Extension Mechanisms for DNS (EDNS0)

 

Network Working Group                                            P. Vixie
Request for Comments: 2671                                            ISC
Category: Standards Track                                     August 1999

                  Extension Mechanisms for DNS (EDNS0)

Status of this Memo

   This document specifies an Internet standards track protocol for the
   Internet community, and requests discussion and suggestions for
   improvements.  Please refer to the current edition of the "Internet
   Official Protocol Standards" (STD 1) for the standardization state
   and status of this protocol.  Distribution of this memo is unlimited.

Copyright Notice

   Copyright (C) The Internet Society (1999).  All Rights Reserved.

Abstract

   The Domain Name System's wire protocol includes a number of fixed
   fields whose range has been or soon will be exhausted and does not
   allow clients to advertise their capabilities to servers.  This
   document describes backward compatible mechanisms for allowing the
   protocol to grow.

1 - Rationale and Scope

1.1. DNS (see [ RFC1035]) specifies a Message Format and within such
     messages there are standard formats for encoding options, errors,
     and name compression.  The maximum allowable size of a DNS Message
     is fixed.  Many of DNS's protocol limits are too small for uses
     which are or which are desired to become common.  There is no way
     for implementations to advertise their capabilities.

1.2. Existing clients will not know how to interpret the protocol
     extensions detailed here.  In practice, these clients will be
     upgraded when they have need of a new feature, and only new
     features will make use of the extensions.  We must however take
     account of client behaviour in the face of extra fields, and design
     a fallback scheme for interoperability with these clients.

2 - Affected Protocol Elements

2.1. The DNS Message Header's (see [ RFC1035 4.1.1]) second full 16-bit
     word is divided into a 4-bit OPCODE, a 4-bit RCODE, and a number of
     1-bit flags.  The original reserved Z bits have been allocated to
     various purposes, and most of the RCODE values are now in use.
     More flags and more possible RCODEs are needed.

2.2. The first two bits of a wire format domain label are used to denote
     the type of the label.  [ RFC1035 4.1.4] allocates two of the four
     possible types and reserves the other two.  Proposals for use of
     the remaining types far outnumber those available.  More label
     types are needed.

2.3. DNS Messages are limited to 512 octets in size when sent over UDP.
     While the minimum maximum reassembly buffer size still allows a
     limit of 512 octets of UDP payload, most of the hosts now connected
     to the Internet are able to reassemble larger datagrams.  Some
     mechanism must be created to allow requestors to advertise larger
     buffer sizes to responders.

3 - Extended Label Types

3.1. The "0 1" label type will now indicate an extended label type,
     whose value is encoded in the lower six bits of the first octet of
     a label.  All subsequently developed label types should be encoded
     using an extended label type.

3.2. The "1 1 1 1 1 1" extended label type will be reserved for future
     expansion of the extended label type code space.

4 - OPT pseudo-RR

4.1. One OPT pseudo-RR can be added to the additional data section of
     either a request or a response.  An OPT is called a pseudo-RR
     because it pertains to a particular transport level message and not
     to any actual DNS data.  OPT RRs shall never be cached, forwarded,
     or stored in or loaded from master files.  The quantity of OPT
     pseudo-RRs per message shall be either zero or one, but not
     greater.

4.2. An OPT RR has a fixed part and a variable set of options expressed
     as {attribute, value} pairs.  The fixed part holds some DNS meta
     data and also a small collection of new protocol elements which we
     expect to be so popular that it would be a waste of wire space to
     encode them as {attribute, value} pairs.

4.3. The fixed part of an OPT RR is structured as follows:

     Field Name   Field Type     Description
     ------------------------------------------------------
     NAME         domain name    empty (root domain)
     TYPE         u_int16_t      OPT
     CLASS        u_int16_t      sender's UDP payload size
     TTL          u_int32_t      extended RCODE and flags
     RDLEN        u_int16_t      describes RDATA
     RDATA        octet stream   {attribute,value} pairs

4.4. The variable part of an OPT RR is encoded in its RDATA and is
     structured as zero or more of the following:

                +0 (MSB)                            +1 (LSB)
     +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
  0: |                          OPTION-CODE                          |
     +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
  2: |                         OPTION-LENGTH                         |
     +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
  4: |                                                               |
     /                          OPTION-DATA                          /
     /                                                               /
     +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+

   OPTION-CODE    (Assigned by IANA.)

   OPTION-LENGTH  Size (in octets) of OPTION-DATA.

   OPTION-DATA    Varies per OPTION-CODE.

4.5. The sender's UDP payload size (which OPT stores in the RR CLASS
     field) is the number of octets of the largest UDP payload that can
     be reassembled and delivered in the sender's network stack.  Note
     that path MTU, with or without fragmentation, may be smaller than
     this.

4.5.1. Note that a 512-octet UDP payload requires a 576-octet IP
       reassembly buffer.  Choosing 1280 on an Ethernet connected
       requestor would be reasonable.  The consequence of choosing too
       large a value may be an ICMP message from an intermediate
       gateway, or even a silent drop of the response message.

4.5.2. Both requestors and responders are advised to take account of the
       path's discovered MTU (if already known) when considering message
       sizes.

