RFC 4704 PROPOSED STANDARD DHCP · DNS · IPV6

DHCPv6 FQDN option: The Dynamic Host Configuration Protocol for IPv6 (DHCPv6) Client Fully Qualified Domain Name (FQDN) Option

How a DHCPv6 client tells the server its domain name, and who then updates the DNS records.

On rfc-editor.org ↗
PUBLISHED
October 2006
AUTHORS
B. Volz
STREAM
IETF
DHCPV6 OPTION CODE
39
MINIMUM OPTION LENGTH
1 octet
OPTION-LEN FORMULA
1 + length of domain name
SUGGESTED RECORD LIFETIME CEILING
One third of the lease lifetime
SUGGESTED RECORD LIFETIME FLOOR
10 minutes
MESSAGES THAT MAY CARRY THE OPTION
SOLICIT, REQUEST, RENEW, REBIND
DOI
10.17487/RFC4704

RFC text

Contents
Original text · RFC Editor View on rfc-editor.org ↗
Network Working Group                                            B. Volz
Request for Comments: 4704                           Cisco Systems, Inc.
Category: Standards Track                                   October 2006


   The Dynamic Host Configuration Protocol for IPv6 (DHCPv6) Client
               Fully Qualified Domain Name (FQDN) Option

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 (2006).

Abstract

   This document specifies a new Dynamic Host Configuration Protocol for
   IPv6 (DHCPv6) option that can be used to exchange information about a
   DHCPv6 client's Fully Qualified Domain Name (FQDN) and about
   responsibility for updating DNS resource records (RRs) related to the
   client's address assignments.

Table of Contents

   1. Introduction ....................................................3
   2. Terminology .....................................................3
   3. Models of Operation .............................................3
   4. The DHCPv6 Client FQDN Option ...................................4
      4.1. The Flags Field ............................................5
      4.2. The Domain Name Field ......................................6
   5. DHCPv6 Client Behavior ..........................................7
      5.1. Client Desires to Update AAAA RRs ..........................7
      5.2. Client Desires Server to Do DNS Updates ....................7
      5.3. Client Desires No Server DNS Updates .......................7
      5.4. Domain Name and DNS Update Issues ..........................8
   6. DHCPv6 Server Behavior ..........................................9
      6.1. When to Perform DNS Updates ................................9
   7. DNS RR TTLs ....................................................10
   8. DNS Update Conflicts ...........................................11
   9. IANA Considerations ............................................11
   10. Security Considerations .......................................12
   11. Acknowledgements ..............................................12
   12. References ....................................................13
      12.1. Normative References .....................................13
      12.2. Informative References ...................................13

1.  Introduction

   DNS ([2], [3]) maintains (among other things) the information about
   mapping between hosts' Fully Qualified Domain Names (FQDNs) [10] and
   IPv6 addresses assigned to the hosts.  The information is maintained
   in two types of Resource Records (RRs): AAAA and PTR [12].  The DNS
   update specification [4] describes a mechanism that enables DNS
   information to be updated over a network.

   The Dynamic Host Configuration Protocol for IPv6 (DHCPv6) [5]
   provides a mechanism by which a host (a DHCPv6 client) can acquire
   certain configuration information, along with its stateful IPv6
   address(es).  This document specifies a new DHCPv6 option, the Client
   FQDN option, which can be used by DHCPv6 clients and servers to
   exchange information about the client's fully qualified domain name
   and about who has the responsibility for updating the DNS with the
   associated AAAA and PTR RRs.

2.  Terminology

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in [1].

   Familiarity with the DNS Update protocol [4] and with DHCPv6 and its
   terminology, as defined in [5], is assumed.

3.  Models of Operation

   When a DHCPv6 client acquires an address, a site's administrator may
   desire that the AAAA RR for the client's FQDN and the PTR RR for the
   acquired address be updated.  Therefore, two separate DNS update
   transactions may occur.  Acquiring an address via DHCPv6 involves two
   entities: a DHCPv6 client and a DHCPv6 server.  In principle, each of
   these entities could perform none, one, or both of the DNS update
   transactions.  However, in practice, not all permutations make sense.
   The DHCPv6 Client FQDN option is primarily intended to operate in the
   following two cases:

   1.  DHCPv6 client updates the AAAA RR; DHCPv6 server updates the PTR
       RR.

   2.  DHCPv6 server updates both the AAAA and the PTR RRs.

   The only difference between these two cases is whether the FQDN-to-
   IPv6-address mapping is updated by a DHCPv6 client or by a DHCPv6
   server.  The IPv6-address-to-FQDN mapping is updated by a DHCPv6
   server in both cases.

   The reason these two are important, while others are unlikely, has to
   do with authority over the respective DNS domain names.  A DHCPv6
   client may be given authority over mapping its own AAAA RRs, or that
   authority may be restricted to a server to prevent the client from
   listing arbitrary addresses or associating its addresses with
   arbitrary domain names.  In all cases, the only reasonable place for
   the authority over the PTR RRs associated with the address is in the
   DHCPv6 server that allocates the address.

