Internet Engineering Task Force J. Bound
INTERNET DRAFT Compaq Computer Corp.
DHC Working Group M. Carney
Obsoletes: draft-ietf-dhc-dhcpv6-14.txt Sun Microsystems, Inc
C. Perkins
Nokia Research Center
5 May 2000
Dynamic Host Configuration Protocol for IPv6 (DHCPv6)
draft-ietf-dhc-dhcpv6-15.txt
Status of This Memo
This document is a submission by the Dynamic Host Configuration
Working Group of the Internet Engineering Task Force (IETF). Comments
should be submitted to the dhcp-v6@bucknell.edu mailing list.
Distribution of this memo is unlimited.
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-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.
Abstract
The Dynamic Host Configuration Protocol for IPv6 (DHCP) enables
DHCP servers to pass configuration parameters using extensions to
IPv6 nodes. It offers the capability of automatic allocation of
reusable network addresses and additional configuration flexibility.
This protocol is a stateful counterpart to ``IPv6 Stateless Address
Autoconfiguration'' [15], and can be used separately or concurrently
with the latter to obtain configuration parameters.
Bound, Carney, Perkins Expires 1 November 2000 [Page i]
Internet Draft DHCP for IPv6 5 May 2000
Contents
Status of This Memo i
Abstract i
1. Introduction 1
2. Terminology 2
2.1. IPv6 Terminology . . . . . . . . . . . . . . . . . . . . 2
2.2. DHCP Terminology . . . . . . . . . . . . . . . . . . . . 3
3. DHCP Constants 5
3.1. Multicast Addresses . . . . . . . . . . . . . . . . . . . 5
3.2. UDP ports . . . . . . . . . . . . . . . . . . . . . . . . 5
3.3. DHCP message types . . . . . . . . . . . . . . . . . . . 6
3.4. Error Values . . . . . . . . . . . . . . . . . . . . . . 8
3.4.1. Generic Error Values . . . . . . . . . . . . . . 8
3.4.2. Server-specific Error Values . . . . . . . . . . 8
3.5. Configuration Variables . . . . . . . . . . . . . . . . . 8
4. Requirements 9
5. Background 9
6. Design Goals 11
7. Non-Goals 11
8. Overview 12
8.1. How does a node know to use DHCP? . . . . . . . . . . . . 12
8.2. How does a client find out about DHCP agents? . . . . . . 12
8.3. What if the client and server(s) are on different links? 12
8.4. How does a client request configuration parameters from
servers? . . . . . . . . . . . . . . . . . . . . . . . 13
8.5. What are releasable resources, and when are they used? . 13
8.6. Can a client release its releasable resources before the lease
expires? . . . . . . . . . . . . . . . . . . . . . . . 14
8.7. What if the client determines its releasable resource is
already being used by another client? . . . . . . . . 14
8.8. How are clients notified of server configuration changes? 14
9. Message Formats 15
9.1. DHCP Solicit Message Format . . . . . . . . . . . . . . . 15
9.2. DHCP Advertise Message Format . . . . . . . . . . . . . . 16
9.3. DHCP Request Message Format . . . . . . . . . . . . . . . 18
Bound, Carney, Perkins Expires 1 November 2000 [Page ii]
Internet Draft DHCP for IPv6 5 May 2000
9.4. DHCP Reply Message Format . . . . . . . . . . . . . . . . 19
9.5. DHCP Release Message Format . . . . . . . . . . . . . . . 20
9.6. DHCP Reconfigure Message Format . . . . . . . . . . . . . 22
9.7. DHCP Reconfigure-reply Message Format . . . . . . . . . . 23
9.8. DHCP Reconfigure-init Message Format . . . . . . . . . . 24
10. DHCP Server Solicitation and Subnet Prefix Discovery 25
10.1. Solicit Message Validation . . . . . . . . . . . . . . . 25
10.2. Advertise Message Validation . . . . . . . . . . . . . . 25
10.3. Client Behavior . . . . . . . . . . . . . . . . . . . . . 26
10.3.1. Creation and sending of the Solicit message . . . 26
10.3.2. Time out and retransmission of Solicit Messages . 27
10.3.3. Receipt of Advertise messages . . . . . . . . . . 27
10.4. Relay Behavior . . . . . . . . . . . . . . . . . . . . . 28
10.4.1. Relaying of Solicit messages . . . . . . . . . . 28
10.4.2. Relaying of Advertise messages . . . . . . . . . 28
10.5. Server Behavior . . . . . . . . . . . . . . . . . . . . . 28
10.5.1. Receipt of Solicit messages . . . . . . . . . . . 28
10.5.2. Creation and sending of Advertise messages . . . 29
11. DHCP Client-Initiated Configuration Exchange 29
11.1. Request Message Validation . . . . . . . . . . . . . . . 29
11.2. Reply Message Validation . . . . . . . . . . . . . . . . 30
