[Docs] [txt|pdf] [Tracker] [WG] [Email] [Diff1] [Diff2] [Nits]

Versions: 00 01 02 03 04 05 06 07 08 09 10 11 RFC 4592

DNSEXT Working Group                                          E. Lewis
INTERNET DRAFT                                                 NeuStar
Expiration Date: January 6, 2006                          July 6, 2005
Updates RFC 1034, RFC 2672

                             The Role of Wildcards
                           in the Domain Name System
                     draft-ietf-dnsext-wcard-clarify-08.txt

Status of this Memo

     By submitting this Internet-Draft, each author represents that
     any applicable patent or other IPR claims of which he or she is
     aware have been or will be disclosed, and any of which he or she
     becomes aware will be disclosed, in accordance with Section 6 of
     BCP 79.

     Internet-Drafts are working documents of the Internet Engineering
     Task Force (IETF), its areas, and its working groups.  Note that
     other groups may also distribute working documents as Internet-
     Drafts.

     Internet-Drafts are draft documents valid for a maximum of six
     months and may be updated, replaced, or obsoleted by other
     documents at any time.  It is inappropriate to use Internet-Drafts
     as reference material or to cite them other than as "work in
     progress."

     The list of current Internet-Drafts can be accessed at
       http://www.ietf.org/ietf/1id-abstracts.txt

     The list of Internet-Draft Shadow Directories can be accessed at
       http://www.ietf.org/shadow.html

     This Internet-Draft will expire on January 6, 2006.

Copyright Notice

     Copyright (C) The Internet Society (2005).

Abstract

     This is an update to the wildcard definition of RFC 1034.  The
     interaction with wildcards and CNAME is changed, an error
     condition removed, and the words defining some concepts central
     to wildcards are changed.  The overall goal is not to change
     wildcards, but to refine the definition of RFC 1034.

Table of Contents

1.    Introduction
1.1   Motivation
1.2   The Original Definition
1.3   Roadmap to This Document
1.3.1 New Terms
1.3.2 Changed Text
1.3.3 Considerations with Special Types
1.4   Standards Terminology
2.    Wildcard Syntax
2.1   Identifying a Wildcard
2.1.1 Wild Card Domain Name and Asterisk Label
2.1.2 Asterisks and Other Characters
2.1.3 Non-terminal Wild Card Domain Names
2.2   Existence Rules
2.2.1 An Example
2.2.2 Empty Non-terminals
2.2.3 Yet Another Definition of Existence
2.3   When is a Wild Card Domain Name Not Special
3.    Impact of a Wild Card Domain Name On a Response
3.1   Step 2
3.2   Step 3
3.3   Part 'c'
3.3.1 Closest Encloser and the Source of Synthesis
3.3.2 Closest Encloser and Source of Synthesis Examples
3.3.3 Type Matching
4.    Considerations with Special Types
4.1   SOA RRSet at a Wild Card Domain Name
4.2   NS RRSet at a Wild Card Domain Name
4.2.1 Discarded Notions
4.3   CNAME RRSet at a Wild Card Domain Name
4.4   DNAME RRSet at a Wild Card Domain Name
4.5   SRV RRSet at a Wild Card Domain Name
4.6   DS RRSet at a Wild Card Domain Name
4.7   NSEC RRSet at a Wild Card Domain Name
4.8   RRSIG at a Wild Card Domain Name
4.9   Empty Non-terminal Wild Card Domain Name
5.    Security Considerations
6.    IANA Considerations
7.    References
8.    Editor
9.    Others Contributing to the Document
10.   Trailing Boilerplate

1. Introduction

     In RFC 1034 [RFC1034], sections 4.3.2 and 4.3.3 describe the
     synthesis of answers from special resource records called
     wildcards.  The definition in RFC 1034 is incomplete and has
     proven to be confusing.  This document describes the wildcard
     synthesis by adding to the discussion and making limited
     modifications.  Modifications are made to close inconsistencies
     that have led to interoperability issues.  This description
     does not expand the service intended by the original definition.

     Staying within the spirit and style of the original documents,
     this document avoids specifying rules for DNS implementations
     regarding wildcards.  The intention is to only describe what is
     needed for interoperability, not restrict implementation choices.
     In addition, consideration is given to minimize any backwards
     compatibility issues with implementations that comply with RFC
     1034's definition.

     This document is focused on the concept of wildcards as defined
     in RFC 1034.  Nothing is implied regarding alternative means of
     synthesizing resource record sets, nor are alternatives discussed.