   Note: A third case is supported in which the client requests that the
   server perform no updates.  However, this case is presumed to be rare
   because of the authority issues.

   In any case, whether a site permits all, some, or no DHCPv6 servers
   and clients to perform DNS updates into the zones that it controls is
   entirely a matter of local administrative policy.  This document does
   not require any specific administrative policy and does not propose
   one.  The range of possible policies is very broad, from sites where
   only the DHCPv6 servers have been given credentials that the DNS
   servers will accept, to sites where each individual DHCPv6 client has
   been configured with credentials that allow the client to modify its
   own domain name.  Compliant implementations MAY support some or all
   of these possibilities.  Furthermore, this specification applies only
   to DHCPv6 client and server processes: it does not apply to other
   processes that initiate DNS updates.

   This document describes a new DHCPv6 option that a client can use to
   convey all or part of its domain name to a DHCPv6 server.  Site-
   specific policy determines whether or not DHCPv6 servers use the
   names that clients offer, and what DHCPv6 servers do in cases where
   clients do not supply domain names.

4.  The DHCPv6 Client FQDN Option

   To update the IPv6-address-to-FQDN mapping, a DHCPv6 server needs to
   know the FQDN of the client for the addresses for the client's IA_NA
   bindings.  To allow the client to convey its FQDN to the server, this
   document defines a new DHCPv6 option called "Client FQDN".  The
   Client FQDN option also contains Flags that DHCPv6 clients and
   servers use to negotiate who does which updates.

   The code for this option is 39.  Its minimum length is 1 octet.

   The format of the DHCPv6 Client FQDN option is shown below:

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |          OPTION_FQDN          |         option-len            |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |   flags       |                                               |
       +-+-+-+-+-+-+-+-+                                               |
       .                                                               .
       .                          domain-name                          .
       .                                                               .
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

         option-code      OPTION_CLIENT_FQDN (39)

         option-len       1 + length of domain name

         flags            flag bits used between client and server to
                          negotiate who performs which updates

         domain-name      the partial or fully qualified domain name
                          (with length option-len - 1)


   The Client FQDN option MUST only appear in a message's options field
   and applies to all addresses for all IA_NA bindings in the
   transaction.

4.1.  The Flags Field

   The format of the Flags field is:

        0 1 2 3 4 5 6 7
       +-+-+-+-+-+-+-+-+
       |  MBZ    |N|O|S|
       +-+-+-+-+-+-+-+-+


   The "S" bit indicates whether the server SHOULD or SHOULD NOT perform
   the AAAA RR (FQDN-to-address) DNS updates.  A client sets the bit to
   0 to indicate that the server SHOULD NOT perform the updates and 1 to
   indicate that the server SHOULD perform the updates.  The state of
   the bit in the reply from the server indicates the action to be taken
   by the server; if it is 1, the server has taken responsibility for
   AAAA RR updates for the FQDN.

   The "O" bit indicates whether the server has overridden the client's
   preference for the "S" bit.  A client MUST set this bit to 0.  A
   server MUST set this bit to 1 if the "S" bit in its reply to the
   client does not match the "S" bit received from the client.

   The "N" bit indicates whether the server SHOULD NOT perform any DNS
   updates.  A client sets this bit to 0 to request that the server
   SHOULD perform updates (the PTR RR and possibly the AAAA RR based on
   the "S" bit) or to 1 to request that the server SHOULD NOT perform
   any DNS updates.  A server sets the "N" bit to indicate whether the
   server SHALL (0) or SHALL NOT (1) perform DNS updates.  If the "N"
   bit is 1, the "S" bit MUST be 0.

   The remaining bits in the Flags field are reserved for future
   assignment.  DHCPv6 clients and servers that send the Client FQDN
   option MUST clear the MBZ bits, and they MUST ignore these bits.

4.2.  The Domain Name Field

   The Domain Name part of the option carries all or part of the FQDN of
   a DHCPv6 client.  The data in the Domain Name field MUST be encoded
   as described in Section 8 of [5].  In order to determine whether the
   FQDN has changed between message exchanges, the client and server
   MUST NOT alter the Domain Name field contents unless the FQDN has
   actually changed.

   A client MAY be configured with a fully qualified domain name or with
   a partial name that is not fully qualified.  If a client knows only
   part of its name, it MAY send a name that is not fully qualified,
   indicating that it knows part of the name but does not necessarily
   know the zone in which the name is to be embedded.

   To send a fully qualified domain name, the Domain Name field is set
   to the DNS-encoded domain name including the terminating zero-length
   label.  To send a partial name, the Domain Name field is set to the
   DNS-encoded domain name without the terminating zero-length label.

   A client MAY also leave the Domain Name field empty if it desires the
   server to provide a name.

   Servers SHOULD send the complete fully qualified domain name in
   Client FQDN options.