11.3. Release Message Validation . . . . . . . . . . . . . . . 31
11.4. Client Behavior . . . . . . . . . . . . . . . . . . . . . 31
11.4.1. Creation and sending of Request messages . . . . 32
11.4.2. Time out and retransmission of Request Messages . 33
11.4.3. Receipt of Reply message in response to a Request 33
11.4.4. Creation and sending of Release messages . . . . 33
11.4.5. Time out and retransmission of Release Messages . 34
11.4.6. Receipt of Reply message in response to a Release 35
11.5. Relay Behavior . . . . . . . . . . . . . . . . . . . . . 35
11.5.1. Relaying of Request or Release messages . . . . . 35
11.6. Server Behavior . . . . . . . . . . . . . . . . . . . . . 35
11.6.1. Receipt of Request messages . . . . . . . . . . . 35
11.6.2. Receipt of Release messages . . . . . . . . . . . 36
11.6.3. Creation and sending of Reply messages . . . . . 36
12. DHCP Server-Initiated Configuration Exchange 37
12.1. Reconfigure Message Validation . . . . . . . . . . . . . 37
12.2. Reconfigure-reply Message Validation . . . . . . . . . . 38
12.3. Reconfigure-init Message Validation . . . . . . . . . . . 38
12.4. Server Behavior . . . . . . . . . . . . . . . . . . . . . 38
12.4.1. Creation and sending of Reconfigure messages . . 39
12.4.2. Time out and retransmission of Reconfigure
messages . . . . . . . . . . . . . . . . . 40
12.4.3. Receipt of Reconfigure-reply messages . . . . . . 40
12.4.4. Creation and sending of Reconfigure-init messages 40
Bound, Carney, Perkins Expires 1 November 2000 [Page iii]
Internet Draft DHCP for IPv6 5 May 2000
12.4.5. Time out and retransmission of Reconfigure-init
messages . . . . . . . . . . . . . . . . . 41
12.4.6. Receipt of Request messages . . . . . . . . . . . 41
12.5. Client Behavior . . . . . . . . . . . . . . . . . . . . . 41
12.5.1. Receipt of Reconfigure messages . . . . . . . . . 42
12.5.2. Creation and sending of Reconfigure-reply messages 42
12.5.3. Receipt of Reconfigure-init messages . . . . . . 43
12.5.4. Creation and sending of Request messages . . . . 43
12.5.5. Time out and retransmission of Request messages . 43
12.5.6. Receipt of Reply messages . . . . . . . . . . . . 43
13. Using DHCP for network renumbering 43
13.1. Passive Renumbering . . . . . . . . . . . . . . . . . . . 44
13.2. Active Renumbering . . . . . . . . . . . . . . . . . . . 44
14. DHCP Client Implementator Notes 44
14.1. Primary Interface . . . . . . . . . . . . . . . . . . . . 45
14.2. Advertise Message and Configuration Parameter Caching . . 45
14.3. Time out and retransmission variables . . . . . . . . . . 45
14.4. Server Preference . . . . . . . . . . . . . . . . . . . . 45
15. DHCP Server Implementator Notes 46
15.1. Client Bindings . . . . . . . . . . . . . . . . . . . . . 46
15.2. Reconfigure Considerations . . . . . . . . . . . . . . . 46
15.3. Server Preference . . . . . . . . . . . . . . . . . . . . 46
15.4. Request Message Transaction-ID Cache . . . . . . . . . . 47
16. DHCP Relay Implementator Notes 47
17. Open Issues for Working Group Discussion 47
17.1. Trade-offs: Optional fields in DHCP messages . . . . . . 47
17.2. Use DHCPv4 authentication or the current DHCPv6 method? . 48
17.3. The Reconfigure Message and Subnet Prefix Extensions . . 48
17.4. ``R'' bit in Request message not needed? . . . . . . . . 48
18. Security Considerations 48
19. Year 2000 considerations 49
20. IANA Considerations 49
21. Acknowledgements 50
A. Comparison between DHCPv4 and DHCPv6 50
B. Full Copyright Statement 52
Chair's Address 55
Bound, Carney, Perkins Expires 1 November 2000 [Page iv]
Internet Draft DHCP for IPv6 5 May 2000
Author's Address 55
Bound, Carney, Perkins Expires 1 November 2000 [Page v]
Internet Draft DHCP for IPv6 5 May 2000
1. Introduction
This document describes DHCP for IPv6 (DHCP), a UDP [14] client /
server protocol designed to reduce the cost of management of IPv6
nodes in environments where network managers require more control
over the allocation of network resources more varied than that
offered by ``IPv6 Stateless Autoconfiguration'' [15]. The DHCP is a
stateful counterpart to stateless autoconfiguration. Note that both
stateful and stateless autoconfiguration can be used concurrently in
the same environment, leveraging the strengths of both mechanisms
in order to reduce the cost of ownership and management of network
nodes.