1.1 Motivation

     Many DNS implementations diverge, in different ways, from the
     original definition of wildcards.  Although there is clearly a
     need to clarify the original documents in light of this alone,
     the impetus for this document lay in the engineering of the DNS
     security extensions [RFC4033].  With an unclear definition of
     wildcards the design of authenticated denial became entangled.

     This document is intended to limit its changes, documenting only
     those based on implementation experience, and to remain as close
     to the original document as possible.  To reinforce that this
     document is meant to clarify and adjust and not redefine wildcards,
     relevant sections of RFC 1034 are repeated verbatim to facilitate
     comparison of the old and new text.

1.2 The Original Definition

     The defintion of the wildcard concept is comprised by the
     documentation of the algorithm by which a name server prepares
     a response (in RFC 1034's section 4.3.2) and the way in which
     a resource record (set) is identified as being a source of
     synthetic data (section 4.3.3).

     This is the definition of the term "wildcard" as it appears in
     RFC 1034, section 4.3.3.

# In the previous algorithm, special treatment was given to RRs with
# owner names starting with the label "*".  Such RRs are called
# wildcards. Wildcard RRs can be thought of as instructions for
# synthesizing RRs.  When the appropriate conditions are met, the name
# server creates RRs with an owner name equal to the query name and
# contents taken from the wildcard RRs.

     This passage follows the algorithm in which the term wildcard
     is first used.   In this definition, wildcard refers to resource
     records.  In other usage, wildcard has referred to domain names,
     and it has been used to describe the operational practice of
     relying on wildcards to generate answers.  It is clear from this
     that there is a need to define clear and unambiguous terminology
     in the process of discussing wildcards.

     The mention of the use of wildcards in the preparation of a
     response is contained in step 3c of RFC 1034's section 4.3.2
     entitled "Algorithm."  Note that "wildcard" does not appear in
     the algorithm, instead references are made to the "*" label.
     The portion of the algorithm relating to wildcards is
     deconstructed in detail in section 3 of this document, this is
     the beginning of the relevant portion of the "Algorithm."

#    c. If at some label, a match is impossible (i.e., the
#       corresponding label does not exist), look to see if [...]
#       the "*" label exists.

     The scope of this document is the RFC 1034 definition of
     wildcards and the implications of updates to those documents,
     such as DNSSEC.  Alternate schemes for synthesizing answers are
     not considered.  (Note that there is no reference listed.  No
     document is known to describe any alternate schemes, although
     there has been some mention of them in mailing lists.)

1.3 Roadmap to This Document

     This document accomplishes these three items.
     o Defines new terms
     o Makes minor changes to avoid conflicting concepts
     o Describes the actions of certain resource records as wildcards

1.3.1 New Terms

     To help in discussing what resource records are wildcards, two
     terms will be defined - "asterisk label" and "wild card domain
     name".  These are defined in section 2.1.1.

     To assist in clarifying the role of wildcards in the name server
     algorithm in RFC 1034, 4.3.2, "source of synthesis" and "closest
     encloser" are defined.  These definitions are in section 3.3.2.
     "Label match" is defined in section 3.2.

     The new terms are used to make discussions of wildcards clearer.
     Terminology doesn't directly have an impact on implementations.

1.3.2 Changed Text

     The definition of "existence" is changed superficially.  This
     change will not be apparent to implementations; it is needed to
     make descriptions more precise.  The change appears in section
     2.2.3.

     RFC 1034, section 4.3.3., seems to prohibit having two asterisk
     labels in a wildcard owner name.  With this document the
     restriction is removed entirely.  This change and its implications
     are in section 2.1.3.

     The actions when a source of synthesis owns a CNAME RR are
     changed to mirror the actions if an exact match name owns a
     CNAME RR.  This is an addition to the words in RFC 1034,
     section 4.3.2, step 3, part c.  The discussion of this is in
     section 3.3.3.

     Only the latter change represents an impact to implementations.
     The definition of existence is not a protocol impact.  The change
     to the restriction on names is unlikely to have an impact, as
     RFC 1034 contained no specification on when and how to enforce the
     restriction.

1.3.3 Considerations with Special Types

     This document describes semantics of wildcard RRSets for
     "interesting" types as well as empty non-terminal wildcards.
     Understanding these situations in the context of wildcards has
     been clouded because these types incur special processing if
     they are the result of an exact match.  This discussion is in
     section 4.

