Network Working Group                                 Alain Durand
INTERNET-DRAFT                              SUN Microsystems, inc.
June 14,
August 21, 2002                           Jun-ichiro itojun Hagino
Expires December February 2002                      IIJ Research Laboratory
                                                       Dave Thaler

        Well known site local unicast addresses for DNS resolver

                          Status of this Memo memo

   This document is an Internet-Draft and is in full conformance with
   all provisions of Section 10 of RFC2026.

   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-

   Internet Drafts are 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 a "work in progress".

   To view the list Internet-Draft Shadow Directories, see


   This documents specifies a method for nodes to find a DNS resolver
   with minimum configuration in the network and without running a
   discovery protocol on the nodes. This method is to be used in last
   resort, when no other information about the addresses of DNS
   resolvers is available.

Copyright Notice notice

   Copyright (C) The Internet Society (2002).  All Rights Reserved.

1. Introduction

   RFC 2462 [ADDRCONF] provides a way to autoconfigure nodes with one or
   more IPv6 address and default routes.

   However, for a node to be fully operational on a network, many other
   parameters are needed, such as the address of a DNS resolver, mail
   relays, web proxies, etc.  Except for name resolution, all the other
   services are usually described using names, not addresses, such as or  For obvious bootstrapping
   reasons, a node needs to be configured with the IP address (and not
   the name) of a DNS resolver.  As IPv6 addresses look much more
   complex than IPv4 ones, there is some incentive to make this
   configuration as automatic and simple as possible.

   Although it would be desirable to have all configuration parameters
   configured/discovered automatically, it is common practice in IPv4
   today to ask the user to do manual configuration for some of them by
   entering server names in a configuration form. So, a solution that
   will allow for automatic configuration of the DNS resolver is seen as
   an important step forward in the autoconfiguration story.

   The intended usage scenario for this proposal is a home or enterprise
   network where IPv6 nodes are plugged/unplugged with minimum
   management and use local resources available on the network to
   autoconfigure. This proposal is also usefull in cellular networks
   where all moble devices are included within the same site.

2. Pre-configuration vs discovery

   Some of the discussions in the past around DNS server discovery have
   been trying to characterize the solution space into stateless versus
   stateful or server-oriented versus severless.  It is not absolutely
   clear how much state if any needs to be kept to perform DNS server
   discovery, and, although the semantic differences between a router
   and a server are well understood from a conceptual perspective, the
   current implementations tend to blur the picture.  In another attempt
   to characterize different approaches, one can look at how much
   intelligence a client needs to have in order to use the service.

   One avenue is to ask the IPv6 node to participate in a discovery
   protocol, such as SLP or DHCP, learn the address of the server and
   send packets to this server. Another one is to pre-configure (hard-
   code) a local scope address on the IPv6 node and let it send packets
   directly to this address, with the underlying assumption that the
   routing system will forward them to the right place.  This document
   explores this later avenue of pre-configuration that does not require
   participation of the end node in the DNS resolver discovery

   The mechanism described here is to be used as a last resort, when no in the
   absence of any information about the addresses of DNS resolvers.  If
   other configuration information mechanisms are available, they should be tried

   Note to implementors:

   Implementing only the mechanism described in this memo may end up
   causing some interoperability problems when operating in networks
   where no DNS resolver is available. configured with the well known addresses.
   Thus, it is recommended to implement also other mechanisms for
   overriding this default, for example: manual configuration, L2
   mechanisms and/or DHCPv6.

3. Reserved Prefix prefix and addresses

   The basic idea of this proposal is to reserve a three well known IPv6
   site local prefix and three well known IPv6 addresses for DNS resolvers and then have the routing
   system forward the DNS request to those DNS
   resolvers. them.

   IPv6 nodes will be stub-resolvers implementing this proposal are pre-configured (hard coded)
   with those three IPv6 addresses as DNS resolver.

   Each local

   Local DNS resolvers should be configured with one of those three
   addresses to enable clients to switch from one to the other if one
   fails.  Host routes for each of those

   A solution to enable clients to reach the DNS resolvers should be injected is to inject
   host routes in the local routing system. Example Examples of methods for
   injecting host routes and a brief discussion of their fate sharing
   properties is are presented here:

      a) Manual injection of routes by a router on the same subnet.
      If the node running the DNS resolver goes down, the router may or
      may not be notified and keep announcing the route.

      b) Running a routing protocol on the same node running the DNS
      If the process running the DNS resolver dies, the routing protocol
      may or may not be notified and keep annoucing the route.

      c) Running a routing protocol within the same process running the
      DNS resolver.
      If the DNS resolver and the routing protocol run in separated
      threads, similar concerns as above are true.

      d) Having an "announcement" protocol that the DNS resolver could
      use to advertize the host route to the nearby router.  Details of
      such a protocols are out of scope of this document, but something
      similar to [MLD] is possible.

