idnits 2.17.1 draft-morishita-dnsop-misbehavior-against-aaaa-00.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Looks like you're using RFC 2026 boilerplate. This must be updated to follow RFC 3978/3979, as updated by RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) == There are 6 instances of lines with non-RFC2606-compliant FQDNs in the document. == There are 6 instances of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. == There are 2 instances of lines with private range IPv4 addresses in the document. If these are generic example addresses, they should be changed to use any of the ranges defined in RFC 6890 (or successor): 192.0.2.x, 198.51.100.x or 203.0.113.x. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (June 17, 2003) is 7619 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) -- Possible downref: Non-RFC (?) normative reference: ref. '2' Summary: 2 errors (**), 0 flaws (~~), 5 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Engineering Task Force Y. Morishita 3 Internet-Draft JPRS 4 Expires: December 16, 2003 T. Jinmei 5 Toshiba 6 June 17, 2003 8 Common Misbehavior against DNS Queries for IPv6 Addresses 9 draft-morishita-dnsop-misbehavior-against-aaaa-00.txt 11 Status of this Memo 13 This document is an Internet-Draft and is in full conformance with 14 all provisions of Section 10 of RFC2026. 16 Internet-Drafts are working documents of the Internet Engineering 17 Task Force (IETF), its areas, and its working groups. Note that 18 other groups may also distribute working documents as Internet- 19 Drafts. 21 Internet-Drafts are draft documents valid for a maximum of six months 22 and may be updated, replaced, or obsoleted by other documents at any 23 time. It is inappropriate to use Internet-Drafts as reference 24 material or to cite them other than as "work in progress." 26 The list of current Internet-Drafts can be accessed at http:// 27 www.ietf.org/ietf/1id-abstracts.txt. 29 The list of Internet-Draft Shadow Directories can be accessed at 30 http://www.ietf.org/shadow.html. 32 This Internet-Draft will expire on December 16, 2003. 34 Copyright Notice 36 Copyright (C) The Internet Society (2003). All Rights Reserved. 38 Abstract 40 There is some known misbehavior of DNS authoritative servers when 41 they are queried for AAAA resource records. Such behavior can block 42 IPv4 communication which should actually be available, cause a 43 significant delay in name resolution, or even make a denial of 44 service attack. This memo describes details of the known cases and 45 discusses the effect. 47 1. Introduction 49 Many DNS clients (resolvers) that support IPv6 first search for AAAA 50 RRs (Resource Records) of a target host name, and then for A RRs of 51 the same name. This fallback mechanism is based on the DNS 52 specifications. Thus, if a DNS server which is responsible for the 53 name is not compliant to the specifications, unpleasant results can 54 happen. In some cases, for example, a web browser fails to connect 55 to a web server otherwise it could. In the following sections, this 56 memo describes some typical cases of the misbehavior, the rationale, 57 and (bad) effects of them. 59 This memo shows concrete implementations and domain names that may 60 cause problematic cases so that the behavior can be reproduced in a 61 practical environment. The examples are for informational purposes 62 only, and the authors do not intend accusation against any 63 implementations or zone administrators described in this memo. 65 2. Network Model 67 In this memo, we assume a typical network model of name resolution 68 environment using DNS. It consists of three components; stub 69 resolvers, caching servers, and authoritative servers. A stub 70 resolver issues a recursive query to a caching server, which then 71 handles the entire name resolution procedure recursively. The 72 caching server caches the result of the query as well as sends the 73 result to the stub resolver. The authoritative servers respond to 74 queries for names for which they have the authority, normally in a 75 non-recursive manner. 77 3. Expected Behavior 79 Suppose that an authoritative server has an A RR but not a AAAA for a 80 host name. Then the server should return a response to a query for a 81 AAAA RR of the name with the RCODE being 0 (indicating no error) and 82 with an empty answer section [1]. Such a response indicates that 83 there is at least one RR of a different type than AAAA for the 84 queried name, and the stub resolver can then look for A RRs. 86 This way, the caching server can cache the fact that the queried name 87 does not have a AAAA RR (but may have other types of RRs), and thus 88 can improve the response time to further queries for a AAAA RR of the 89 name. 91 4. Problematic Behaviors 93 There are some known cases not compliant to the expected behavior. 94 This section describes those problematic cases. 96 4.1 Return NXDOMAIN 98 This type of server returns a response with the RCODE being 3 99 (NXDOMAIN) to a query for a AAAA RR, indicating it does not have any 100 RRs of any type for the queried name. In fact, such a server 101 apparently returns NXDOMAIN to all queries except those for an A RR. 103 With this response, the stub resolver may immediately give up and 104 never fall back. Even if the resolver retries with a query for an A 105 RR, the negative response for the name has been cached in the caching 106 server, and the caching server will simply return the negative 107 response. As a result, the stub resolver considers this as a fatal 108 error in name resolution. 110 An example of this case was found by looking for a AAAA RR of 111 www.css.vtext.com at 66.174.3.4, although the implementation of the 112 authoritative server seemed to change to that described in the next 113 section. 115 4.2 Return NOTIMP 117 Other authoritative servers return a response with the RCODE being 4 118 (NOTIMP), indicating the servers do not support the requested type of 119 query. 121 This case is less harmful than the previous one; if the stub resolver 122 falls back to querying for an A RR, the caching server will process 123 the query correctly and return an appropriate response. 125 In this case, the caching server does not cache the fact that the 126 queried name has no AAAA RR, resulting in redundant queries for AAAA 127 RRs in the future. The behavior will waste network bandwidth and 128 increase the load of the authoritative server. 