     These discussions do not have an implementation impact, they cover
     existing knowledge of the types, but to a greater level of detail.

1.4 Standards Terminology

     This document does not use terms as defined in "Key words for use
     in RFCs to Indicate Requirement Levels." [RFC2119]

     Quotations of RFC 1034 are denoted by a '#' in the leftmost
     column.  References to section "4.3.2" are assumed to refer
     to RFC 1034's section 4.3.2, simply titled "Algorithm."

2. Wildcard Syntax

     The syntax of a wildcard is the same as any other DNS resource
     record, across all classes and types.  The only significant
     feature is the owner name.

     Because wildcards are encoded as resource records with special
     names, they are included in zone transfers and incremental zone
     transfers[RFC1995] just as non-wildcard resource records are.
     This feature has been underappreciated until discussions on
     alternative approaches to wildcards appeared on mailing lists.

2.1 Identifying a Wildcard

     To provide a more accurate description of wildcards, the
     definition has to start with a discussion of the domain names
     that appear as owners.  Two new terms are needed, "Asterisk
     Label" and "Wild Card Domain Name."

2.1.1 Wild Card Domain Name and Asterisk Label

     A "wild card domain name" is defined by having its initial
     (i.e., left-most or least significant) label be, in binary format:

          0000 0001 0010 1010 (binary) = 0x01 0x2a (hexadecimal)

     The first octet is the normal label type and length for a 1 octet
     long label, the second octet is the ASCII representation [RFC20]
     for the '*' character.

     A descriptive name of a label equaling that value is an "asterisk
     label."

     RFC 1034's definition of wildcard would be "a resource record
     owned by a wild card domain name."

2.1.2 Asterisks and Other Characters

     No label values other than that in section 2.1.1 are asterisk
     labels, hence names beginning with other labels are never wild
     card domain names.  Labels such as 'the*' and '**' are not
     asterisk labels so these labels do not start wild card domain
     names.

2.1.3 Non-terminal Wild Card Domain Names

     In section 4.3.3, the following is stated:

# ..........................  The owner name of the wildcard RRs is of
# the form "*.<anydomain>", where <anydomain> is any domain name.
# <anydomain> should not contain other * labels......................

     The restriction is now removed.  The original documentation of it
     is incomplete and the restriction does not serve any purpose given
     years of operational experience.

     There are three possible reasons for putting the restriction in
     place, but none of the three has held up over time.  One is
     that the restriction meant that there would never be subdomains
     of wild card domain names, but the restriciton as stated still
     permits "example.*.example." for instance.  Another is that
     wild card domain names are not intended to be empty non-terminals,
     but this situation does not disrupt the algorithm in 4.3.2.
     Finally, "nested" wild card domain names are not ambiguous once
     the concept of the closest encloser had been documented.

     A wild card domain name can have subdomains.  There is no need
     to inspect the subdomains to see if there is another asterisk
     label in any subdomain.

     A wild card domain name can be an empty non-terminal.  (See the
     upcoming sections on empty non-terminals.)  In this case, any
     lookup encountering it will terminate as would any empty
     non-terminal match.

2.2 Existence Rules

     The notion that a domain name 'exists' is mentioned in the
     definition of wildcards.  In section 4.3.3 of RFC 1034:

# Wildcard RRs do not apply:
#
...
#   - When the query name or a name between the wildcard domain and
#     the query name is know[n] to exist.  For example, if a wildcard

     "Existence" is therefore an important concept in the understanding
     of wildcards.  Unfortunately, the definition of what exists, in RFC
     1034, is unlcear.  So, in sections 2.2.2. and 2.2.3, another look is
     taken at the definition of existence.

2.2.1 An Example

     To illustrate what is meant by existence consider this complete
     zone:

       $ORIGIN example.
       example.                 3600 IN  SOA   <SOA RDATA>
       example.                 3600     NS    ns.example.com.
       example.                 3600     NS    ns.example.net.
       *.example.               3600     TXT   "this is a wild card"
       *.example.               3600     MX    10 host1.example.
       sub.*.example.           3600     TXT   "this is not a wild card"
       host1.example.           3600     A     192.0.4.1
       _ssh._tcp.host1.example. 3600     SRV  <SRV RDATA>
       _ssh._tcp.host2.example. 3600     SRV  <SRV RDATA>
       subdel.example.          3600     NS   ns.example.com.
       subdel.example.          3600     NS   ns.example.net.

