RFC 6804

Independent Submission                                        B. Manning
Request for Comments: 6804                                 November 2012
Category: Historic
ISSN: 2070-1721

               DISCOVER: Supporting Multicast DNS Queries


   This document describes the DISCOVER opcode, an experimental
   extension to the Domain Name System (DNS) to use multicast queries
   for resource discovery.  This opcode was tested in experiments run
   during 1995 and 1996 for the Topology Based Domain Search (TBDS)
   project.  This project is no longer active and there are no current
   plans to restart it.  TBDS was the first known use of multicast
   transport for DNS.  A client multicasts a DNS query using the
   DISCOVER opcode and processes the multiple responses that may result.

Status of This Memo

   This document is not an Internet Standards Track specification; it is
   published for the historical record.

   This document defines a Historic Document for the Internet community.
   This is a contribution to the RFC Series, independently of any other
   RFC stream.  The RFC Editor has chosen to publish this document at
   its discretion and makes no statement about its value for
   implementation or deployment.  Documents approved for publication by
   the RFC Editor are not a candidate for any level of Internet
   Standard; see Section 2 of RFC 5741.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at

Copyright Notice

   Copyright (c) 2012 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (http://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.

Manning                         Historic                        [Page 1]

RFC 6804                        DISCOVER                   November 2012

1.  Introduction

   The TBDS project developed extensions to existing network services to
   enable software for clients and servers of an application to become
   more resilient to changes in topology by dynamically sensing changes
   and switching between client/server and peer-peer methods for both
   end-system-to-server and server-to-server communications.

   The first existing network service to be investigated was the Domain
   Name Systems (DNS), which is used to map symbolic Internet names to
   numeric Internet addresses.  Based upon a hierarchical tree
   structure, the DNS relies upon uninterrupted connectivity of nodes to
   a special set of static, manually configured root servers.  To
   improve the robustness and availability of the DNS service, TBDS
   developed and defined enhancements that enable nodes to map names to
   numbers without the need for uninterrupted connectivity to the
   Internet root servers.  These techniques were automated, allowing
   transition between connected and unconnected operations to be done
   without direct human intervention.

   These enhancements to the DNS server code are based on the open
   source BIND to support reception and processing of multicast packets.

   Proof-of-concept modifications to BIND 8.1.2 were made to show that
   multicast awareness could be added to BIND.  An analysis was made of
   the existing DNS code deployment and the schedule of new feature
   deployment so that we could synchronize TBDS with a more appropriate
   code base.  Testing identified a race condition due to overloading
   the semantics of the DNS opcode that was used to communicate to

   This race condition was explored within the IETF regarding use of
   existing DNS opcodes.  Discussion within the team and with others in
   the IETF led to the idea that we needed a new opcode that would not
   overload the semantics of existing opcodes.  The original DNS design
   specification presumes that few clients exist that would share common
   DNS data.  To correct this problem, a new opcode was designed to
   disambiguate TBDS requests from normal nameserver requests.

   In the standard Domain Name System (DNS) [1] [2], queries are always
   unicast using the QUERY opcode.  The TBDS research project [5],
   funded under DARPA grant F30602-99-1-0523, explored the use of
   multicast DNS [1] [2] queries for resource discovery by autonomous,
   mobile nodes in disconnected networks.  The operations model is
   covered in the TBDS documentation.  Multicast queries may return
   multiple replies, while the standard DNS QUERY operation (see
   Sections 3.7, 4.3, and 5 of RFC 1034 [1]; and Section 4.1.1 of RFC
   1035 [2]) expects a single reply.  Instead of extending the QUERY

Manning                         Historic                        [Page 2]

RFC 6804                        DISCOVER                   November 2012

   opcode, the project developed and tested a new query operation,
   DISCOVER, that was designed to accommodate multiple responses from a
   multicast query.  The ability to accept multiple replies provides a
   basis for discrimination of man-in-the-middle attacks, which succeed
   by being the first to respond.  Use of DISCOVER requires the use of
   caching in the receiver, so the ephemeral nature of stub resolvers is

   This memo documents the processing rules for DISCOVER, for possible
   incorporation in a future revision of the DNS specification.

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   document are to be interpreted as described in BCP 14, RFC 2119 [3].

2.  DISCOVER Processing Rules

   A requester will send a DISCOVER query message to a multicast
   destination address, with some particular multicast scope.  The
   requester must be prepared to receive multiple replies from multiple
   responders, although we expect that there will be a single reply per

   DISCOVER responses (i.e., response messages from DISCOVER queries)
   have standard Answer, Authority, and Additional sections.  For
   example, the DISCOVER response is the same as the response to a QUERY
   operation.  Zero-content answers should not be sent, to avoid badly
   formed or unfulfilled requests.  Responses should be sent to the
   unicast address of the requester, and the source address should
   reflect the unicast address of the responder.  DISCOVER responses may
   echo the request's Question section or leave it blank, just as for

   DISCOVER works like QUERY, with the following exceptions:

      1. The Question section of a DISCOVER operation contains
         <QNAME=zonename,QTYPE=SOA> tuples, if the section is present.

