . . . . . . . . . . . . . . . . 4 72 5. Aggressive Negative Caching . . . . . . . . . . . . . . . . . 5 73 5.1. NSEC . . . . . . . . . . . . . . . . . . . . . . . . . . 5 74 5.2. NSEC3 . . . . . . . . . . . . . . . . . . . . . . . . . . 6 75 5.3. Consideration on TTL . . . . . . . . . . . . . . . . . . 6 76 6. Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . 6 77 7. Update to RFC 4035 . . . . . . . . . . . . . . . . . . . . . 7 78 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 7 79 9. Security Considerations . . . . . . . . . . . . . . . . . . . 8 80 10. Implementation Status . . . . . . . . . . . . . . . . . . . . 8 81 11. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 8 82 12. Change History . . . . . . . . . . . . . . . . . . . . . . . 8 83 12.1. Version draft-fujiwara-dnsop-nsec-aggressiveuse-01 . . . 10 84 12.2. Version draft-fujiwara-dnsop-nsec-aggressiveuse-02 . . . 10 85 12.3. Version draft-fujiwara-dnsop-nsec-aggressiveuse-03 . . . 10 86 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 11 87 13.1. Normative References . . . . . . . . . . . . . . . . . . 11 88 13.2. Informative References . . . . . . . . . . . . . . . . . 11 89 Appendix A. Detailed implementation notes . . . . . . . . . . . 12 90 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 12 92 1. Introduction 94 A DNS negative cache exists, and is used to cache the fact that a 95 name does not exist. This method of negative caching requires exact 96 matching; this leads to unnecessary additional lookups, increases 97 latency, leads to extra resource utilization on both authoritative 98 and recursive servers, and decreases privacy by leaking queries. 100 This document updates RFC 4035 to allow recursive resolvers to use 101 NSEC/NSEC3 resource records to synthetize negative answers from the 102 information they have in the cache. This allows validating resolvers 103 to respond with NXDOMAIN immediately if the name in question falls 104 into a range expressed by a NSEC/NSEC3 resource record already in the 105 cache. 107 Aggressive Negative Caching was first proposed in Section 6 of DNSSEC 108 Lookaside Validation (DLV) [RFC5074] in order to find covering NSEC 109 records efficiently. 111 Section 3 of [I-D.vixie-dnsext-resimprove] "Stopping Downward Cache 112 Search on NXDOMAIN" and [I-D.ietf-dnsop-nxdomain-cut] proposed 113 another approach to use NXDOMAIN information effectively. 115 2. Terminology 117 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 118 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 119 document are to be interpreted as described in RFC 2119 [RFC2119]. 121 Many of the specialized terms used in this document are defined in 122 DNS Terminology [RFC7719]. 124 The key words "Closest Encloser" and "Source of Synthesis" in this 125 document are to be interpreted as described in [RFC4592]. 127 "Closest Encloser" is also defined in NSEC3 [RFC5155], as is "Next 128 closer name". 130 3. Problem Statement 132 The DNS negative cache caches negative (non-existent) information, 133 and requires an exact match in most instances [RFC2308]. 135 Assume that the (DNSSEC signed) "example.com" zone contains: 137 apple.example.com IN A 192.0.2.1 139 elephant.example.com IN A 192.0.2.2 141 zebra.example.com IN A 192.0.2.3 143 If a validating resolver gets a query for cat.example.com, it will 144 query the example.com servers and will get back an NSEC (or NSEC3) 145 record starting that there are no records between apple and elephant. 146 The resolver then knows that cat.example.com does not exist; however, 147 it does not use the fact that the proof covers a range (apple to 148 elephant) to suppress queries for other labels that fall within this 149 range. This means that if the validating resolver gets a query for 150 ball.example.com (or dog.example.com) it will once again go off and 151 query the example.com servers for these names. 153 Apart from wasting bandwidth, this also wastes resources on the 154 recursive server (it needs to keep state for outstanding queries), 155 wastes resources on the authoritative server (it has to answer 156 additional questions), increases latency (the end user has to wait 157 longer than necessary to get back an NXDOMAIN answer), can be used by 158 attackers to cause a DoS (see additional resources), and also has 159 privacy implications (e.g: typos leak out further than necessary). 