     A look at the domain names in a tree structure is helpful:

                                   |
                   -------------example------------
                  /           /         \          \
                 /           /           \          \
                /           /             \          \
               *          host1          host2      subdel
               |            |             |
               |            |             |
              sub         _tcp          _tcp
                            |             |
                            |             |
                          _ssh          _ssh

     The following responses would be synthesized from one of the
     wildcards in the zone:

         QNAME=host3.example. QTYPE=MX, QCLASS=IN
              the answer will be a "host3.example. IN MX ..."

         QNAME=host3.example. QTYPE=A, QCLASS=IN
              the answer will reflect "no error, but no data"
              because there is no A RR set at '*.example.'

         QNAME=foo.bar.example. QTYPE=TXT, QCLASS=IN
              the answer will be "foo.bar.example. IN TXT ..."
              because bar.example. does not exist, but the wildcard
              does.

     The following responses would not be synthesized from any of the
     wildcards in the zone:

         QNAME=host1.example., QTYPE=MX, QCLASS=IN
              because host1.example. exists

         QNAME=sub.*.example., QTYPE=MX, QCLASS=IN
              because sub.*.example. exists

         QNAME=_telnet._tcp.host1.example., QTYPE=SRV, QCLASS=IN
              because _tcp.host1.example. exists (without data)

         QNAME=host.subdel.example., QTYPE=A, QCLASS=IN
              because subdel.example. exists (and is a zone cut)

         QNAME=ghost.*.example., QTYPE=MX, QCLASS=IN
              because *.example. exists

     The final example highlights one common misconception about
     wildcards.  A wildcard "blocks itself" in the sense that a
     wildcard does not match its own subdomains.  I.e. "*.example."
     does not match all names in the "example." zone, it fails to
     match the names below "*.example." To cover names under
     "*.example.", another wild card domain name is needed -
     "*.*.example." - which covers all but it's own subdomains.

2.2.2 Empty Non-terminals

     Empty non-terminals [RFC2136, Section 7.16] are domain names
     that own no resource records but have subdomains that do.  In
     section 2.2.1, "_tcp.host1.example." is an example of a empty
     non-terminal name.  Empty non-terminals are introduced by this
     text in section 3.1 of RFC 1034:

# The domain name space is a tree structure.  Each node and leaf on
# the tree corresponds to a resource set (which may be empty).  The
# domain system makes no distinctions between the uses of the
# interior nodes and leaves, and this memo uses the term "node" to
# refer to both.

     The parenthesized "which may be empty" specifies that empty non-
     terminals are explicitly recognized, and that empty non-terminals
     "exist."

     Pedantically reading the above paragraph can lead to an
     interpretation that all possible domains exist - up to the
     suggested limit of 255 octets for a domain name [RFC1035].
     For example, www.example. may have an A RR, and as far as is
     practically concerned, is a leaf of the domain tree.  But the
     definition can be taken to mean that sub.www.example. also
     exists, albeit with no data.  By extension, all possible domains
     exist, from the root on down.

     As RFC 1034 also defines "an authoritative name error indicating
     that the name does not exist" in section 4.3.1, so this apparently
     is not the intent of the original definition, justifying the
     need for an updated definition in the next section.

2.2.3 Yet Another Definition of Existence

     RFC1034's wording is fixed by the following paragraph:

     The domain name space is a tree structure.  Nodes in the tree
     either own at least one RRSet and/or have descendants that
     collectively own at least one RRSet.  A node may exist with no
     RRSets only if it has descendents that do, this node is an empty
     non-terminal.

     A node with no descendants is a leaf node.  Empty leaf nodes do
     not exist.

     Note that at a zone boundary, the domain name owns data,
     including the NS RR set.  In the delegating zone, the NS RR
     set is not authoritative, but that is of no consequence here.
     The domain name owns data, therefore, it exists.

2.3 When is a Wild Card Domain Name Not Special

     When a wild card domain name appears in a message's query section,
     no special processing occurs.  An asterisk label in a query name
     only matches a single, corresponding asterisk label in the
     existing zone tree when the 4.3.2 algorithm is being followed.

     When a wild card domain name appears in the resource data of a
     record, no special processing occurs.  An asterisk label in that
     context literally means just an asterisk.

3. Impact of a Wild Card Domain Name On a Response

     RFC 1034's description of how wildcards impact response
     generation is in its section 4.3.2.  That passage contains the
     algorithm followed by a server in constructing a response.
     Within that algorithm, step 3, part 'c' defines the behavior of
     the wildcard.

