3 Network Working Group S. Woolf
4 Internet-Draft Internet Systems Consortium, Inc.
5 Expires: September 14, 2005 D. Conrad
10 Identifying an Authoritative Name Server
11 draft-ietf-dnsop-serverid-04
15 This document is an Internet-Draft and is subject to all provisions
16 of Section 3 of RFC 3667. By submitting this Internet-Draft, each
17 author represents that any applicable patent or other IPR claims of
18 which he or she is aware have been or will be disclosed, and any of
19 which he or she become aware will be disclosed, in accordance with
22 Internet-Drafts are working documents of the Internet Engineering
23 Task Force (IETF), its areas, and its working groups. Note that
24 other groups may also distribute working documents as
27 Internet-Drafts are draft documents valid for a maximum of six months
28 and may be updated, replaced, or obsoleted by other documents at any
29 time. It is inappropriate to use Internet-Drafts as reference
30 material or to cite them other than as "work in progress."
32 The list of current Internet-Drafts can be accessed at
33 http://www.ietf.org/ietf/1id-abstracts.txt.
35 The list of Internet-Draft Shadow Directories can be accessed at
36 http://www.ietf.org/shadow.html.
38 This Internet-Draft will expire on September 14, 2005.
42 Copyright (C) The Internet Society (2005).
46 With the increased use of DNS anycast, load balancing, and other
47 mechanisms allowing more than one DNS name server to share a single
48 IP address, it is sometimes difficult to tell which of a pool of name
49 servers has answered a particular query. A standardized mechanism to
50 determine the identity of a name server responding to a particular
54 Woolf & Conrad Expires September 14, 2005 [Page 1]
56 Internet-Draft Identifying an Authoritative Name Server March 2005
59 query would be useful, particularly as a diagnostic aid. Existing ad
60 hoc mechanisms for addressing this concern are not adequate. This
61 document attempts to describe the common ad hoc solution to this
62 problem, including its advantages and disadvantages, and to
63 characterize an improved mechanism.
110 Woolf & Conrad Expires September 14, 2005 [Page 2]
112 Internet-Draft Identifying an Authoritative Name Server March 2005
117 With the increased use of DNS anycast, load balancing, and other
118 mechanisms allowing more than one DNS name server to share a single
119 IP address, it is sometimes difficult to tell which of a pool of name
120 servers has answered a particular query. A standardized mechanism to
121 determine the identity of a name server responding to a particular
122 query would be useful, particularly as a diagnostic aid.
124 Unfortunately, existing ad-hoc mechanisms for providing such
125 identification have some shortcomings, not the least of which is the
126 lack of prior analysis of exactly how such a mechanism should be
127 designed and deployed. This document describes the existing
128 convention used in one widely deployed implementation of the DNS
129 protocol and discusses requirements for an improved solution to the
166 Woolf & Conrad Expires September 14, 2005 [Page 3]
168 Internet-Draft Identifying an Authoritative Name Server March 2005
173 Identifying which name server is responding to queries is often
174 useful, particularly in attempting to diagnose name server
175 difficulties. However, relying on the IP address of the name server
176 has become more problematic due the deployment of various load
177 balancing solutions, including the use of shared unicast addresses as
178 documented in [RFC3258].
180 An unfortunate side effect of these load balancing solutions, and
181 some changes in management practices as the public Internet has
182 evolved, is that traditional methods of determining which server is
183 responding can be unreliable. Specifically, non-DNS methods such as
184 ICMP ping, TCP connections, or non-DNS UDP packets (such as those
185 generated by tools such as "traceroute"), etc., can end up going to a
186 different server than that which receives the DNS queries.
188 There is a well-known and frequently-used technique for determining
189 an identity for a nameserver more specific than the
190 possibly-non-unique "server that answered my query". The widespread
191 use of the existing convention suggests a need for a documented,
192 interoperable means of querying the identity of a nameserver that may
193 be part of an anycast or load-balancing cluster. At the same time,
194 however, it also has some drawbacks that argue against standardizing
195 it as it's been practiced so far.