         Within TBDS, this structure was augmented with:
         <QNAME=service,QTYPE=SRV>.  While this worked, it would be
         cleaner to ask the SRV question in a separate pass; any future
         work should take this into consideration.

      2. If QDCOUNT equals 0, then only servers willing to do recursion
         should answer; other servers must silently discard a DISCOVER
         request with QDCOUNT equals 0.

Manning                         Historic                        [Page 3]

RFC 6804                        DISCOVER                   November 2012

      3. If QDCOUNT is not equal to 0, then only servers that are
         authoritative for the zones named by some QNAME should answer.

      Hence, replies to DISCOVER queries will always be authoritative or
      else have RA (Recursion Available) set.

3.  Using DISCOVER Queries

3.1.  Performing Host Lookups

   To perform a hostname lookup using DISCOVER, a node could:

      o  Compute the zone name of the enclosing in-addr.arpa, ip6.int,
         or ip6.arpa domain.

      o  DISCOVER whether any in-scope server(s) are authoritative for
         this zone.

            If so, query these authoritative servers for local
            in-addr/ip6 names.

      o  If not, DISCOVER whether there are recursive servers available.

            If so, query these recursive servers for local in-addr/ip6

         The requester can determine from the replies whether there are
         any DNS servers that are authoritative (or support recursion)
         for the zone.

      o  Once the host's Fully Qualified Domain Name (FQDN) is known,
         repeat the process to discover the closest enclosing
         authoritative server for this local name.

      o  Cache all NS and A data learned in this process, respecting
         Times To Live (TTLs).

3.2.  Performing Service Lookups

   To lookup a service name using DISCOVER, the following steps may be

      o  Use DISCOVER as outlined in Section 3.1 to perform
         gethostbyaddr() and then gethostbyname() on one's own link-
         local address.  This gives a list of local authoritative

Manning                         Historic                        [Page 4]

RFC 6804                        DISCOVER                   November 2012

      o  Assume that the closest enclosing zone for which an
         authoritative server responds to an in-scope DISCOVER message
         is this host's "parent domain", and compute the SRV name as


         This is a change to the definition provided in RFC 1034 [1].  A
         wildcard label ("*") in the QNAME used in a DNS message with
         the opcode DISCOVER should be evaluated with special rules: the
         wildcard should match any label for which the DNS server data
         is authoritative.  For example 'x.*.example.com.' would match
         'x.y.example.com.' and 'x.yy.example.com.', provided that the
         server was authoritative for 'example.com.'

      o  Finally, send an SRV query for this SRV name to the discovered
         local authoritative servers to complete the getservbyname()

         This call returns a structure that can be populated by response
         values, as follows:

         s_name    The name of the service, "_service" without the
                   preceding underscore.

         s_aliases The names returned in the SRV Resource Records (RRs)
                   in replies to the query.

         s_port    The port number in the SRV RRs replies to the query.
                   If these port numbers do not match, one of the port
                   numbers is chosen, and only those names that
                   correspond are returned.

         s_proto   The transport protocol passed from the DNS process
                   using the "_transport" label, without the preceding

3.3.  Using DISCOVER for Disconnected Names

   DISCOVER allows discovery of a host (for example, a printer offering
   LPD services) whose DNS server answers authoritatively for a domain
   name that hasn't been delegated to it, but is defined within some
   local scope.  Since DISCOVER is explicitly defined to discover
   undelegated zones for tightly scoped queries, this behavior isn't a
   violation of DNS's coherency principles.  Note that a responder to
   DISCOVER might not be traditional DNS software, it could be special-
   purpose software.

Manning                         Historic                        [Page 5]

RFC 6804                        DISCOVER                   November 2012

   DISCOVER usage for disconnected networks with no authoritative
   servers can be achieved using the following conditions:

      o  Hosts run a "stub authoritative server" that acts as though its
         FQDN were a zone name.

      o  The computed SOA gives the host's FQDN as the MNAME, "." as the
         ANAME, seconds-since-1Jan2000 as the SERIAL, and low constants
         for EXPIRE and the other SOA timers.

      o  NS is used as the host's FQDN.

      o  The glue is computed as the host's link-local address, or hosts
         may run a "DNS stub server" that acts as though its FQDN were a
         zone name.

   The rules governing the behavior of this server consist of a single
   change to the method of use, and no change whatsoever to the current
   format of DNS packets.  Specifically, this extension allows UDP DNS
   queries, as documented in RFC 1035, Sections 4.1.1, 4.1.2, and 4.2.1,
   to be addressed to port 53 of statically assigned relative offset -4
   within the range of multicast addresses defined as "administratively
   scoped" by Section 9 of RFC 2365 [6].  Within the full /8 of
   administratively scoped addresses, this corresponds to the
   destination address  Until the Multicast-Scope Zone
   Announcement Protocol (MZAP) or a similar protocol is implemented to
   allow hosts to discover the extent of the local multicast scopes that
   enclose them, it is anticipated that implementations will simply
   utilize the destination address  Queries sent via
   multicast MUST NOT request recursion.