     The algorithm in section 4.3.2. is not intended to be pseudo-code,
     i.e., its steps are not intended to be followed in strict order.
     The "algorithm" is a suggested means of implementing the
     requirements.  As such, in step 3, parts a, b, and c, do not have
     to be implemented in that order, provided that the result of the
     implemented code is compliant with the protocol's specification.

3.1 Step 2

     Step 2 of the section 4.3.2 reads:

#   2. Search the available zones for the zone which is the nearest
#      ancestor to QNAME.  If such a zone is found, go to step 3,
#      otherwise step 4.

     In this step, the most appropriate zone for the response is
     chosen.  The significance of this step is that it means all of
     step 3 is being performed within one zone.  This has significance
     when considering whether or not an SOA RR can be ever be used for
     synthesis.

3.2 Step 3

     Step 3 is dominated by three parts, labelled 'a', 'b', and 'c'.
     But the beginning of the step is important and needs explanation.

#   3. Start matching down, label by label, in the zone.  The
#      matching process can terminate several ways:

     The word 'matching' refers to label matching.  The concept
     is based in the view of the zone as the tree of existing names.
     The query name is considered to be an ordered sequence of
     labels - as if the name were a path from the root to the owner
     of the desired data.  (Which it is - 3rd paragraph of RFC 1034,
     section 3.1.)

     The process of label matching a query name ends in exactly one of
     three choices, the parts 'a', 'b', and 'c'.  Either the name is
     found, the name is below a cut point, or the name is not found.

     Once one of the parts is chosen, the other parts are not
     considered.  (E.g., do not execute part 'c' and then change
     the execution path to finish in part 'b'.)  The process of label
     matching is also done independent of the query type (QTYPE).

     Parts 'a' and 'b' are not an issue for this clarification as they
     do not relate to record synthesis.  Part 'a' is an exact match
     that results in an answer, part 'b' is a referral.

3.3 Part 'c'

     The context of part 'c' is that the process of label matching the
     labels of the query name has resulted in a situation in which
     there is no corresponding label in the tree.  It is as if the
     lookup has "fallen off the tree."

#     c. If at some label, a match is impossible (i.e., the
#        corresponding label does not exist), look to see if [...]
#        the "*" label exists.

     To help describe the process of looking 'to see if [...] the "*"
     label exists' a term has been coined to describe the last domain
     (node) matched.  The term is "closest encloser."

3.3.1 Closest Encloser and the Source of Synthesis

     The closest encloser is the node in the zone's tree of existing
     domain names that has the most labels matching the query name
     (consecutively, counting from the root label downward). Each match
     is a "label match" and the order of the labels is the same.

     The closest encloser is, by definition, an existing name in the
     zone.  The closest encloser might be an empty non-terminal or even
     be a wild card domain name itself.  In no circumstances is the
     closest encloser to be used to synthesize records for the current
     query.

     The source of synthesis is defined in the context of a query
     process as that wild card domain name immediately descending
     from the closest encloser, provided that this wild card domain
     name exists.  "Immediately descending" means that the source
     of synthesis has a name of the form:
           <asterisk label>.<closest encloser>.
     A source of synthesis does not guarantee having a RRSet to use
     for synthesis.  The source of synthesis could be an empty
     non-terminal.

     If the source of synthesis does not exist (not on the domain
     tree), there will be no wildcard synthesis.  There is no search
     for an alternate.

     The important concept is that for any given lookup process, there
     is at most one place at which wildcard synthetic records can be
     obtained.  If the source of synthesis does not exist, the lookup
     terminates, the lookup does not look for other wildcard records.

3.3.2 Closest Encloser and Source of Synthesis Examples

     To illustrate, using the example zone in section 2.2.1 of this
     document, the following chart shows QNAMEs and the closest
     enclosers.

     QNAME                       Closest Encloser    Source of Synthesis
     host3.example.              example.            *.example.
     _telnet._tcp.host1.example. _tcp.host1.example. no source
     _telnet._tcp.host2.example. host2.example.      no source
     _telnet._tcp.host3.example. example.            *.example.
     _chat._udp.host3.example.   example.            *.example.
     foobar.*.example.           *.example.          no source

3.3.3 Type Matching

      RFC 1034 concludes part 'c' with this:

#            If the "*" label does not exist, check whether the name
#            we are looking for is the original QNAME in the query
#            or a name we have followed due to a CNAME.  If the name
#            is original, set an authoritative name error in the
#            response and exit.  Otherwise just exit.
#
#            If the "*" label does exist, match RRs at that node
#            against QTYPE.  If any match, copy them into the answer
#            section, but set the owner of the RR to be QNAME, and
#            not the node with the "*" label.  Go to step 6.