5.  DHCPv6 Client Behavior

   The following describes the behavior of a DHCPv6 client that
   implements the Client FQDN option.

   A client MUST only include the Client FQDN option in SOLICIT,
   REQUEST, RENEW, or REBIND messages.

   A client that sends the Client FQDN option MUST also include the
   option in the Option Request option if it expects the server to
   include the Client FQDN option in any responses.

5.1.  Client Desires to Update AAAA RRs

   If a client that owns/maintains its own FQDN wants to be responsible
   for updating the FQDN-to-IPv6-address mapping for the FQDN and
   address(es) used by the client, the client MUST include the Client
   FQDN option in the SOLICIT with Rapid Commit, REQUEST, RENEW, and
   REBIND message originated by the client.  A client MAY choose to
   include the Client FQDN option in its SOLICIT messages.  The "S",
   "O", and "N" bits in the Flags field in the option MUST be 0.

   Once the client's DHCPv6 configuration is completed (the client
   receives a REPLY message and successfully completes a final check on
   the parameters passed in the message), the client MAY originate an
   update for the AAAA RRs (associated with the client's FQDN) unless
   the server has set the "S" bit to 1.  If the "S" is 1, the DHCPv6
   client SHOULD NOT initiate an update for the name in the server's
   returned Client FQDN option Domain Name field.  However, a DHCPv6
   client that is explicitly configured with a FQDN MAY ignore the state
   of the "S" bit if the server's returned name matches the client's
   configured name.

5.2.  Client Desires Server to Do DNS Updates

   A client can choose to delegate the responsibility for updating the
   FQDN-to-IPv6-address mapping for the FQDN and address(es) used by the
   client to the server.  In order to inform the server of this choice,
   the client SHOULD include the Client FQDN option in its SOLICIT with
   Rapid Commit, REQUEST, RENEW, and REBIND messages and MAY include the
   Client FQDN option in its SOLICIT.  The "S" bit in the Flags field in
   the option MUST be 1, and the "O" and "N" bits MUST be 0.

5.3.  Client Desires No Server DNS Updates

   A client can choose to request that the server perform no DNS updates
   on its behalf.  In order to inform the server of this choice, the
   client SHOULD include the Client FQDN option in its SOLICIT with
   Rapid Commit, REQUEST, RENEW, and REBIND messages and MAY include the
   Client FQDN option in its SOLICIT.  The "N" bit in the Flags field in
   the option MUST be 1, and the "S" and "O" bits MUST be 0.

   Once the client's DHCPv6 configuration is completed (the client
   receives a REPLY message and successfully completes a final check on
   the parameters passed in the message), the client MAY originate its
   DNS updates provided the server's "N" bit is 1.  If the server's "N"
   bit is 0, the server MAY perform the PTR RR updates; it MAY also
   perform the AAAA RR updates if the "S" bit is 1.

5.4.  Domain Name and DNS Update Issues

   As there is a possibility that the DHCPv6 server is configured to
   complete or replace a domain name that the client sends, the client
   MAY find it useful to send the Client FQDN option in its SOLICIT
   messages.  If the DHCPv6 server returns different Domain Name data in
   its ADVERTISE message, the client could use that data in performing
   its own eventual AAAA RR update, or in forming the Client FQDN option
   that it sends in its subsequent messages.  There is no requirement
   that the client send identical Client FQDN option data in its
   SOLICIT, REQUEST, RENEW, or REBIND messages.  In particular, if a
   client has sent the Client FQDN option to its server, and the
   configuration of the client changes so that its notion of its domain
   name changes, it MAY send the new name data in a Client FQDN option
   when it communicates with the server again.  This MAY cause the
   DHCPv6 server to update the name associated with the PTR records and,
   if the server updated the AAAA record representing the client, to
   delete that record and attempt an update for the client's current
   domain name.

   A client that delegates the responsibility for updating the FQDN-to-
   IPv6-address mapping to a server will not receive any indication
   (either positive or negative) from the server as to whether the
   server was able to perform the update.  The client MAY use a DNS
   query to check whether the mapping is up to date.  However, depending
   on the load on the DHCPv6 and DNS servers and the DNS propagation
   delays, the client can only infer success.  If the information is not
   found to be up to date in DNS, the authoritative servers might not
   have completed the updates or zone transfers, or caching resolvers
   may yet have updated their caches.

   If a client releases an address prior to the expiration of the valid
   lifetime and the client is responsible for updating its AAAA RR, the
   client SHOULD delete the AAAA RR associated with the address before
   sending a RELEASE message.  Similarly, if a client is responsible for
   updating its AAAA RRs, but is unable to renew the lifetimes for an
   address, the client SHOULD attempt to delete the AAAA RR before the
   lifetime on the address is no longer valid.  A DHCPv6 client that has
   not been able to delete an AAAA RR that it added SHOULD attempt to
   notify its administrator, perhaps by emitting a log message.