222 Woolf & Conrad Expires September 14, 2005 [Page 4]
224 Internet-Draft Identifying an Authoritative Name Server March 2005
227 3. Existing Conventions
229 Recent versions of the commonly deployed Berkeley Internet Name
230 Domain implementation of the DNS protocol suite from the Internet
231 Software Consortium [BIND] support a way of identifying a particular
232 server via the use of a standard, if somewhat unusual, DNS query.
233 Specifically, a query to a late model BIND server for a TXT resource
234 record in class 3 (CHAOS) for the domain name "HOSTNAME.BIND." will
235 return a string that can be configured by the name server
236 administrator to provide a unique identifier for the responding
237 server (defaulting to the value of a gethostname() call). This
238 mechanism, which is an extension of the BIND convention of using
239 CHAOS class TXT RR queries to sub-domains of the "BIND." domain for
240 version information, has been copied by several name server vendors.
242 For reference, the other well-known name used by recent versions of
243 BIND within the CHAOS class "BIND." domain is "VERSION.BIND." A
244 query for a TXT RR for this name will return an administratively
245 defined string which defaults to the version of the server
246 responding. This is, however, not generally implemented by other
251 There are several valuable attributes to this mechanism, which
252 account for its usefulness.
253 1. The "hostname.bind" query response mechanism is within the DNS
254 protocol itself. An identification mechanism that relies on the
255 DNS protocol is more likely to be successful (although not
256 guaranteed) in going to the same machine as a "normal" DNS query.
257 2. Since the identity information is requested and returned within
258 the DNS protocol, it doesn't require allowing any other query
259 mechanism to the server, such as holes in firewalls for
260 otherwise-unallowed ICMP Echo requests. Thus it does not require
261 any special exceptions to site security policy.
262 3. It is simple to configure. An administrator can easily turn on
263 this feature and control the results of the relevant query.
264 4. It allows the administrator complete control of what information
265 is given out in the response, minimizing passive leakage of
266 implementation or configuration details. Such details are often
267 considered sensitive by infrastructure operators.
271 At the same time, there are some forbidding drawbacks to the
272 VERSION.BIND mechanism that argue against standardizing it as it
278 Woolf & Conrad Expires September 14, 2005 [Page 5]
280 Internet-Draft Identifying an Authoritative Name Server March 2005
283 1. It requires an additional query to correlate between the answer
284 to a DNS query under normal conditions and the supposed identity
285 of the server receiving the query. There are a number of
286 situations in which this simply isn't reliable.
287 2. It reserves an entire class in the DNS (CHAOS) for what amounts
288 to one zone. While CHAOS class is defined in [RFC1034] and
289 [RFC1035], it's not clear that supporting it solely for this
290 purpose is a good use of the namespace or of implementation
292 3. It is implementation specific. BIND is one DNS implementation.
293 At the time of this writing, it is probably the most prevalent
294 for authoritative servers. This does not justify standardizing
295 on its ad hoc solution to a problem shared across many operators
298 The first of the listed disadvantages is technically the most
299 serious. It argues for an attempt to design a good answer to the
300 problem that "I need to know what nameserver is answering my
301 queries", not simply a convenient one.
334 Woolf & Conrad Expires September 14, 2005 [Page 6]
336 Internet-Draft Identifying an Authoritative Name Server March 2005
339 4. Characteristics of an Implementation Neutral Convention
341 The discussion above of advantages and disadvantages to the
342 HOSTNAME.BIND mechanism suggest some requirements for a better
343 solution to the server identification problem. These are summarized
344 here as guidelines for any effort to provide appropriate protocol
346 1. The mechanism adopted MUST be in-band for the DNS protocol. That
347 is, it needs to allow the query for the server's identifying
348 information to be part of a normal, operational query. It SHOULD
349 also permit a separate, dedicated query for the server's
350 identifying information.