161 4. Background 163 DNSSEC [RFC4035] and [RFC5155] both provide "authenticated denial of 164 existence"; this is a cryptographic proof that the queried for name 165 does not exist, accomplished by providing a (DNSSEC secured) record 166 containing the names which appear alphabetically before and after the 167 queried for name. In the example above, if the (DNSSEC validating) 168 recursive server were to query for lion.example.com it would receive 169 a (signed) NSEC/NSEC3 record stating that there are no labels between 170 "elephant" and "zebra". This is a signed, cryptographic proof that 171 these names are the ones before and after the queried for label. As 172 lion.example.com falls within this range, the recursive server knows 173 that lion.example.com really does not exist. This document specifies 174 that this NSEC/NSEC3 record should be used to generate negative 175 answers for any queries that the recursive server receives that fall 176 within the range covered by the record (for the TTL for the record). 178 Section 4.5 of [RFC4035] says: 180 "In theory, a resolver could use wildcards or NSEC RRs to generate 181 positive and negative responses (respectively) until the TTL or 182 signatures on the records in question expire. However, it seems 183 prudent for resolvers to avoid blocking new authoritative data or 184 synthesizing new data on their own. Resolvers that follow this 185 recommendation will have a more consistent view of the namespace." 186 and "The reason for these recommendations is that, between the 187 initial query and the expiration of the data from the cache, the 188 authoritative data might have been changed (for example, via dynamic 189 update).". In other words, if a resolver generates negative answers 190 from an NSEC record, it will not send any queries for names within 191 that NSEC range (for the TTL). If a new name is added to the zone 192 during this interval the resolver will not know this. 194 We believe this recommendation can be relaxed because, in the absense 195 of this technique, a lookup for the exact name could have come in 196 during this interval, and so this could already be cached (see 197 [RFC2308] for more background). This means that zone operators 198 should have no expectation that an added name would work immediately. 199 With DNSSEC and Aggressive NSEC, the TTL of the NSEC record is the 200 authoritative statement of how quickly a name can start working 201 within a zone. 203 5. Aggressive Negative Caching 205 Section 4.5 of [RFC4035] shows that "In theory, a resolver could use 206 wildcards or NSEC RRs to generate positive and negative responses 207 (respectively) until the TTL or signatures on the records in question 208 expire. However, it seems prudent for resolvers to avoid blocking 209 new authoritative data or synthesizing new data on their own. 210 Resolvers that follow this recommendation will have a more consistent 211 view of the namespace". 213 This document relaxes this this restriction, as follows: 215 +--------------------------------------------------------------+ 216 | Once the records are validated, DNSSEC enabled validating | 217 | resolvers SHOULD use NSEC/NSEC3 resource records to generate | 218 | negative responses until their effective TTLs or signatures | 219 | for those records expire. | 220 +--------------------------------------------------------------+ 222 If the validating resolver's cache has sufficient information to 223 validate the query, the resolver SHOULD use NSEC/NSEC3/wildcard 224 records aggressively. Otherwise, it MUST fall back to send the query 225 to the authoritative DNS servers. 227 If the query name has the matching NSEC/NSEC3 RR proving the 228 information requested does not exist, the validating resolver may 229 respond with a NODATA (empty) answer. 231 5.1. NSEC 233 Implementations which support aggressive use of NSEC SHOULD enable 234 this by default. Implementations MAY provide a configuration switch 235 to disable aggressive use of NSEC and allow it to be enabled or 236 disabled per domain. 