   A client SHOULD NOT perform DNS updates to AAAA RRs for its non-
   Global Unicast addresses [7] or temporary addresses [6].

6.  DHCPv6 Server Behavior

   The following describes the behavior of a DHCPv6 server that
   implements the Client FQDN option when the client's message includes
   the Client FQDN option.

   Servers MUST only include a Client FQDN option in ADVERTISE and REPLY
   messages if the client included a Client FQDN option and the Client
   FQDN option is requested by the Option Request option in the client's
   message to which the server is responding.

   The server examines its configuration and the Flag bits in the
   client's Client FQDN option to determine how to respond:

   o  The server sets to 0 the "S", "O", and "N" bits in its copy of the
      option it will return to the client.

   o  If the client's "N" bit is 1 and the server's configuration allows
      it to honor the client's request for no server-initiated DNS
      updates, the server sets the "N" bit to 1.

   o  Otherwise, if the client's "S" bit is 1 and the server's
      configuration allows it to honor the client's request for the
      server to initiate AAAA RR DNS updates, the server sets the "S" to
      1.  If the server's "S" bit does not match the client's "S" bit,
      the server sets the "O" bit to 1.

   The server MAY be configured to use the name supplied in the client's
   Client FQDN option, or it MAY be configured to modify the supplied
   name or to substitute a different name.  The server SHOULD send its
   notion of the complete FQDN for the client in the Domain Name field.
   The server MAY simply copy the Domain Name field from the Client FQDN
   option that the client sent to the server.

6.1.  When to Perform DNS Updates

   The server SHOULD NOT perform any DNS updates if the "N" bit is 1 in
   the Flags field of the Client FQDN option in the REPLY messages (to
   be) sent to the client.  However, the server SHOULD delete any RRs
   that it previously added via DNS updates for the client.

   The server MAY perform the PTR RR DNS update (unless the "N" bit is
   1).

   The server MAY perform the AAAA RR DNS update if the "S" bit is 1 in
   the Flags field of the Client FQDN option in the REPLY message (to
   be) sent to the client.

   The server MAY perform these updates even if the client's message did
   not carry the Client FQDN option.  The server MUST NOT initiate DNS
   updates when responding with an ADVERTISE message to the client.

   The server MAY complete its DNS updates (PTR RR or PTR and AAAA RR)
   before or after sending the REPLY message to the client.

   If the server's AAAA RR DNS update does not complete until after the
   server has replied to the DHCPv6 client, the server's interaction
   with the DNS server MAY cause the DHCPv6 server to change the domain
   name that it associates with the client.  This can occur, for
   example, if the server detects and resolves a domain-name conflict
   [8].  In such cases, the domain name that the server returns to the
   DHCPv6 client would change between two DHCPv6 exchanges.

   If the server previously performed DNS updates for the client and the
   client's information has not changed, the server MAY skip performing
   additional DNS updates.

   When a server receives a RELEASE or DECLINE for an address, detects
   that the valid lifetime on an address that the server bound to a
   client has expired, or terminates a binding on an address prior to
   the binding's expiration time (for instance, by sending a REPLY with
   a zero valid lifetime for an address), the server SHOULD delete any
   PTR RR that it associated with the address via DNS update.  In
   addition, if the server took responsibility for AAAA RRs, the server
   SHOULD also delete the AAAA RR.

7.  DNS RR TTLs

   RRs associated with DHCP clients may be more volatile than statically
   configured RRs.  DHCP clients and servers that perform dynamic
   updates should attempt to specify resource record TTLs that reflect
   this volatility, in order to minimize the possibility that answers to
   DNS queries will return records that refer to DHCP IP address
   assignments that have expired or been released.

   The coupling among primary, secondary, and caching DNS servers is
   'loose'; that is a fundamental part of the design of the DNS.  This
   looseness makes it impossible to prevent all possible situations in
   which a resolver may return a record reflecting a DHCP-assigned IP
   address that has expired or been released.  In deployment, this
   rarely, if ever, represents a significant problem.  Most DHCP-managed
   clients are infrequently looked up by name in the DNS, and the
   deployment of IXFR [13] and NOTIFY [14] can reduce the latency
   between updates and their visibility at secondary servers.

   We suggest these basic guidelines for implementers.  In general, the
   TTLs for RRs added as a result of DHCP IP address assignment activity
   SHOULD be less than the initial lifetime.  The RR TTL on a DNS record
   added SHOULD NOT exceed 1/3 of the lifetime, but SHOULD NOT be less
   than 10 minutes.  We recognize that individual administrators will
   have varying requirements: DHCP servers and clients SHOULD allow
   administrators to configure TTLs and upper and lower bounds on the
   TTL values, either as an absolute time interval or as a percentage of
   the lease lifetime.

   While clients and servers MAY update the TTL of the records as the
   lifetime is about to expire, there is no requirement that they do so
   as this puts additional load on the DNS system with likely little
   benefit.