130 The current implementation of an authoritative server for 131 css.vtext.com looks to belong to this category. 133 Using SERVFAIL or FORMERR would cause the same effect, though the 134 authors have not seen such implementations yet. 136 4.3 Ignore Queries for AAAA 138 Some authoritative severs seem to ignore queries for a AAAA RR, 139 causing a delay to fall back to a query for an A RR. This behavior 140 may even cause a fatal timeout at the stub resolver. 142 This can be seen by trying to ask for a AAAA RR of "ftp-mozilla.gftp- 143 mozilla.netscape.com," which is an alias of ftp.mozilla.org, at 144 205.188.139.70. 146 Again, these servers apparently ignore all queries except those for 147 an A RR. 149 4.4 Return a Broken Response 151 Some other type of authoritative servers return broken responses to 152 AAAA queries. 154 An example of such a response can be seen by querying for a AAAA RR 155 of "www.gslb.mainichi.co.jp" at 210.173.172.2. This authoritative 156 server returns a response whose RR type is AAAA, but the length of 157 the RDATA is 4 bytes. The 4-byte data looks like the IPv4 address of 158 the queried host name. That is, the RR in the answer section would 159 be described like this: 161 www.gslb.mainichi.co.jp. 600 IN AAAA 210.158.208.73 163 which is, of course, bogus (or at least meaningless). 165 The same behavior can be found with the name vip.alt.ihp.sony.co.jp 166 (which is an alias of www.sony.co.jp) at 210.139.255.204. 168 BIND 8 caching servers transparently return the broken response (as 169 well as cache it) to the stub resolver. BIND 9 caching servers parse 170 the response by themselves, and send a separate response with the 171 RCODE being 2 (SERVFAIL). 173 In the former case, many stub resolvers consider this as a fatal 174 error, and do not fall back to querying for an A RR. This is the 175 case for the BIND resolver library and (reportedly) that implemented 176 in Internet Explorer on Windows XP SP1. 178 In the latter case, if the stub resolver retries the query for an A 179 RR, it will get an appropriate response. 181 There are reportedly other kinds of resolver implementations that can 182 fall back to queries for an A RR even in the first case, but the 183 authors actually do not know of such implementations. 185 4.5 Make a Delegation Loop 187 Some authoritative servers constantly indicate a (loop) delegation 188 for any queries except those for an A RR. 190 For example, such a server would return a response to a query for a 191 AAAA RR of "www.bad.example" as follows: 193 www.bad.example. IN NS ns.foo.bad.example. 194 ns.foo.bad.example. IN A 10.0.0.1 196 Then the caching server will ask 10.0.0.1 for a AAAA RR of 197 "www.bad.example" and see the same answer. 199 Caching servers interpret this as a lame delegation, and return a 200 response with the RCODE being 2 (SERVFAIL) to the stub resolver. 201 Furthermore, BIND 8 caching servers record the authoritative server 202 as lame and will not use it for a certain period of time. BIND 9 203 caching servers relax the rule a little bit. They basically try to 204 avoid using the lame server, but still continue to try it as a last 205 resort. 207 With a BIND 8 caching server, even if the stub resolver falls back to 208 querying for an A RR, the caching server will simply return a 209 response with the RCODE being SERVFAIL, since all the servers are 210 known to be "lame." 212 This behavior was previously found by asking for a AAAA RR of 213 "www.united.com" at 64.95.89.4, which has recently been fixed. 215 5. Security Considerations 217 The CERT/CC pointed out that the response with NXDOMAIN described in 218 Section 4.1 can be used for a denial of service attack [2]. The same 219 argument applies to the cases of "broken responses" and "delegation 220 loop" described in Section 4.4 and Section 4.5, respectively. 222 6. Acknowledgements 224 Erik Nordmark encouraged the authors to publish this document as an 225 Internet Draft. Akira Kato and Paul Vixie reviewed a preliminary 226 version of this draft. 228 Normative References 230 [1] Mockapetris, P., "DOMAIN NAMES - CONCEPTS AND FACILITIES", RFC 231 1034, November 1987. 233 [2] The CERT Coordination Center, "Incorrect NXDOMAIN responses from 234 AAAA queries could cause denial-of-service conditions", March 235 2003, . 237 Authors' Addresses 239 MORISHITA Orange Yasuhiro 240 Research and Development Department, Japan Registry Service Co.,Ltd. 241 Fuundo Bldg 3F, 1-2 Kanda-Ogawamachi 242 Chiyoda-ku, Tokyo 101-0052 243 Japan 245 EMail: yasuhiro@jprs.co.jp 247 JINMEI Tatuya 248 Corporate Research & Development Center, Toshiba Corporation 249 1 Komukai Toshiba-cho, Saiwai-ku 250 Kawasaki-shi, Kanagawa 212-8582 251 Japan 253 EMail: jinmei@isl.rdc.toshiba.co.jp 255 Full Copyright Statement 257 Copyright (C) The Internet Society (2003). All Rights Reserved. 259 This document and translations of it may be copied and furnished to 260 others, and derivative works that comment on or otherwise explain it 261 or assist in its implementation may be prepared, copied, published 262 and distributed, in whole or in part, without restriction of any 263 kind, provided that the above copyright notice and this paragraph are 264 included on all such copies and derivative works. However, this 265 document itself may not be modified in any way, such as by removing 266 the copyright notice or references to the Internet Society or other 267 Internet organizations, except as needed for the purpose of 268 developing Internet standards in which case the procedures for 269 copyrights defined in the Internet Standards process must be 270 followed, or as required to translate it into languages other than 271 English. 273 The limited permissions granted above are perpetual and will not be 274 revoked by the Internet Society or its successors or assigns. 276 This document and the information contained herein is provided on an 277 "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING 278 TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING 279 BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION 280 HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF 281 MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 283 Acknowledgement 285 Funding for the RFC Editor function is currently provided by the 286 Internet Society.