238 The validating resolver needs to check the existence of an NSEC RR 239 matching/covering the source of synthesis and an NSEC RR covering the 240 query name. 242 If the validating resolver's cache contains an NSEC RR covering the 243 source of synthesis and the covering NSEC RR of the query name, the 244 validating resolver may respond with NXDOMAIN error immediately. 246 5.2. NSEC3 248 NSEC3 aggressive negative caching is more difficult. If the zone is 249 signed with NSEC3, the validating resolver needs to check the 250 existence of non-terminals and wildcards which derive from query 251 names. 253 If the validating resolver's cache contains an NSEC3 RR matching the 254 closest encloser, an NSEC3 RR covering the next closer name, and an 255 NSEC3 RR covering the source of synthesis, it is possible for the 256 validating resolver to respond with NXDOMAIN immediately. 258 If a covering NSEC3 RR has Opt-Out flag, the covering NSEC3 RR does 259 not prove the non-existence of the domain name and the aggressive 260 negative caching is not possible for the domain name. 262 A validating resolver implementation MAY support aggressive use of 263 NSEC3. If it does aggressive use of NSEC3, it MAY provide a 264 configuration switch to disable aggressive use of NSEC3 and allow it 265 to be enabled or disabled for specific zones. 267 5.3. Consideration on TTL 269 The TTL value of negative information is especially important, 270 because newly added domain names cannot be used while the negative 271 information is effective. Section 5 of RFC 2308 states that the 272 maximum number of negative cache TTL value is 3 hours (10800). It is 273 RECOMMENDED that validating resolvers limit the maximum effective TTL 274 value of negative responses (NSEC/NSEC3 RRs) to this same value. 276 6. Benefits 278 The techniques described in this document provide a number of 279 benefits, including (in no specific order): 281 Reduced latency By answering directly from cache, validating 282 resolvers can immediately inform clients that the name they are 283 looking for does not exist, improving the user experience. 285 Decreased recursive server load By answering negative queries from 286 the cache, validating servers avoid having send a query and wait 287 for a response. In addition to decreasing the bandwidth used, it 288 also means that the server does not need to allocate and maintain 289 state, thereby decreasing memory and CPU load. 291 Decreased authorative server load Because recursive servers can 292 answer (negative) queries without asking the authoritative server, 293 the authoritative servers receive less queries. This decreases 294 the authoritative server bandwidth, queries per second and CPU 295 utilization. 297 The scale of the benefit depends upon multiple factors, including the 298 query distribution. For example, currently around 65% of queries to 299 Root Name servers result in NXDOMAIN responses; this technique will 300 eliminate a sizable quantity of these. 302 The technique described in this document may also mitigate so-called 303 "random QNAME attacks", in which attackers send many queries for 304 random sub-domains to resolvers. As the resolver will not have the 305 answers cached it has to ask external servers for each random query, 306 leading to a DoS on the authoritative servers (and often resolvers). 307 Aggressive NSEC may help mitigate these attacks by allowing the 308 resolver to answer directly from cache for any random queries which 309 fall within already requested ranges. It will not always work as an 310 effective defense, not least because not many zones are DNSSEC signed 311 at all, but it will still provide an additional layer of defense. 313 7. Update to RFC 4035 315 Section 4.5 of [RFC4035] shows that "In theory, a resolver could use 316 wildcards or NSEC RRs to generate positive and negative responses 317 (respectively) until the TTL or signatures on the records in question 318 expire. However, it seems prudent for resolvers to avoid blocking 319 new authoritative data or synthesizing new data on their own. 320 Resolvers that follow this recommendation will have a more consistent 321 view of the namespace". 