8.  DNS Update Conflicts

   This document does not resolve how a DHCPv6 client or server prevent
   name conflicts.  This document addresses only how a DHCPv6 client and
   server negotiate the fully qualified domain name and who will perform
   the DNS updates.

   Implementers of this work will need to consider how name conflicts
   will be prevented.  If a DNS updater needs a security token in order
   to successfully perform DNS updates on a specific name, name
   conflicts can only occur if multiple updaters are given a security
   token for that name.  Or, if the fully qualified domains are based on
   the specific address bound to a client, conflicts will not occur.
   Or, a name conflict resolution technique as described in "Resolving
   Name Conflicts" [8]) SHOULD be used.

9.  IANA Considerations

   The IANA has assigned DHCPv6 option code 39 for the Client FQDN
   option.

10.  Security Considerations

   Unauthenticated updates to the DNS can lead to tremendous confusion,
   through malicious attack or through inadvertent misconfiguration.
   Administrators need to be wary of permitting unsecured DNS updates to
   zones that are exposed to the global Internet.  Both DHCPv6 clients
   and servers SHOULD use some form of update request origin
   authentication procedure (e.g., Secure DNS Dynamic Update [11]) when
   performing DNS updates.

   Whether a DHCPv6 client is responsible for updating an FQDN-to-IPv6-
   address mapping or whether this is the responsibility of the DHCPv6
   server is a site-local matter.  The choice between the two
   alternatives is likely based on the security model that is used with
   the DNS update protocol (e.g., only a client may have sufficient
   credentials to perform updates to the FQDN-to-IPv6-address mapping
   for its FQDN).

   Whether a DHCPv6 server is always responsible for updating the FQDN-
   to-IPv6-address mapping (in addition to updating the IPv6-to-FQDN
   mapping), regardless of the wishes of an individual DHCPv6 client, is
   also a site-local matter.  The choice between the two alternatives is
   likely based on the security model that is being used with DNS
   updates.  In cases where a DHCPv6 server is performing DNS updates on
   behalf of a client, the DHCPv6 server SHOULD be sure of the DNS name
   to use for the client, and of the identity of the client.

   Depending on the presence of or type of authentication used with the
   Authentication option, a DHCPv6 server may not have much confidence
   in the identities of its clients.  There are many ways for a DHCPv6
   server to develop a DNS name to use for a client, but only in certain
   circumstances will the DHCPv6 server know for certain the identity of
   the client.

   It is critical to implement proper conflict resolution, and the
   security considerations of conflict resolution apply [8].

11.  Acknowledgements

   Many thanks to Mark Stapp and Yakov Rekhter, as this document is
   based on the DHCPv4 Client FQDN option [9], and to Ralph Droms, Ted
   Lemon, Josh Littlefield, Kim Kinnear, Pekka Savola, and Mark Stapp
   for their review and comments.

12.  References

12.1.  Normative References

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

   [2]  Mockapetris, P., "Domain names - concepts and facilities",
        STD 13, RFC 1034, November 1987.

   [3]  Mockapetris, P., "Domain names - implementation and
        specification", STD 13, RFC 1035, November 1987.

   [4]  Vixie, P., Thomson, S., Rekhter, Y., and J. Bound, "Dynamic
        Updates in the Domain Name System (DNS UPDATE)", RFC 2136,
        April 1997.

   [5]  Droms, R., Bound, J., Volz, B., Lemon, T., Perkins, C., and M.
        Carney, "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)",
        RFC 3315, July 2003.

   [6]  Narten, T. and R. Draves, "Privacy Extensions for Stateless
        Address Autoconfiguration in IPv6", RFC 3041, January 2001.

   [7]  Hinden, R. and S. Deering, "IP Version 6 Addressing
        Architecture", RFC 4291, February 2006.

   [8]  Stapp, M. and B. Volz, "Resolution of Fully Qualified Domain
        Name (FQDN) Conflicts among Dynamic Host Configuration Protocol
        (DHCP) Clients", RFC 4703, October 2006.

12.2.  Informative References

   [9]   Stapp, M., Volz, B., and Y. Rekhter, "The Dynamic Host
         Configuration Protocol (DHCP) Client Fully Qualified Domain
         Name (FQDN) Option", RFC 4702, October 2006.

   [10]  Marine, A., Reynolds, J., and G. Malkin, "FYI on Questions and
         Answers - Answers to Commonly asked "New Internet User"
         Questions", FYI 4, RFC 1594, March 1994.

   [11]  Wellington, B., "Secure Domain Name System (DNS) Dynamic
         Update", RFC 3007, November 2000.

   [12]  Thomson, S., Huitema, C., Ksinant, V., and M. Souissi, "DNS
         Extensions to Support IP Version 6", RFC 3596, October 2003.

   [13]  Ohta, M., "Incremental Zone Transfer in DNS", RFC 1995,
         August 1996.

   [14]  Vixie, P., "A Mechanism for Prompt Notification of Zone Changes
         (DNS NOTIFY)", RFC 1996, August 1996.