   In order to receive multicasted queries, DNS server implementations
   MUST listen on the -4 offset to their local scope (as above, in the
   absence of a method of determining the scope, this will be assumed to
   be relative to the full /8 allocated for administratively scoped
   multicast use, or and respond via ordinary unicast
   UDP to ONLY those queries for which they have a positive answer that
   originated within a locally-configured zone file.  That is, a server
   MUST NOT answer a multicasted query with cached information that it
   received from another server, nor may it request further resolution
   from other servers on behalf of a multicasted query.  A multicast-
   capable server may, however, utilize multicast queries to perform
   further resolution on behalf of queries received via ordinary
   unicast.  This is referred to as "proxy" operation.  Multicast-
   enabled DNS servers MUST answer multicasted queries non-
   authoritatively.  That is, when responding to a query that was

Manning                         Historic                        [Page 6]

RFC 6804                        DISCOVER                   November 2012

   received via multicast, they MUST NOT include an NS record that
   contains data that resolves back to their own IP address and MUST NOT
   set the AA bit.

   Resolvers MUST anticipate receiving no replies to some multicasted
   queries, in the event that no multicast-enabled DNS server
   implementations are active within the local scope, or in the event
   that no positive responses exist to the transmitted query.  That is,
   a query for the MX record for host.domain.com would go unanswered if
   no local server was able to resolve that request, if no MX record
   exists for host.domain.com, or if no local servers were capable of
   receiving multicast queries.  The resolver that initiated the query
   MUST treat such non-response as a non-cacheable negative response.
   Since this multicast transmission does not provide reliable delivery,
   resolvers MAY repeat the transmission of a query in order to assure
   themselves that is has been received by any hosts capable of
   answering; however, any resolvers that repeat a query MUST increase
   the interval by a factor of two between each repetition.  It is more
   likely, however, that any repeated queries will be performed under
   the explicit direction of the application driving the query, rather
   than autonomously by the resolver implementation.

   It will often be the case that multicast queries will result in
   responses from multiple servers.  In the event that the multicast
   query was generated via a current API such as gethostbyname, or as
   the result of a proxy operation, the first response received must be
   passed to the requesting application or host, and all subsequently
   received responses must be discarded.  Future multicast-aware APIs
   that use DISCOVER should anticipate receiving multiple independent RR
   sets in response to queries and using external heuristics for
   selecting the most appropriate RR set.

   Such servers should answer DISCOVER packets for its zone, and will be
   found by the iterative "discover closest enclosing authority server"
   by DISCOVER clients, in either the gethostbyname() or SRV cases
   described above.  Note that stub servers answer only with zone names
   that exactly match QNAME's, not with zone names that are owned by

4.  IANA Considerations

   At such time as this idea might be considered for a future addition
   to the DNS protocol, IANA would need to assign a value for the

Manning                         Historic                        [Page 7]

RFC 6804                        DISCOVER                   November 2012

5.  Security Considerations

   The following paragraph on security considerations was written very
   early in the use and exploration of IP multicast and, as such,
   represents a fairly naive view on the type and scope of exploits that
   are enabled through the use of IP multicast.  A more up-to-date
   understanding of multicast security considerations may be found in
   RFC 5294 [4].

   No new security considerations are known to be introduced with a new
   DNS query operation.  However, using multicast for service discovery
   has the potential for denial of service from flooding attacks.  How
   to scope multicast is not part of the DISCOVER processing rules.  It
   may also be possible to enable deliberate misconfiguration of clients
   simply by running a malicious DNS server that falsely claims to be
   authoritative for delegations.  One possible way to mitigate this
   threat is to use credentials, such as CERT resource records within an
   RR set.  The TBDS project took this approach.  TBDS did not directly
   utilize DNS Security (DNSSEC), so possible interactions with DNSSEC-
   aware/-capable servers are unknown.

6.  Acknowledgments

   This material was generated in discussions on the mdns mailing list
   hosted by Zocalo in March 2000 and updated by discussions in
   September/October 2003 on a closed mailing list.  David Lawrence,
   Scott Rose, Stuart Cheshire, Bill Woodcock, and Erik Guttman were
   active contributors.  Suzanne Woolf was part of the original
   implementation team and an invaluable sanity checker.  Funding for
   the RFC Editor function is currently provided by the Internet

7.  References

7.1.  Normative References

        13, RFC 1034, November 1987.

        SPECIFICATION", STD 13, RFC 1035, November 1987.

   [3]  Bradner, S., "Key words for use in RFCs to Indicate Requirement
        Levels", BCP 14, RFC 2119, March 1997.

   [4]  Savola, P. and J. Lingard, "Host Threats to Protocol Independent
        Multicast (PIM)", RFC 5294, August 2008.

Manning                         Historic                        [Page 8]

RFC 6804                        DISCOVER                   November 2012

7.2.  Informative References

   [5]  Manning, B., "Topology Based Domain Search (TBDS)", Final
        Report, June 2002,

   [6]  Meyer, D., "Administratively Scoped IP Multicast", BCP 23, RFC
        2365, July 1998.

Authors' Addresses

   Bill Manning
   PO 12317
   Marina del Rey, CA. 90295
   United States

   EMail: bmanning@sfc.keio.ac.jp

Manning                         Historic                        [Page 9]