The DHCP reduces the cost of ownership by centralizing the management
of network resources such as IP addresses, routing information, OS
installation information, directory service information, and other
such information on a few DHCP servers, rather than distributing such
information in local configuration files among each network node.
The DHCP is designed to be easily extended to carry new configuration
parameters through the addition of new DHCP ``extensions'' defined to
carry this information. See this document's companion specification,
``Extensions for the Dynamic Host Configuration Protocol for
IPv6'' [2] for specifications of existing extensions as well as
information on the process by which an interested party might specify
new extensions.
Those readers familiar with DHCP for IPv4 [7] will find DHCP for IPv6
provides a superset of features, and benefits from the additional
features of IPv6 and freedom from BOOTP [5]-backward compatibility
constraints. For more information about the differences between DHCP
for IPv6 and DHCP for IPv4, see Appendix A.
This document is organized as follows. Section 2 defines terminology
used throughout this document. Section 3 defines constant values
used by DHCP. Section 4 briefly discusses requirement levels.
Section 5 points the reader to helpful background specifications
covering related IPv6 protocols. Section 6 discusses the design
goals that influenced DHCP. Section 7 identifies some of the
non-goals of this specification. Section 8 gives a high level
overview of DHCP, its message types, and identifies DHCP functional
entities (client, relay, server). Section 9 describes in detail the
format of each DHCP message type. Section 10 discusses DHCP server
solicitation and subnet prefix discovery. Section 11 discusses DHCP
client-initiated configuration information exchange. Section 12
discusses DHCP server-initiated configuration information exchange.
Section 13 describes how DHCP can be used to renumber networks.
Section 14 presents helpful notes for DHCP client implementators.
Section 15 presents helpful notes for DHCP server implementors.
Bound, Carney, Perkins Expires 1 November 2000 [Page 1]
Internet Draft DHCP for IPv6 5 May 2000
Section 16 presents helpful notes for DHCP relay implementors.
Section 18 discusses security considerations for DHCP.
2. Terminology
2.1. IPv6 Terminology
IPv6 terminology relevant to this specification from the IPv6
Protocol [6], IPv6 Addressing Architecture [8], and IPv6 Stateless
Address Autoconfiguration [15] is included below.
address An IP layer identifier for an interface or a set of
interfaces.
unicast address
An identifier for a single interface. A packet sent
to a unicast address is delivered to the interface
identified by that address.
multicast address
An identifier for a set of interfaces (typically
belonging to different nodes). A packet sent to a
multicast address is delivered to all interfaces
identified by that address.
host Any node that is not a router.
IP Internet Protocol Version 6 (IPv6). The terms IPv4 and
IPv6 are used only in contexts where it is necessary to
avoid ambiguity.
interface
A node's attachment to a link.
link A communication facility or medium over which nodes
can communicate at the link layer, i.e., the layer
immediately below IP. Examples are Ethernet (simple or
bridged); Token Ring; PPP links, X.25, Frame Relay, or
ATM networks; and Internet (or higher) layer "tunnels",
such as tunnels over IPv4 or IPv6 itself.
link-layer identifier
a link-layer identifier for an interface. Examples
include IEEE 802 addresses for Ethernet or Token Ring
network interfaces, and E.164 addresses for ISDN links.
link-local address
An IP address having link-only scope, indicated by
Bound, Carney, Perkins Expires 1 November 2000 [Page 2]
Internet Draft DHCP for IPv6 5 May 2000
having the subnet prefix (FE80::0000/64), that can be
used to reach neighboring nodes attached to the same
link. Every interface has a link-local address.
message A unit of data carried in a packet, exchanged between
DHCP agents and clients.
neighbor A node attached to the same link.
node A device that implements IP.
packet An IP header plus payload.
prefix A bit string that consists of some number of initial
bits of an address.
router A node that forwards IP packets not explicitly
addressed to itself.