Author's Address

   Bernard Volz
   Cisco Systems, Inc.
   1414 Massachusetts Ave.
   Boxborough, MA  01719
   USA

   Phone: +1 978 936 0382
   EMail: volz@cisco.com

Full Copyright Statement

   Copyright (C) The Internet Society (2006).

   This document is subject to the rights, licenses and restrictions
   contained in BCP 78, and except as set forth therein, the authors
   retain all their rights.

   This document and the information contained herein are provided on an
   "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
   OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
   ENGINEERING TASK FORCE DISCLAIM 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.

Intellectual Property

   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at
   http://www.ietf.org/ipr.

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights that may cover technology that may be required to implement
   this standard.  Please address the information to the IETF at
   ietf-ipr@ietf.org.

Acknowledgement

   Funding for the RFC Editor function is provided by the IETF
   Administrative Support Activity (IASA).

ISO Plain Language

On this page
PROBLEMWhen a site hands out an IPv6 address with DHCPv6, the DNS records that map the host's name to its address and back have to be kept in step, and nobody had defined how the client and the server should divide that job.
MECHANISMA new DHCPv6 option, code 39, carries the client's fully qualified domain name and three flag bits. The S, O and N bits tell the server whether it should update the AAAA record, the PTR record, both or neither.
WHY IT MATTERSClient and server agree on who updates which DNS record, and the RFC suggests record lifetimes that match the lease. Name conflicts are left to a separate document.

§1What the Client FQDN Option Is For

DNS maintains, among other things, the information about the mapping between hosts' fully qualified domain names and the IPv6 addresses assigned to the hosts. That information is held in two types of resource record: AAAA and PTR. The DNS update specification in RFC 2136 describes a mechanism that enables DNS information to be updated over a network.

The Dynamic Host Configuration Protocol for IPv6 provides a mechanism by which a host, a DHCPv6 client, can acquire configuration information along with its stateful IPv6 addresses. This document specifies a new DHCPv6 option, the Client FQDN option, which DHCPv6 clients and servers use to exchange information about the client's fully qualified domain name and about who has the responsibility for updating the DNS with the associated AAAA and PTR records.

DNS holds a record that maps a host's name to its address and another that maps the address back. DHCPv6 hands out the address, so something has to keep those records in step. This RFC adds an option through which the client and the server say who will do it.

§3Two Models of Operation

When a DHCPv6 client acquires an address, the site's administrator may want the AAAA record for the client's fully qualified domain name and the PTR record for the acquired address to be updated. Acquiring an address involves two entities, a client and a server, and in principle each could perform none, one or both of the DNS update transactions. In practice, the option is primarily intended to work in two cases: the client updates the AAAA record and the server updates the PTR record, or the server updates both. The only difference between the two cases is whether the FQDN-to-address mapping is updated by the client or by the server. The address-to-FQDN mapping is updated by the server in both cases.

These two cases matter because of authority over the DNS domain names. A client may be given authority over mapping its own AAAA records, or that authority may be restricted to a server to stop the client listing arbitrary addresses or associating its addresses with arbitrary domain names. In all cases, the only reasonable place for the authority over the PTR records associated with the address is in the server that allocates the address.

A third case is supported, in which the client asks the server to perform no updates, but it is presumed to be rare because of the authority issues. Whether a site permits all, some or no DHCPv6 servers and clients to perform DNS updates into the zones it controls is entirely a matter of local administrative policy. The range of policies is very broad, from sites where only the servers hold credentials the DNS servers accept, to sites where each individual client has been configured with credentials that let it modify its own domain name. The specification applies only to DHCPv6 client and server processes; it does not apply to other processes that initiate DNS updates.

There are two sensible arrangements. Either the client keeps its own name record up to date and the server keeps the reverse record, or the server does both. The split matters for authority: a client that can edit its own name record could also claim other names, so some sites keep all the editing at the server.

THE NEGOTIATION AND THE DNS UPDATE; ILLUSTRATIVE, NOT FROM THE RFC
DHCPv6 clientDHCPv6 serverDNS server
  1. SOLICIT WITH CLIENT FQDN OPTIONCarries the client's domain name and the flag bits.
  2. ADVERTISE OR REPLY WITH CLIENT FQDN OPTIONThe flags in the reply say who will do the updates.
  3. DNS UPDATEAdds or removes the AAAA and PTR records.

§4The Option and Its Flags

To update the address-to-FQDN mapping, a DHCPv6 server needs to know the FQDN of the client for the addresses in the client's IA_NA bindings. To let the client convey its FQDN, this document defines a new DHCPv6 option called Client FQDN. The code for this option is 39 and its minimum length is 1 octet. The option holds the option code, the option length, a flags field and the domain name, whose length is the option length minus one. The Client FQDN option must only appear in a message's options field, and it applies to all addresses for all IA_NA bindings in the transaction.

