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

Versions: 00 01 02 03

INTERNET DRAFT                                                   M. Ohta
draft-ietf-dnsop-ohta-shared-root-server-02.txt
                                           Tokyo Institute of Technology
                                                           November 2002

         Root Name Servers with Inter Domain Anycast Addresses

Status of this Memo

   This document is an Internet-Draft and is subject to 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-
   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/1id-abstracts.html

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

Abstract

   This memo describes an operational guideline for millions of name
   servers to share an interdomain anycast address.

   It enables people operate as many root name servers as they want and
   still make them traceable.

1. Motivation

   DNS root servers are the essential component of the Internet that all
   the ISPs in the world want to run several root servers.

   To satisfy them, we need to have thousands or millions of root
   servers.

   However, because of the restriction on DNS message size over UDP, the
   number of unicast IP addresses of root servers is severely limited.

   Thus, it is necessary to increase the number of root servers by



M. Ohta                  Expires on May 1, 2003                 [Page 1]


INTERNET DRAFT         Anycast Root Name Servers           November 2002


   assigning an IP address to a lot of root servers.

   Even if DNS were designed to allow a lot of root servers, it is
   difficult for DNS clients to choose the best (with regard to
   availability, credibility of zone content, delay, domain policy etc.)
   root servers among so many root servers. It is not practical to ping
   millions of servers to find which has the smallest RTT.

   This memo proposes a mechanism of policy based selection of a root
   server sharing an IP address (anycast IP address) with other root
   servers and discusses operational issues related to it.

   Because the selection is policy based, domain administrators  have
   some control over the selection of the best root server among root
   servers sharing an IP address.

   Note that operations similar to that described in this memo are
   possible today locally without global coordination by any operator
   who may be irritated by the lack of his control on (sufficiently
   many) root servers, which may be a source of some operational
   problems. This memo is an attempt to document the way to solve the
   problem in a least harmful manner.

   Similar operation described in this memo may be applicable to gTLD or
   other global servers but it is outside the scope of this memo.

2. Suggested Operation

   As is demonstrated by proliferated private use addresses, it is easy
   to set up routers to let unicast addresses have local scopes. It is
   also easy to let the unicast addresses have nested local scopes. The
   important difference between the addresses for private use and root
   servers is in their semantics that the root servers sharing an
   address also share the globally unique semantics of the address. The
   root servers may share a globally unique DNS host name, too.

   A possible problem of such addresses is that the shared addresses can
   not be used for global communication.

   So, the root name servers with the anycast addresses must have
   additional globally unique unicast address (or addresses), which may
   be used for global communication such as zone transfer.

   The other possible problem of such addresses is that the shared
   addresses are not managed by a single entity that the mapping from
   the shared address of a root server to some operational entity is
   impossible.




M. Ohta                  Expires on May 1, 2003                 [Page 2]


INTERNET DRAFT         Anycast Root Name Servers           November 2002


   However, if a router adjacent to (or near) the root server has a
   globally unique address, it is possible to map from the global
   address to an operational entity, which is expected to be operating
   the root server.  That is, tools like "traceroute" work to uniquely
   identify the operational entity of the root servers sharing a anycast
   address.

   To be compatible with the current practice that a single address
   belong to a single AS, each anycast address is assigned its own AS
   number. There will be multiple ASes of the AS number containing the
   same address ranges.

   ASes, still, can be identified by adjacent ASes.  For example,
   network operators may choose their favorite root server based on the
   AS numbers of the next hop ASes with, for example, AS path and BGP
   policy.

   It is required that operators of an AS adjacent to the root servers'
   AS be fully responsible to the operation of the root servers.  If a
   root server's AS is adjacent to multiple ASes, operators of all the
   ASes must be fully responsible to the operation of the root server.
   Thus, if there is a routing problem related to a root server,
   operators of the next hop AS(es) should be contacted.

   To avoid complex routing tricks, globally unique unicast address(es)
   of the root name servers must be taken from the AS(es) adjacent to
   the root server's AS. Then, in a likely simple case that the root
   server's AS consists of a single host, which acts as a name server
   and a BGP router, the host should peer with adjacent AS(es) through
   an interface(s) address(es) of which belongs to the adjacent AS(es).
   If the root server's AS has more complex structure, special IGP
   arrangement of globally unique unicast address(es) is necessary in
   the AS and at the border router(s) of the adjacent AS(es). The border
   router(s) must accept IGP information advertised from the root
   server's AS.

3. Assignment

   When a server with an anycast address fails but a route to it is
   still available, there is no way for clients use other servers with
   the same anycast address, anycast does not improve availability of
   servers so much.

   So, even with anycast addresses, there should be multiple root
   servers.

   However, as anycast solves the problem of load concentration, we
   don't need so many anycast IP addresses,



M. Ohta                  Expires on May 1, 2003                 [Page 3]


INTERNET DRAFT         Anycast Root Name Servers           November 2002


   We should have at least 3 and at most 7 anycast addresses for root
   servers.

4. Ongoing Experiment

   An experiment is ongoing with a block of IP addresses 192.83.230/24
   in AS 4128 and DNS servers sharing an IP address of 192.83.230.1.

   Domains served by the DNS servers are not root domain but are: real-
   internet.org. and psg.com.

   If one is interested in joining (or just watching) the test, send a
   mail containing a single line of:

      subscribe aroot

   to

      majordomo@ops.ietf.org

   the mailing list is located at:

      aroot@ops.ietf.org

   Archive of the list is available at
   ftp://ops.ietf.org:/pub/lists/aroot.

   If there is something wrong with route information from AS 4128, the
   author of the memo or the mailing list for the test may be contacted.

   However, for direct contact to the source of the problem, the contact
   person of an AS next to AS 4128 in the AS path of problematic route
   information should be contacted.

5. Security Considerations

   This memo describes just an operational guideline with no protocol
   change. As such, the guideline does not introduce any security issues
   of the protocol level.

   As the route forgery to the root servers can be implemented today
   without this memo by anyone including local intruders, the guideline
   does not introduce any security issues of the operational level,
   either.

   A guideline to track down and verify a route or an AS path to a valid
   or a forged root server is described in section 2.




M. Ohta                  Expires on May 1, 2003                 [Page 4]


INTERNET DRAFT         Anycast Root Name Servers           November 2002


   Furthermore, if an ISP or a site operate its own anycast root server,
   hosts of the ISP or the site using the root server is protected from
   external forged route.

   In addition, if a lot of local root servers share an anycast address,
   it reduce the effect of distributed denial of service attack on the
   anycast address.

6. Author's Address

   Masataka Ohta
   Graduate School of Information Science and Engineering
   Tokyo Institute of Technology
   2-12-1, O-okayama, Meguro-ku
   Tokyo 152-8552, JAPAN

   Phone: +81-3-5734-3299
   Fax: +81-3-5734-3299
   EMail: mohta@necom830.hpcl.titech.ac.jp
































M. Ohta                  Expires on May 1, 2003                 [Page 5]


Html markup produced by rfcmarkup 1.129d, available from https://tools.ietf.org/tools/rfcmarkup/