Network Working Group                                         R. Austein
draft-ietf-dnsind-sigalgopt-00.txt                         On Sabbatical
                                                                P. Vixie
                                            Internet Software Consortium
                                                            October 1999


                             DNS SIGALGOPT


Status of this document

   This document is an Internet-Draft and is in full conformance with
   all provisions of Section 10 of RFC 2026.

   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>

   Distribution of this document is unlimited. Please send comments to
   the namedroppers@internic.net mailing list.

Abstract

   This document describes a mechanism for conserving packet space in a
   DNS response message in the presence of multiple DNSSEC signature
   algorithms.

Motivation and Scope

   DNSSEC [DNSSEC] specifies a general framework for attaching
   cryptographic signatures to DNS resource records.  The framework
   includes provisions for multiple signature protocols, possibly even
   on a per-name basis.  While this open-ended framework is good and
   useful, it poses a problem when multiple signature protocols are in
   use and DNS message sizes are limited by the underlying UDP transport
   packet size.  EDNS0 [EDNS0] provides a way to specify a larger



Austein & Vixie           Expires 18 April 2000                 [Page 1]


draft-ietf-dnsind-sigalgopt-00.txt                          October 1999


   payload size, but this still does not entirely solve the problem for
   large RRsets.  Worse, in cases where multiple signature algorithms
   generate a response packet so large that it must be truncated, the
   signatures that fit into the truncated response will be useless if
   the resolver doesn't know how to verify signatures generated with
   that algorithm.

   This note proposes a way for a resolver to indicate which signature
   algorithms it understands to a name server in the form of an ordered
   list.  When this mechanism is in use, the name server can conserve
   packet space by (a) not sending signatures with algorithms that the
   resolver will not understand, and (b) not sending multiple signatures
   for the same resource records.

Mechanism

   [DNSSEC] SIG RRs include a one-octet code indicating the algorithm
   associated with a particular signature.  The SIGALGOPT option defined
   below allows the resolver to specify an ordered list of signature
   algorithms using the same one-octet codes that DNSSEC uses.

   SIGALGOPT is encoded n the variable RDATA part of the OPT pseudo-RR
   in the DNS request (see [EDNS0]).

   The OPTION-CODE for SIGALGOPT is [TBD].

   The OPTION-DATA for SIGALGOPT is an ordered list of the one-octet
   codes used by DNSSEC.

   If the SIGALGOPT option in a query specifies multiple signature
   algorithms and signatures using more than one of those algorithms are
   available in the zone, the server must respond with the signatures
   corresponding to the first algorithm on the SIGALGOPT list that
   matches, omitting any signatures corresponding to the remaining
   algorithms.

   We have deliberately not provided a mechanism to return all the
   matching signatures, because the purpose of the SIGALGOPT mechanism
   is to minimize packet size.  If the resolver wants to see all
   available signatures, it should just leave off the SIGALGOPT option
   entirely.

Security Considerations

   Good question.  What horrible things could a bad guy do by
   creating/altering/deleting SIGALGOPT?  Are any of the possible
   attacks more interesting than denial of service?




Austein & Vixie           Expires 18 April 2000                 [Page 2]


draft-ietf-dnsind-sigalgopt-00.txt                          October 1999


IANA Considerations

   SIGALGOPT will need an option code.

   The signature algorithm codes themselves are borrowed from DNSSEC and
   do not create any new issues for IANA.

References

   [DNSSEC]  Eastlake, D., "Domain Name System Security Extensions", RFC
        2535, March 1999.

   [DNS-CONCEPTS]  Mockapetris, P., "Domain names - concepts and
        facilities", RFC 1034, November 1987.

   [DNS-IMPLEMENTATION]  Mockapetris, P., "Domain names - implementation
        and specification", RFC 1035, November 1987.

   [EDNS0]  Vixie, P., "Extension Mechanisms for DNS (EDNS0)", RFC 2671,
        August 1999.

Author's addresses:

      Rob Austein
      On Sabbatical
      sra@hactrn.net

      Paul Vixie
      Internet Software Consortium
      950 Charter Street
      Redwood City, CA 94063
      +1 650 779 7001
      vixie@isc.org


















Austein & Vixie           Expires 18 April 2000                 [Page 3]