The flags field is one octet holding four flag bits and the rest reserved. The S bit indicates whether the server should or should not perform the AAAA record, the FQDN-to-address, DNS updates. A client sets the bit to 0 to say the server should not perform the updates and to 1 to say it should. The state of the bit in the server's reply indicates the action taken by the server; if it is 1, the server has taken responsibility for AAAA record updates for the FQDN.

The O bit indicates whether the server has overridden the client's preference for the S bit. A client must set this bit to 0, and a server must set it to 1 if the S bit in its reply does not match the S bit it received.

The N bit indicates whether the server should not perform any DNS updates. A client sets it to 0 to request that the server should perform updates, the PTR record and possibly the AAAA record based on the S bit, or to 1 to request that the server should not perform any DNS updates. A server sets the N bit to say whether the server shall, at 0, or shall not, at 1, perform DNS updates. If the N bit is 1, the S bit must be 0.

The remaining bits are reserved for future assignment. Clients and servers that send the option must clear the reserved bits and must ignore them when reading.

The option is number 39 and holds the client's name plus three flags. The S flag asks the server to look after the name record. The N flag asks the server to do no DNS work at all. The O flag is the server saying that it has ignored the client's preference. Any spare bits are set to zero and ignored.

THE OPTION FIELDS AS PRINTED IN THE RFC
DHCPv6 Client FQDN option
option-code      OPTION_CLIENT_FQDN (39)
option-len       1 + length of domain name
flags            flag bits used between client and server to
                 negotiate who performs which updates
domain-name      the partial or fully qualified domain name
                 (with length option-len - 1)
THE THREE FLAG BITS IN THE FLAGS FIELD
  • SSet by the client to 1 to ask the server to do the AAAA record update, the one that maps the name to the address.
  • OSet by the server to 1 when the S bit in its reply does not match the S bit the client sent.
  • NSet to 1 to ask for no server DNS updates at all. When it is 1, the S bit must be 0.

§4.2The Domain Name Field

The domain name part of the option carries all or part of the fully qualified domain name of a DHCPv6 client. The data must be encoded as described in section 8 of RFC 3315. In order to determine whether the FQDN has changed between message exchanges, the client and server must not alter the domain name field contents unless the FQDN has actually changed.

A client may be configured with a fully qualified domain name, or with a partial name that is not fully qualified. If a client knows only part of its name, it may send a name that is not fully qualified, indicating that it knows part of the name but does not necessarily know the zone in which the name is to be embedded. To send a fully qualified domain name, the domain name field is set to the DNS-encoded name including the terminating zero-length label. To send a partial name, it is set to the DNS-encoded name without the terminating zero-length label. A client may also leave the domain name field empty if it wants the server to provide a name. Servers should send the complete fully qualified domain name in Client FQDN options.

The name travels in the DNS wire format. A client that knows only part of its name can send that part, leaving off the final empty label, and a client with no name at all can send an empty field and let the server choose one.

§5What the Client Does

A client must only include the Client FQDN option in SOLICIT, REQUEST, RENEW or REBIND messages. A client that sends the option must also include the option in the Option Request option if it expects the server to include the Client FQDN option in any responses.

A client that owns and maintains its own fully qualified domain name, and wants to be responsible for updating the FQDN-to-address mapping, must include the option in its SOLICIT with Rapid Commit, REQUEST, RENEW and REBIND messages, and may include it in its other SOLICIT messages. The S, O and N bits must all be 0. Once the client's configuration is complete, meaning it has received a REPLY and successfully completed a final check on the parameters in it, the client may start an update for its AAAA records, unless the server set the S bit to 1. If the S bit is 1, the client should not update the name in the server's returned domain name field. A client explicitly configured with a fully qualified domain name may ignore the state of the S bit if the server's returned name matches the client's configured name.

A client that wants the server to do the updates sets the S bit to 1 and the O and N bits to 0. A client that wants no server-side DNS updates sets the N bit to 1 and the S and O bits to 0. In that last case, once its configuration is complete the client may do its own DNS updates as long as the server's N bit is 1. If the server's N bit is 0, the server may perform the PTR record updates, and may also perform the AAAA record updates if the S bit is 1.

A client that delegates the FQDN-to-address mapping to a server gets no indication, positive or negative, of whether the server managed the update. The client may use a DNS query to check whether the mapping is up to date, but because of load on the servers and DNS propagation delays it can only infer success. If a client releases an address before the valid lifetime expires and the client is responsible for updating its AAAA record, it should delete that record before sending the RELEASE message. A client that has not been able to delete an AAAA record it added should try to notify its administrator. A client should not perform DNS updates to AAAA records for its non-global unicast addresses or its temporary addresses.

The client says in the option whether it will do the name update itself or wants the server to do it. After the lease is confirmed, the client may then update the DNS, unless the server took that job on. The server never reports back whether the update worked, so a client that wants to check has to query DNS and can only guess from the answer.

§6What the Server Does

Servers must only include a Client FQDN option in ADVERTISE and REPLY messages if the client included the option and the option is requested in the Option Request option of the client's message.