     The final paragraph covers the role of the QTYPE in the lookup
     process.

     Based on implementation feedback and similarities between step
     'a' and step 'c' a change to this passage has been made.

     The change is to add the following text to step 'c' prior to the
     instructions to "go to step 6":

              If the data at the source of synthesis is a CNAME, and
              QTYPE doesn't match CNAME, copy the CNAME RR into the
              answer section of the response changing the owner name
              to the QNAME, change QNAME to the canonical name in the
              CNAME RR, and go back to step 1.

     This is essentially the same text in step a covering the
     processing of CNAME RRSets.

4. Considerations with Special Types

     Sections 2 and 3 of this document discuss wildcard synthesis
     with respect to names in the domain tree and ignore the impact
     of types.  In this section, the implication of wildcards of
     specific types are discussed.  The types covered are those
     that have proven to be the most difficult to understand.  The
     types are SOA, NS, CNAME, DNAME, SRV, DS, NSEC, RRSIG and
     "none," i.e., empty non-terminal wild card domain names.

4.1 SOA RRSet at a Wild Card Domain Name

     A wild card domain name owning an SOA RRSet means that the
     domain is at the root of the zone (apex).  The domain can not
     be a source of synthesis because that is, by definition, a
     descendent node (of the closest encloser) and a zone apex is
     at the top of the zone.

     Although a wild card domain name owning an SOA RRSet can never
     be a source of synthesis, there is no reason to forbid the
     ownership of an SOA RRSet.

     E.g., given this zone:
            $ORIGIN *.example.
            @                 3600 IN  SOA   <SOA RDATA>
                              3600     NS    ns1.example.com.
                              3600     NS    ns1.example.net.
            www               3600     TXT   "the www txt record"

     A query for www.*.example.'s TXT record would still find the
     "the www txt record" answer.  The reason is that the asterisk
     label only becomes significant when section's 4.3.2, step 3
     part 'c' in in effect.

     Of course, there would need to be a delegation in the parent
     zone, "example." for this to work too.  This is covered in the
     next section.

4.2 NS RRSet at a Wild Card Domain Name

     With the definition of DNSSEC [RFC4033, RFC4034, RFC4035] now
     in place, the semantics of a wild card domain name owning an
     NS RRSet has come to be poorly defined.  The dilemma relates to
     a conflict between the rules for synthesis in part 'c' and the
     fact that the resulting synthesis generates a record for which
     the zone is not authoritative.  In a DNSSEC signed zone, the
     mechanics of signature management (generation and inclusion
     in a message) become unclear.

     After some lengthy discussions, there has been no clear "best
     answer" on how to document the semantics of such a situation.
     Barring such records from the DNS would require definition of
     rules for that, as well as introducing a restriction on records
     that were once legal.  Allowing such records and amending the
     process of signature management would entail complicating the
     DNSSEC definition.

     There is one more ingredient to the discussion, that being the
     utility of a wild card domain name owned NS RRSet.  Although
     there are cases of this use, it is an operational rarity.
     Expending effort to close this topic has proven to be an
     exercise in diminishing returns.

     In summary, there is no definition given for wild card domain
     names owning an NS RRSet.  The semantics are left undefined until
     there is a clear need to have a set defined, and until there is
     a clear direction to proceed.  Operationally, inclusion of wild
     card NS RRSets in a zone is discouraged, but not barred.

4.2.1 Discarded Notions

     Prior to DNSSEC, a wild card domain name owning a NS RRSet
     appeared to be workable, and there are some instances in which
     it is found in deployments using implementations that support
     this.  Continuing to allow this in the specificaion is not
     tenable with DNSSEC.  The reason is that the synthesis of the
     NS RRSet is being done in a zone that has delegated away the
     responsibility for the name.  This "unauthorized" synthesis is
     not a problem for the base DNS protocol, but DNSSEC, in affirming
     the authorization model for DNS exposes the problem.

     Outright banning of wildcards of type NS is also untenable as
     the DNS protocol does not define how to handle "illegal" data.
     Implementations may choose not to load a zone, but there is no
     protocol definition.  The lack of the definition is complicated
     by having to cover dynamic update [RFC 2136], zone transfers,
     as well as loading at the master server.  The case of a client
     (resolver, cacheing server) getting a wildcard of type NS in
     a reply would also have to be considered.