351 2. The new mechanism SHOULD not require dedicated namespaces or
352 other reserved values outside of the existing protocol mechanisms
353 for these, i.e. the OPT pseudo-RR. In particular, it should not
354 propagate the existing drawback of requiring support for a CLASS
355 and top level domain in the authoritative server (or the querying
357 3. Support for the identification functionality SHOULD be easy to
358 implement and easy to enable. It MUST be easy to disable and
359 SHOULD lend itself to access controls on who can query for it.
360 4. It should be possible to return a unique identifier for a server
361 without requiring the exposure of information that may be
362 non-public and considered sensitive by the operator, such as a
363 hostname or unicast IP address maintained for administrative
365 5. The identification mechanism SHOULD NOT be
366 implementation-specific.
390 Woolf & Conrad Expires September 14, 2005 [Page 7]
392 Internet-Draft Identifying an Authoritative Name Server March 2005
395 5. IANA Considerations
397 This document proposes no specific IANA action. Protocol extensions,
398 if any, to meet the requirements described are out of scope for this
399 document. Should such extensions be specified and adopted by normal
400 IETF process, the specification will include appropriate guidance to
446 Woolf & Conrad Expires September 14, 2005 [Page 8]
448 Internet-Draft Identifying an Authoritative Name Server March 2005
451 6. Security Considerations
453 Providing identifying information as to which server is responding to
454 a particular query from a particular location in the Internet can be
455 seen as information leakage and thus a security risk. This motivates
456 the suggestion above that a new mechanism for server identification
457 allow the administrator to disable the functionality altogether or
458 partially restrict availability of the data. It also suggests that
459 the serverid data should not be readily correlated with a hostname or
460 unicast IP address that may be considered private to the nameserver
461 operator's management infrastructure.
463 Propagation of protocol or service meta-data can sometimes expose the
464 application to denial of service or other attack. As DNS is a
465 critically important infrastructure service for the production
466 Internet, extra care needs to be taken against this risk for
467 designers, implementors, and operators of a new mechanism for server
502 Woolf & Conrad Expires September 14, 2005 [Page 9]
504 Internet-Draft Identifying an Authoritative Name Server March 2005
509 The technique for host identification documented here was initially
510 implemented by Paul Vixie of the Internet Software Consortium in the
511 Berkeley Internet Name Daemon package. Comments and questions on
512 earlier drafts were provided by Bob Halley, Brian Wellington, Andreas
513 Gustafsson, Ted Hardie, Chris Yarnell, Randy Bush, and members of the
514 ICANN Root Server System Advisory Committee. The newest version
515 takes a significantly different direction from previous versions,
516 owing to discussion among contributors to the DNSOP working group and
517 others, particularly Olafur Gudmundsson, Ed Lewis, Bill Manning, Sam
518 Weiler, and Rob Austein.
558 Woolf & Conrad Expires September 14, 2005 [Page 10]
560 Internet-Draft Identifying an Authoritative Name Server March 2005
563 Intellectual Property Statement
565 The IETF takes no position regarding the validity or scope of any
566 Intellectual Property Rights or other rights that might be claimed to
567 pertain to the implementation or use of the technology described in
568 this document or the extent to which any license under such rights
569 might or might not be available; nor does it represent that it has
570 made any independent effort to identify any such rights. Information
571 on the procedures with respect to rights in RFC documents can be
572 found in BCP 78 and BCP 79.
574 Copies of IPR disclosures made to the IETF Secretariat and any
575 assurances of licenses to be made available, or the result of an
576 attempt made to obtain a general license or permission for the use of
577 such proprietary rights by implementers or users of this
578 specification can be obtained from the IETF on-line IPR repository at
579 http://www.ietf.org/ipr.
581 The IETF invites any interested party to bring to its attention any
582 copyrights, patents or patent applications, or other proprietary
583 rights that may cover technology that may be required to implement
584 this standard. Please address the information to the IETF at
588 Disclaimer of Validity
590 This document and the information contained herein are provided on an
591 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
592 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
593 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED,
594 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
595 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
596 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
601 Copyright (C) The Internet Society (2005). This document is subject
602 to the rights, licenses and restrictions contained in BCP 78, and
603 except as set forth therein, the authors retain all their rights.
608 Funding for the RFC Editor function is currently provided by the
614 Woolf & Conrad Expires September 14, 2005 [Page 11]