2.2. DHCP Terminology
Terminology specific to DHCP can be found below.
abort status
A status value returned to the application that has
invoked a DHCP client operation, indicating anything
other than success.
agent address
The address of a neighboring DHCP Agent on the same
link as the DHCP client.
binding A binding (or, client binding) is a group of server
data records indexed by <client's link-local address,
subnet prefix> containing the releasable resource data
which a DHCP server has assigned to a client.
Note that the transaction-ID from the Request message
that produced the assignment of the releasable resource
is also stored in the server data record including the
releasable resource identifier.
DHCP Dynamic Host Configuration Protocol for IPv6. The
terms DHCPv4 and DHCPv6 are used only in contexts where
it is necessary to avoid ambiguity.
Bound, Carney, Perkins Expires 1 November 2000 [Page 3]
Internet Draft DHCP for IPv6 5 May 2000
configuration parameter
An element of the configuration information set on the
server and delivered to the client using DHCP. Such
parameters may be used to carry information to be used
by a node to configure its network subsystem and enable
communication on a link or internetwork, for example.
DHCP client (or client)
A node that initiates requests on a link to obtain
configuration parameters from one or more DHCP servers.
DHCP domain
A chunk of network topology managed by DHCP and
operated by a single administrative entity.
DHCP server (or server)
A server is a node that responds to requests from
clients, and may or may not be on the same link as the
client(s).
DHCP relay (or relay)
A node that acts as an intermediary to deliver DHCP
messages between clients and servers, and is on the
same link as a client.
DHCP agent (or agent)
Either a DHCP server on the same link as a client, or a
DHCP relay.
Releasable resource
Any configuration resource allocated by a server for
a finite period of time. As of this writing, the
only example of such a resource is the IP address.
Releasable resources are carried in extensions
allocated out of the 1--8192 range.
solicit-ID
An unsigned integer generated by the client and
inserted into its DHCP Solicit messages, and echoed
back to the client by the server in its resultant DHCP
Advertise message(s). The client uses the solicit-ID
to match received Advertise messages to Solicit
messages it has generated.
transaction-ID
An unsigned integer to match responses with replies
initiated either by a client or server. Servers
Bound, Carney, Perkins Expires 1 November 2000 [Page 4]
Internet Draft DHCP for IPv6 5 May 2000
allocate their transaction-IDs from the range of
0--1023, and clients allocate their transaction-IDs
from the range of 1024--65535. Limiting clients and
servers to different ranges prevents transaction-ID
collisions (e.g. client and server happen to use the
same transaction-ID for unrelated transactions (e.g.
client Request, server Reconfigure-init).
3. DHCP Constants
This section describes various program and networking constants used
by DHCP.
3.1. Multicast Addresses
The DHCP makes use of the following multicast addresses:
All DHCP Agents address: FF02::1:2
This link-local multicast address is used by clients to
communicate with the on-link agent(s) when they do not
know those agents' link-local address(es). All agents
(servers and relays) are members of this multicast
group.
All DHCP Servers address: FF05::1:3
This site-local multicast address is used by clients or
relays to communicate with server(s), either because
they want to send messages to all servers or because
they do not know the server(s) unicast address(es).
Note that in order for a client to use this address,
it must have an address of sufficient scope to be
reachable by the server(s). All servers within the
site are members of this multicast group.
3.2. UDP ports
The DHCP uses the following destination UDP [14] port numbers. While
source ports MAY be arbitrary, client implementations SHOULD permit
their specification through a local configuration parameter to
facilitate the use of DHCP through firewalls.
546 Client port. Used by agents to send messages to
clients. Also used by servers to send messages to
relays.
Bound, Carney, Perkins Expires 1 November 2000 [Page 5]
Internet Draft DHCP for IPv6 5 May 2000
547 Agent port. Used by clients to send messages to
agents. Also used by relays to send messages to
servers.