     Given the daunting challenge of a complete definition of how to
     ban such records, dealing with existing implementations that
     permit the records today is a further complication.  There are
     uses of wild card domain name owning NS RRSets.

     One compromise proposed would have redefined wildcards of type
     NS to not be used in synthesis, this compromise fell apart
     because it would have required significant edits to the DNSSEC
     signing and validation work.  (Again, DNSSEC catches
     unauthorized data.)

     With no clear consensus forming on the solution to this dilemma,
     and the realization that wildcards of type NS are a rarity in
     operations, the best course of action is to leave this open-ended
     until "it matters."

4.3 CNAME RRSet at a Wild Card Domain Name

     The issue of a CNAME RRSet owned by a wild card domain name has
     prompted a suggested change to the last paragraph of step 3c of
     the algorithm in 4.3.2.  The changed text appears in section
     3.3.3 of this document.

4.4 DNAME RRSet at a Wild Card Domain Name

     Ownership of a DNAME [RFC2672] RRSet by a wild card domain name
     represents a threat to the coherency of the DNS and is to be
     avoided or outright rejected.  Such a DNAME RRSet represents
     non-deterministic synthesis of rules fed to different caches.
     As caches are fed the different rules (in an unpredictable
     manner) the caches will cease to be coherent.  ("As caches
     are fed" refers to the storage in a cache of records obtained
     in responses by recursive or iterative servers.)

     For example, assume one cache, responding to a recursive
     request, obtains the record:
        "a.b.example. DNAME foo.bar.example.net."
     and another cache obtains:
        "b.example.  DNAME foo.bar.example.net."
     both generated from the record:
        "*.example. DNAME foo.bar.example.net."
     by an authoritative server.

     The DNAME specification is not clear on whether DNAME records
     in a cache are used to rewrite queries.  In some interpretations,
     the rewrite occurs, in some, it is not.  Allowing for the
     occurrence of rewriting, queries for "sub.a.b.example. A" may
     be rewritten as "sub.foo.bar.tld. A" by the former caching
     server and may be rewritten as "sub.a.foo.bar.tld. A" by the
     latter.  Coherency is lost, an operational nightmare ensues.

     Another justification for banning or avoiding wildcard DNAME
     records is the observation that such a record could synthesize
     a DNAME owned by "sub.foo.bar.example." and "foo.bar.example."
     There is a restriction in the DNAME definition that no domain
     exist below a DNAME-owning domain, hence, the wildcard DNAME
     is not to be permitted.

4.5 SRV RRSet at a Wild Card Domain Name

     The definition of the SRV RRset is RFC 2782 [RFC2782].  In the
     definition of the record, there is some confusion over the term
     "Name."  The definition reads as follows:

# The format of the SRV RR
...
#    _Service._Proto.Name TTL Class SRV Priority Weight Port Target
...
#  Name
#   The domain this RR refers to.  The SRV RR is unique in that the
#   name one searches for is not this name; the example near the end
#   shows this clearly.

     Do not confuse the definition "Name" with the owner name.  I.e.,
     once removing the _Service and _Proto labels from the owner name
     of the SRV RRSet, what remains could be a wild card domain name
     but this is immaterial to the SRV RRSet.

     E.g.,  If an SRV record is:
        _foo._udp.*.example. 10800 IN SRV 0 1 9 old-slow-box.example.

     *.example is a wild card domain name and although it it the Name
     of the SRV RR, it is not the owner (domain name).  The owner
     domain name is "_foo._udp.*.example." which is not a wild card
     domain name.

     The confusion is likely based on the mixture of the specification
     of the SRV RR and the description of a "use case."

4.6 DS RRSet at a Wild Card Domain Name

     A DS RRSet owned by a wild card domain name is meaningless and
     harmless.  This statement is made in the context that an NS RRSet
     at a wild card domain name is undefined.  At a non-delegation
     point, a DS RRSet has no value (no corresponding DNSKEY RRSet
     will be used in DNSSEC validation).  If there is a synthesized
     DS RRSet, it alone will not be very useful as it exists in the
     context of a delegation point.

4.7 NSEC RRSet at a Wild Card Domain Name

     Wild card domain names in DNSSEC signed zones will have an NSEC
     RRSet.  Synthesis of these records will only occur when the
     query exactly matches the record.  Synthesized NSEC RR's will not
     be harmful as they will never be used in negative caching or to
     generate a negative response.