   An alternate solution is to configure a link with the well known
   prefix and position the three DNS resolvers on that link.  The
   advantage of this method is that host routes are not necessary , the
   well known prefix is advertized to the routing system by the routers
   on the link. However, in the event of a problem on the physical link,
   all resolvers will become unreachable.

   IANA considerations for this prefix are covered in Section 6.

4. Site local versus global scope considerations

   The rationales for having a site local prefix are:

      -a) Using a site local prefix will ensure that the traffic to the
      DNS resolver stays local to the site. This will prevent the DNS
      requests from accidentally leaking out of the site.  However, the
      local resolver can implement a policy to forward DNS resolution of
      non-local addresses to an external DNS resolver.

      -b) Reverse DNS resolution of site local addresses is only
      meaningful within the site. Thus, making sure that such queries
      are first sent to a DNS resolver located within the site perimeter
      increase their likelyhood of success.

   Note: there is currently some discussions about the usefulness of
   site local addresses in the IPv6 architecture. Depending on the
   outcome of this discussion, this section will need to be revisited.
   If a global prefix was chosen for this mechanism, concerns raised in
   a) could be addressed using a simple access list on the site exit
   routers and concerns raised in b) would disappear.

5. Examples of use

   This section presents example scenarios showing how the mechanism
   described in this memo can co-exist with other techniques, namely
   manual configuration and DHCPv6 discovery.

5.1 Simple case, general purpose DNS resolver

   This example shows the case of an enterprise or a cellular network
   that manages a full flavor general purpose DNS resolver and a large
   number of nodes running DNS stub resolvers.  The DNS resolver is
   performing (and caching) all the recursive queries on behalf of the
   stub resolvers.  Those stub resolvers are either manually configured
   with the IPv6 address of the resolver or with one (or several) of the
   well known site local unicast addresses defined in this memo.

            |                                         |
            |                  ---------------------  |
            |                  |manually configured|  |
            |                  |DNS stub resolver  |  |
            |                  ---------------------  |
            |  ----------           |                 |
            |  |DNS     |<-----------                 |
            |  |resolver|<-----------                 |
            |  ----------           |                 |
            |                  ---------------------  |
            |                  |DNS stub resolver  |  |
            |                  |configured with    |  |
            |                  |well known address |  |
            |                  ---------------------  |
            |                                         |

            (The DNS resolver is configured to listen both on
            its IPv6 address and on the well known address)

5.2 DNS forwarder

   A drawback of the choice of site local scope for the reserved
   addresses for DNS resolver is that, in the case of a home/small
   office network connected to an ISP, DNS traffic cannot be sent
   directly to the ISP DNS resolver without having the ISP and all its
   customers share the same definition of site.

   In this scenario, the home/small office network is connected to the
   ISP router (PE) via an edge router (CPE). Prefix delegation is
   performed out of band is is out of scope of this memo.

                                                    /            |
            --------           --------------      /             |
            |ISP PE|           |customer CPE|     /    Customer  |
            |      |===========|            |====<     site      |
            |      |           |            |     \              |
            --------           --------------      \             |
                                                    \            |

   The customer router CPE could be configured on its internal interface
   with one of the reserved site local addresses and listen for DNS
   queries. It would be configured to use one (or several) of the well
   known site local unicast addresses within the ISP's site to send its
   own queries to.  It would act as a DNS forwarder, forwarding queries
   received on its internal interface to the ISP's DNS resolver.

                                                  /            |
        ----------           --------------      /             |
        |ISP     |           |customer CPE|     /    Customer  |
        |DNS     |===========|         DNS|====<     site      |
        |resolver|    <------|---forwarder|-----\----          |
        ----------           --------------      \             |
                                                  \            |

       In this configuration, the CPE is acting as a multi-sited router.

5.3 DNS forwarder with DHCPv6 interactions

   In this variant scenario, DHCPv6 could is be used between the PE and CPE to
   do prefix delegation [DELEG] and DNS resolver discovery.

                                                    /            |
            --------           --------------      /             |
            |ISP   |           |customer CPE|     /    Customer  |
            |DHCPv6|===========|      DHCPv6|====<     site      |
            |server|    <------|------client|     \              |
            --------           --------------      \             |
                                                    \            |

   This example will show how DHCPv6 and well known site local unicast
   addresses can be used at the same time within a site cooperate to discover the
   address of enable the DNS forwarder. internal nodes to access DNS.

   The customer router CPE could be is configured on its internal interface with
   one of the reserved site local addresses and listen for DNS queries.
   It would act as a DNS forwarder, as in 5.2,  forwarding those queries
   to the DNS resolver pointed out by the ISP in the DHCPv6 exchange.

                                                  /            |
        ----------           --------------      /             |
        |ISP     |           |customer CPE|     /    Customer  |
        |DNS     |===========|         DNS|====<     site      |
        |resolver|    <------|---forwarder|-----\----          |
        ----------           --------------      \             |
                                                  \            |

   The same CPE router could also act as implement a local DHCPv6 server,
   advertising  either server and
   advertises itself as DNS forwarder.