4.5.3. The requestor's maximum payload size can change over time, and
       should therefore not be cached for use beyond the transaction in
       which it is advertised.

4.5.4. The responder's maximum payload size can change over time, but
       can be reasonably expected to remain constant between two
       sequential transactions; for example, a meaningless QUERY to
       discover a responder's maximum UDP payload size, followed
       immediately by an UPDATE which takes advantage of this size.
       (This is considered preferrable to the outright use of TCP for
       oversized requests, if there is any reason to suspect that the
       responder implements EDNS, and if a request will not fit in the
       default 512 payload size limit.)

4.5.5. Due to transaction overhead, it is unwise to advertise an
       architectural limit as a maximum UDP payload size.  Just because
       your stack can reassemble 64KB datagrams, don't assume that you
       want to spend more than about 4KB of state memory per ongoing
       transaction.

4.6. The extended RCODE and flags (which OPT stores in the RR TTL field)
     are structured as follows:

                 +0 (MSB)                            +1 (LSB)
      +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
   0: |         EXTENDED-RCODE        |            VERSION            |
      +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
   2: |                               Z                               |
      +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+

   EXTENDED-RCODE  Forms upper 8 bits of extended 12-bit RCODE.  Note
                   that EXTENDED-RCODE value "0" indicates that an
                   unextended RCODE is in use (values "0" through "15").

   VERSION         Indicates the implementation level of whoever sets
                   it.  Full conformance with this specification is
                   indicated by version "0."  Requestors are encouraged
                   to set this to the lowest implemented level capable
                   of expressing a transaction, to minimize the
                   responder and network load of discovering the
                   greatest common implementation level between
                   requestor and responder.  A requestor's version
                   numbering strategy should ideally be a run time
                   configuration option.

                   If a responder does not implement the VERSION level
                   of the request, then it answers with RCODE=BADVERS.
                   All responses will be limited in format to the

                   VERSION level of the request, but the VERSION of each
                   response will be the highest implementation level of
                   the responder.  In this way a requestor will learn
                   the implementation level of a responder as a side
                   effect of every response, including error responses,
                   including RCODE=BADVERS.

   Z               Set to zero by senders and ignored by receivers,
                   unless modified in a subsequent specification.

5 - Transport Considerations

5.1. The presence of an OPT pseudo-RR in a request should be taken as an
     indication that the requestor fully implements the given version of
     EDNS, and can correctly understand any response that conforms to
     that feature's specification.

5.2. Lack of use of these features in a request must be taken as an
     indication that the requestor does not implement any part of this
     specification and that the responder may make no use of any
     protocol extension described here in its response.

5.3. Responders who do not understand these protocol extensions are
     expected to send a response with RCODE NOTIMPL, FORMERR, or
     SERVFAIL.  Therefore use of extensions should be "probed" such that
     a responder who isn't known to support them be allowed a retry with
     no extensions if it responds with such an RCODE.  If a responder's
     capability level is cached by a requestor, a new probe should be
     sent periodically to test for changes to responder capability.

6 - Security Considerations

     Requestor-side specification of the maximum buffer size may open a
     new DNS denial of service attack if responders can be made to send
     messages which are too large for intermediate gateways to forward,
     thus leading to potential ICMP storms between gateways and
     responders.

7 - IANA Considerations

     The IANA has assigned RR type code 41 for OPT.

     It is the recommendation of this document and its working group
     that IANA create a registry for EDNS Extended Label Types, for EDNS
     Option Codes, and for EDNS Version Numbers.

     This document assigns label type 0b01xxxxxx as "EDNS Extended Label
     Type."  We request that IANA record this assignment.

     This document assigns extended label type 0bxx111111 as "Reserved
     for future extended label types."  We request that IANA record this
     assignment.

     This document assigns option code 65535 to "Reserved for future
     expansion."

     This document expands the RCODE space from 4 bits to 12 bits.  This
     will allow IANA to assign more than the 16 distinct RCODE values
     allowed in [ RFC1035].

     This document assigns EDNS Extended RCODE "16" to "BADVERS".

     IESG approval should be required to create new entries in the EDNS
     Extended Label Type or EDNS Version Number registries, while any
     published RFC (including Informational, Experimental, or BCP)
     should be grounds for allocation of an EDNS Option Code.

8 - Acknowledgements

     Paul Mockapetris, Mark Andrews, Robert Elz, Don Lewis, Bob Halley,
     Donald Eastlake, Rob Austein, Matt Crawford, Randy Bush, and Thomas
     Narten were each instrumental in creating and refining this
     specification.

9 - References

    [ RFC1035]  Mockapetris, P., "Domain Names - Implementation and
               Specification", STD 13, RFC 1035, November 1987.

10 - Author's Address

   Paul Vixie
   Internet Software Consortium
   950 Charter Street
   Redwood City, CA 94063

   Phone: +1 650 779 7001
   EMail: vixie@isc.org

11 - Full Copyright Statement

   Copyright (C) The Internet Society (1999).  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 Standards process must be
   followed, or as required to translate it into 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
   "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
   TASK FORCE DISCLAIMS 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.

Acknowledgement

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