323 The paragraph is updated as follows: 325 +--------------------------------------------------------------+ 326 | Once the records are validated, DNSSEC enabled validating | 327 | resolvers MAY use wildcards and NSEC/NSEC3 resource records | 328 | to generate negative responses until their effective TTLs | 329 | or signatures for those records expire. | 330 +--------------------------------------------------------------+ 332 8. IANA Considerations 334 This document has no IANA actions. 336 9. Security Considerations 338 Newly registered resource records may not be used immediately. 339 However, choosing suitable TTL value and negative cache TTL value 340 (SOA MINIMUM field) will mitigate the delay concern, and it is not a 341 security problem. 343 It is also suggested to limit the maximum TTL value of NSEC / NSEC3 344 resource records in the negative cache to, for example, 10800 seconds 345 (3hrs), to mitigate this issue. Implementations which comply with 346 this proposal are recommended to have a configurable maximum value of 347 NSEC RRs in the negative cache. 349 Aggressive use of NSEC / NSEC3 resource records without DNSSEC 350 validation may create serious security issues, and so this technique 351 requires DNSSEC validation. 353 10. Implementation Status 355 Unbound currenty implements aggressive negative caching, as does 356 Google Public DNS. 358 11. Acknowledgments 360 The authors gratefully acknowledge DLV [RFC5074] author Samuel Weiler 361 and the Unbound developers. 363 The authors would like to specifically thank Tatuya JINMEI for 364 extensive review and comments, and also Mark Andrews, Stephane 365 Bortzmeyer, Casey Deccio, Alexander Dupuy, Olafur Gudmundsson, Bob 366 Harold, Shumon Huque, John Levine, Pieter Lexis and Matthijs Mekking 367 (who even sent pull requests!). 369 12. Change History 371 RFC Editor: Please remove this section prior to publication. 373 -02 to -03: 375 o Integrated a bunch of comments from Matthijs Mekking - details in: 376 https://github.com/wkumari/draft-ietf-dnsop-nsec-aggressiveuse/ 377 pull/1. I decided to keep "Aggressive Negative Caching" instead 378 of "Aggressive USE OF Negative Caching" for readability. 380 o Attempted to address Bob Harold's comment on the readability 381 issues with "But, it will be more effective when both are 382 enabled..." in Section 5.4 - https://www.ietf.org/mail- 383 archive/web/dnsop/current/msg17997.html 385 o MAYs and SHOULD drifted in the text block. Fixed - thanks to 386 https://mailarchive.ietf.org/arch/msg/ 387 dnsop/2ljmmzxtIMCFMLOZmWcSbTYVOy4 389 o A number of good edits from Stephane in: https://www.ietf.org/ 390 mail-archive/web/dnsop/current/msg18109.html 392 o A bunch more edits from Jinmei, as in: https://www.ietf.org/mail- 393 archive/web/dnsop/current/msg18206.html 395 -01 to -02: 397 o Added Section 6 - Benefits (as suggested by Jinmei). 399 o Removed Appendix B (Jinmei) 401 o Replaced "full-service" with "validating" (where applicable) 403 o Integrated other comments from Jinmei from https://www.ietf.org/ 404 mail-archive/web/dnsop/current/msg17875.html 406 o Integrated comment from co-authors, including re-adding parts of 407 Appendix B, terminology, typos. 409 o Tried to explain under what conditions this may actually mitigate 410 attacks. 412 -00 to -01: 414 o Comments from DNSOP meeting in Berlin. 416 o Changed intended status to Standards Track (updates RFC 4035) 418 o Added a section "Updates to RFC 4035" 420 o Some language clarification / typo / cleanup 422 o Cleaned up the TTL section a bit. 424 o Removed Effects section, Additional proposal section, and pseudo 425 code. 427 o Moved "mitigation of random subdomain attacks" to Appendix. 429 From draft-fujiwara-dnsop-nsec-aggressiveuse-03 -> draft-ietf-dnsop- 430 nsec-aggressiveuse 432 o Document adopted by DNSOP WG. 434 o Adoption comments 436 o Changed main purpose to performance 438 o Use NSEC3/Wildcard keywords 440 o Improved wordings (from good comments) 442 o Simplified pseudo code for NSEC3 444 o Added Warren as co-author. 446 o Reworded much of the problem statement 448 o Reworked examples to better explain the problem / solution. 450 12.1. Version draft-fujiwara-dnsop-nsec-aggressiveuse-01 452 o Added reference to DLV [RFC5074] and imported some sentences. 454 o Added Aggressive Negative Caching Flag idea. 456 o Added detailed algorithms. 