                                                    /            |
            --------           --------------      /   Customer  |
            |ISP PE|           |customer CPE|     /    site      |
            |      |===========|DHCPv6      |====<               |
            |      |           |server------|-----\--->          |
            --------           --------------      \             |
                                                    \            |

   Within the site:

      a) DHCPv6 aware clients could use DHCPv6 to obtain the address of the
      DNS forwarder...

                                                  /            |
        ----------           --------------      /   Customer  |
        |ISP     |           |customer CPE|     /    site      |
        |DNS     |===========|         DNS|====<               |
        |resolver|    <------|---forwarder|-----\----DHCPv6    |
        ----------           --------------      \   client    |
                                                  \            |
          (The address of the DNS forwarder is aquired via DHCPv6.)

      b) other nodes may simply send their DNS request to the reserved site
      local addresses.

                                                  /            |
        ----------           --------------      /   customer  |
        |ISP     |           |customer CPE|     /    site      |
        |DNS     |===========|         DNS|====<               |
        |resolver|    <------|---forwarder|-----\----non DHCPv6|
        ----------           --------------      \   node      |
                                                  \            |
          (Internal nodes use the reserved site local unicast address.)

   A variant of this scenario is the CPE can decide to pass the global
   address of the ISP DNS resolver in the DHCPv6 exchange with the
   internal nodes.

6. IANA considerations

   The site local prefix fec0:0000:0000:ffff::/64 is to be reserved out
   of the site local fec0::/10 prefix.

   The unicast addresses fec0:000:0000:ffff::1, fec0:000:0000:ffff::2
   and fec0:000:0000:ffff::2 fec0:000:0000:ffff::3 are to be reserved for DNS resolver

   All other addresses within the fec0:0000:0000:ffff::/64 are reserved
   for future use and are expected to be assigned only with IESG

7.  Security Considerations

   Ensuring that queries reach a legitimate DNS server relies on the
   security of the IPv6 routing infrastructure.  The issues here are the
   same as those for protecting basic IPv6 connectivity.

   IPsec/IKE can be used as the well-known addresses are used as unicast

   The payload can be protected using standard DNS security techniques.
   If the client can preconfigure a well known private or public key
   then TSIG [TSIG] can be used with the same packets presented for the
   query.  If this is not the case, then TSIG keys will have to be
   negotiated using [TKEY].  After the client has the proper key then
   the query can be performed.

   The use of site local addresses instead of global addresses will
   ensure the DNS queries issued by host using this mechanism will not
   leak out of the site.

8.  References

        Thomson, S., and T. Narten, "IPv6 Stateless Address
        Autoconfiguration", RFC 2462, December 1998.

        Deering, S., Fenner, W., Haberman, B.,
        "Multicast Listener Discovery (MLD) for IPv6",
        RFC2710, October 1999.

        Vixie, P., Gudmundsson, O., Eastlake, D. and B. Wellington,
        "Secret Key Transaction Authentication for DNS (TSIG)",
        RFC2845, May 2000.

        D. Eastlake, "Secret Key Establishment for DNS (TKEY RR)",
        RFC2930, September 2000.

        Bound, J., Carney, M., Perkins, C., Lemon, T., Volz, B. and
        Droms, R. (ed.), "Dynamic host Configuration Protocol for IPv6
        (DHCPv6)", draft-ietf-dhc-dhcpv6-23 (work in progress),
        Februray 2002.

        Troan, O., Droms, R., "IPv6 Prefix Options for DHCPv6",
        draft-troan-dhcpv6-opt-prefix-delegation-00.txt (work in progress),
        February 2002.

9.  Authors' Addresses

   Alain Durand
   SUN microsystems, inc.
   901 San Antonio rd UMPK 17-202
   Palo Alto, CA 94303, USA.

   Jun-ichiro itojun HAGINO
   Research Laboratory, Internet Initiative Japan Inc.
   Takebashi Yasuda Bldg.,
   3-13 Kanda Nishiki-cho,
   Chiyoda-ku, Tokyo 101-0054, JAPAN

   Dave Thaler
   One Microsoft Way
   Redmond, CA 98052, USA

10.  Full Copyright Statement

Copyright (C) The Internet Society (2002).  All Rights Reserved.

   This document and translations of it may be copied and furnished to
   others, and derivative works that comment on or otherwise explain it
   or assist in its implementation may be prepared, copied, published
   and distributed, in whole or in part, without restriction of any
   kind, provided that the above copyright notice and this paragraph are
   included on all such copies and derivative works.  However, this
   document itself may not be modified in any way, such as by removing
   the copyright notice or references to the Internet Society or other
   Internet organizations, except as needed for the purpose of
   developing Internet standards in which case the procedures for
   copyrights defined in the Internet languages other than English.

   The limited permissions granted above are perpetual and will not be
   revoked by the Internet Society or its successors or assigns.

   This document and the information contained herein is provided on an