The server examines its configuration and the flag bits in the client's option to decide how to respond. It sets the S, O and N bits to 0 in the copy of the option it returns. If the client's N bit is 1 and the server's configuration allows it to honour the request for no server-initiated DNS updates, the server sets the N bit to 1. Otherwise, if the client's S bit is 1 and the server's configuration allows it to honour the request for the server to do the AAAA updates, the server sets the S bit to 1, and if that does not match the client's S bit it also sets the O bit to 1.

The server may be configured to use the name supplied in the client's option, or to modify the supplied name or substitute a different one. It should send its notion of the complete fully qualified domain name for the client in the domain name field, and may simply copy the field from the client's option.

The server should not perform any DNS updates if the N bit is 1 in the option of the REPLY it is about to send, but it should delete any records it previously added for the client. It may perform the PTR record update unless the N bit is 1, and it may perform the AAAA record update if the S bit is 1 in the option of the REPLY it is about to send. It may perform these updates even if the client's message did not carry the option, and it must not initiate DNS updates when responding with an ADVERTISE message. It may complete its updates before or after sending the REPLY. If the AAAA update finishes only after the reply, the server's later interaction with the DNS server may cause it to change the domain name it associates with the client, for example if it detects and resolves a domain-name conflict, in which case the name it returns to the client would change between two exchanges. If the server previously performed updates for the client and the client's information has not changed, it may skip further updates.

When a server receives a RELEASE or DECLINE for an address, detects that the valid lifetime has expired, or ends a binding early by sending a REPLY with a zero valid lifetime, it should delete any PTR record it associated with the address. If the server took responsibility for AAAA records, it should delete those as well.

The server returns the option only if the client asked for it, and sets the flags to match what its own policy allows. It can keep the client's name, change it or replace it, and should send back the full name it settled on. When the lease ends, the server removes the DNS records it added.

§7Choosing DNS Record Lifetimes

Records associated with DHCP clients may be more volatile than statically configured records. DHCP clients and servers that perform dynamic updates should try to specify resource record lifetimes that reflect this volatility, in order to minimise the possibility that answers to DNS queries return records that refer to address assignments that have expired or been released.

The coupling among primary, secondary and caching DNS servers is loose, and it is impossible to prevent all situations in which a resolver may return a record reflecting an expired or released address. In deployment this rarely represents a significant problem. Most DHCP-managed clients are infrequently looked up by name in the DNS, and the deployment of incremental zone transfer and NOTIFY can reduce the latency between updates and their visibility at secondary servers.

The RFC suggests basic guidelines for implementers. In general, the lifetimes for records added as a result of address assignment should be less than the initial lifetime. The record lifetime on a DNS record added should not exceed one third of the lifetime, but should not be less than 10 minutes. Individual administrators will have varying requirements, so DHCP servers and clients should allow administrators to configure the lifetime and its upper and lower bounds, either as an absolute time interval or as a percentage of the lease lifetime. While clients and servers may update the lifetime of the records as the lease is about to expire, there is no requirement that they do so, as this puts additional load on the DNS system with likely little benefit.

DNS records for DHCP clients go stale quickly, so they should carry short lifetimes. The RFC suggests a lifetime of no more than a third of the lease and no less than ten minutes, and asks implementations to let the administrator set those bounds.

§10Security Considerations and Name Conflicts

Unauthenticated updates to the DNS can lead to tremendous confusion, through malicious attack or through inadvertent misconfiguration. Administrators need to be wary of permitting unsecured DNS updates to zones that are exposed to the global Internet. Both DHCPv6 clients and servers should use some form of update request origin authentication, such as secure DNS dynamic update, when performing DNS updates.

Whether a client is responsible for updating its name mapping or whether that is the responsibility of the server is a site-local matter, likely based on the security model used with the DNS update protocol, for example where only the client holds sufficient credentials. Whether the server is always responsible for the mapping, regardless of the wishes of an individual client, is also a site-local matter. In cases where a server is performing DNS updates on behalf of a client, it should be sure of the DNS name to use for the client and of the identity of the client. Depending on the presence of or type of authentication used with the DHCPv6 Authentication option, a server may not have much confidence in the identities of its clients.

This document does not resolve how a client or server prevents name conflicts; it addresses only how they negotiate the fully qualified domain name and who performs the DNS updates. Implementers will need to consider how name conflicts will be prevented. If a DNS updater needs a security token to perform updates on a specific name, conflicts can only occur if multiple updaters are given a token for that name. Or, if the fully qualified domains are based on the specific address bound to a client, conflicts will not occur. Or, a name conflict resolution technique should be used. It is critical to implement proper conflict resolution, and the security considerations of conflict resolution apply.

Anyone allowed to update the DNS without authenticating can cause a great deal of confusion, so both sides should authenticate their updates. This RFC does not say what happens when two machines claim the same name; that is left to a separate document on resolving name conflicts.

§9References