4.8 RRSIG at a Wild Card Domain Name

     RRSIG records will be present at a wild card domain name in a
     signed zone, and will be synthesized along with data sought in a
     query.  The fact that the owner name is synthesized is not a
     problem as the label count in the RRSIG will instruct the
     verifying code to ignore it.

4.9 Empty Non-terminal Wild Card Domain Name

     If a source of synthesis is an empty non-terminal, then the
     response will be one of no error in the return code and no RRSet
     in the answer section.

5. Security Considerations

     This document is refining the specifications to make it more
     likely that security can be added to DNS.  No functional
     additions are being made, just refining what is considered
     proper to allow the DNS, security of the DNS, and extending
     the DNS to be more predictable.

6. IANA Considerations

      None.

7. References

     Normative References

     [RFC20]   ASCII Format for Network Interchange, V.G. Cerf,
               Oct-16-1969

     [RFC1034] Domain Names - Concepts and Facilities,
               P.V. Mockapetris, Nov-01-1987

     [RFC1035] Domain Names - Implementation and Specification, P.V
               Mockapetris, Nov-01-1987

     [RFC1995] Incremental Zone Transfer in DNS, M. Ohta, August 1996

     [RFC2119] Key Words for Use in RFCs to Indicate Requirement
               Levels, S Bradner, March 1997

     [RFC2181] Clarifications to the DNS Specification, R. Elz and
               R. Bush, July 1997

     [RFC2308] Negative Caching of DNS Queries (DNS NCACHE),
               M. Andrews, March 1998

     [RFC2672] Non-Terminal DNS Name Redirection, M. Crawford,
               August 1999.

     [RFC2782] A DNS RR for specifying the location of services (DNS
               SRV), A. Gulbrandsen, et.al., February 2000

     [RFC4033] DNS Security Introduction and Requirements, R. Arends,
               et.al., March 2005

     [RFC4034] Resource Records for the DNS Security Extensions,
               R. Arends, et.al., March 2005

     [RFC4035] Protocol Modifications for the DNS Security Extensions,
               R. Arends, et.al., March 2005

     [RFC2672] Non-Terminal DNS Name Redirection, M. Crawford,
               August 1999

     Informative References

     [RFC2136] Dynamic Updates in the Domain Name System (DNS UPDATE),
               P. Vixie, Ed., S. Thomson, Y. Rekhter, J. Bound,
               April 1997

8. Editor

          Name:         Edward Lewis
          Affiliation:  NeuStar
          Address:      46000 Center Oak Plaza, Sterling, VA, 20166, US
          Phone:        +1-571-434-5468
          Email:        ed.lewis@neustar.biz

     Comments on this document can be sent to the editor or the mailing
     list for the DNSEXT WG, namedroppers@ops.ietf.org.

9. Others Contributing to the Document

     This document represents the work of a large working group.  The
     editor merely recorded the collective wisdom of the working group.

10. Trailing Boilerplate

     Copyright (C) The Internet Society (2005).

     This document is subject to the rights, licenses and restrictions
     contained in BCP 78, and except as set forth therein, the authors
     retain all their rights.

     This document and the information contained herein are provided
     on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION
     HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET
     SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL
     WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO
     ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT
     INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
     MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Intellectual Property

     The IETF takes no position regarding the validity or scope of
     any Intellectual Property Rights or other rights that might
     be claimed to pertain to the implementation or use of the
     technology described in this document or the extent to which
     any license under such rights might or might not be available;
     nor does it represent that it has made any independent effort
     to identify any such rights.  Information on the procedures
     with respect to rights in RFC documents can be found in BCP 78
     and BCP 79.

     Copies of IPR disclosures made to the IETF Secretariat and any
     assurances of licenses to be made available, or the result of an
     attempt made to obtain a general license or permission for the
     use of such proprietary rights by implementers or users of this
     specification can be obtained from the IETF on-line IPR
     repository at http://www.ietf.org/ipr.  The IETF invites any
     interested party to bring to its attention any copyrights,
     patents or patent applications, or other proprietary rights
     that may cover technology that may be required to implement
     this standard.  Please address the information to the IETF at
     ietf-ipr@ietf.org.

Acknowledgement

     Funding for the RFC Editor function is currently provided by the
     Internet Society.

Expiration

     This document expires on or about January 6, 2006.


Html markup produced by rfcmarkup 1.108, available from http://tools.ietf.org/tools/rfcmarkup/