458 12.2. Version draft-fujiwara-dnsop-nsec-aggressiveuse-02 460 o Added reference to [I-D.vixie-dnsext-resimprove] 462 o Added considerations for the CD bit 464 o Updated detailed algorithms. 466 o Moved Aggressive Negative Caching Flag idea into Additional 467 Proposals 469 12.3. Version draft-fujiwara-dnsop-nsec-aggressiveuse-03 471 o Added "Partial implementation" 473 o Section 4,5,6 reorganized for better representation 475 o Added NODATA answer in Section 4 477 o Trivial updates 479 o Updated pseudo code 481 13. References 483 13.1. Normative References 485 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 486 Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/ 487 RFC2119, March 1997, 488 . 490 [RFC2308] Andrews, M., "Negative Caching of DNS Queries (DNS 491 NCACHE)", RFC 2308, DOI 10.17487/RFC2308, March 1998, 492 . 494 [RFC4035] Arends, R., Austein, R., Larson, M., Massey, D., and S. 495 Rose, "Protocol Modifications for the DNS Security 496 Extensions", RFC 4035, DOI 10.17487/RFC4035, March 2005, 497 . 499 [RFC4592] Lewis, E., "The Role of Wildcards in the Domain Name 500 System", RFC 4592, DOI 10.17487/RFC4592, July 2006, 501 . 503 [RFC5074] Weiler, S., "DNSSEC Lookaside Validation (DLV)", RFC 5074, 504 DOI 10.17487/RFC5074, November 2007, 505 . 507 [RFC5155] Laurie, B., Sisson, G., Arends, R., and D. Blacka, "DNS 508 Security (DNSSEC) Hashed Authenticated Denial of 509 Existence", RFC 5155, DOI 10.17487/RFC5155, March 2008, 510 . 512 [RFC7719] Hoffman, P., Sullivan, A., and K. Fujiwara, "DNS 513 Terminology", RFC 7719, DOI 10.17487/RFC7719, December 514 2015, . 516 13.2. Informative References 518 [I-D.ietf-dnsop-nxdomain-cut] 519 Bortzmeyer, S. and S. Huque, "NXDOMAIN really means there 520 is nothing underneath", draft-ietf-dnsop-nxdomain-cut-03 521 (work in progress), May 2016. 523 [I-D.vixie-dnsext-resimprove] 524 Vixie, P., Joffe, R., and F. Neves, "Improvements to DNS 525 Resolvers for Resiliency, Robustness, and Responsiveness", 526 draft-vixie-dnsext-resimprove-00 (work in progress), June 527 2010. 529 Appendix A. Detailed implementation notes 531 o Previously, cached negative responses were indexed by QNAME, 532 QCLASS, QTYPE, and the setting of the CD bit (see RFC 4035, 533 Section 4.7), and only queries matching the index key would be 534 answered from the cache. With aggressive negative caching, the 535 validator, in addition to checking to see if the answer is in its 536 cache before sending a query, checks to see whether any cached and 537 validated NSEC record denies the existence of the sought 538 record(s). Using aggressive negative caching, a validator will 539 not make queries for any name covered by a cached and validated 540 NSEC record. Furthermore, a validator answering queries from 541 clients will synthesize a negative answer whenever it has an 542 applicable validated NSEC in its cache unless the CD bit was set 543 on the incoming query. (Imported from Section 6 of [RFC5074]). 545 o Implementing aggressive negative caching suggests that a validator 546 will need to build an ordered data structure of NSEC and NSEC3 547 records for each signer domain name of NSEC / NSEC3 records in 548 order to efficiently find covering NSEC / NSEC3 records. Call the 549 table as NSEC_TABLE. (Imported from Section 6.1 of [RFC5074] and 550 expanded.) 552 o The aggressive negative caching may be inserted at the cache 553 lookup part of the recursive resolvers. 555 o If errors happen in aggressive negative caching algorithm, 556 resolvers MUST fall back to resolve the query as usual. "Resolve 557 the query as usual" means that the resolver must process the query 558 as though it does not implement aggressive negative caching. 560 Authors' Addresses 562 Kazunori Fujiwara 563 Japan Registry Services Co., Ltd. 564 Chiyoda First Bldg. East 13F, 3-8-1 Nishi-Kanda 565 Chiyoda-ku, Tokyo 101-0065 566 Japan 568 Phone: +81 3 5215 8451 569 Email: fujiwara@jprs.co.jp 570 Akira Kato 571 Keio University/WIDE Project 572 Graduate School of Media Design, 4-1-1 Hiyoshi 573 Kohoku, Yokohama 223-8526 574 Japan 576 Phone: +81 45 564 2490 577 Email: kato@wide.ad.jp 579 Warren Kumari 580 Google 581 1600 Amphitheatre Parkway 582 Mountain View, CA 94043 583 US 585 Email: warren@kumari.net