3.3. DHCP message types
The DHCP defines the following message types. More detail on these
message types can be found in Section 9. Message types 0 and 9--255
are reserved and MUST be silently ignored.
01 DHCP Solicit
The DHCP Solicit (or Solicit) message is used by clients to
locate servers and (optionally) learn about the subnet prefixes
on the client's link for networks that are managed by DHCP.
This message is multicast using the All-DHCP-Agents address.
Relay(s) forward Solicits as necessary to off-link servers.
Section 9.1 contains more details about the Solicit message.
02 DHCP Advertise
The DHCP Advertise (or Advertise) message is used by servers
responding to Solicits. This message is unicast to the
client's link-local address (if the server and client are
on the same link) or unicast to the relay through which the
Solicit was sent for final delivery to the client.
Section 9.2 contains more details about the Advertise message.
03 DHCP Request
The DHCP Request (or Request) message is used by clients to
request configuration parameters from servers. This message
is unicast to the server if the client has an address with
sufficient scope to be reachable by the server, otherwise it
is unicast to the on-link relay through which the Advertise
message was relayed.
Section 9.3 contains more details about the Request message.
04 DHCP Reply
The DHCP Reply (or Reply) message is used by servers responding
to Request and Release messages. In the case of responding to
a Request message, the Reply contains configuration parameters
destined for the client. This message is unicast to the client
if the client has an address of sufficient scope that is
Bound, Carney, Perkins Expires 1 November 2000 [Page 6]
Internet Draft DHCP for IPv6 5 May 2000
reachable by the server. Otherwise, it is unicast to the relay
through which the Request or Release message was sent for final
delivery to the client.
Section 9.4 contains more details about the Reply message.
05 DHCP Release
The DHCP Release (or Release) message is used by clients to
return one or more instances of releasable resources (e.g. IP
addresses) to servers. This message is unicast to the server
if the client will have an address of sufficient scope after
the Release operation to receive a Reply message. Otherwise,
the Release message is sent through the relay. The server will
acknowledge the receipt of the Release message by sending the
client a Reply message.
Section 9.5 contains more details about the Release message.
06 DHCP Reconfigure
The DHCP Reconfigure (or Reconfigure) message is sent by
servers to client(s). It contains new or updated configuration
parameters for use by the client(s). This message may be
unicast or multicast to the client(s).
Section 9.6 contains more details about the Reconfigure
message.
07 DHCP Reconfigure-reply
The DHCP Reconfigure-reply (or Reconfigure-reply) message is
unicast by client(s) to the server to acknowledge the receipt
of a Reconfigure message.
Section 9.7 contains more details about the Reconfigure-reply
message.
08 DHCP Reconfigure-init
The DHCP Reconfigure-init (or Reconfigure-init) message is set
by server(s) to inform client(s) that the server(s) has new or
updated configuration parameters, and that the client(s) are
to initiate a Request/Reply transaction with the server(s) in
order to receive the updated information.
Section 9.8 contains more details about the Reconfigure-init
message.
Bound, Carney, Perkins Expires 1 November 2000 [Page 7]
Internet Draft DHCP for IPv6 5 May 2000
3.4. Error Values
This section describes error values exchanged between DHCP
implementations.
3.4.1. Generic Error Values
The following symbolic names are used between client and server
implementations to convey error conditions. The following table
contains the actual numeric values for each name. Note that the
numeric values do not start at 1, nor are they consecutive. The
errors are organized in logical groups.
_______________________________________________________________
|_Error_Name___|Error_ID_|Description__________________________|
|_Success______|00_______|Success______________________________|
|_UnspecFail___|16_______|Failure,_reason_unspecified__________|
|_AuthFailed___|17_______|Authentication_failed_or_nonexistent_|
|_PoorlyFormed_|18_______|Poorly_formed_message________________|
|_Unavail______|19_______|Resources_unavailable________________|
3.4.2. Server-specific Error Values
The following symbolic names are used by server implementations to
convey error conditions to clients. The following table contains the
actual numeric values for each name.
_______________________________________________________________
|_Error_Name____|Error_ID_|Description_________________________|
|_NoBinding_____|20_______|Client_record_(binding)_unavailable_|
|_InvalidSource_|21_______|Invalid_Client_IP_address___________|
|_NoServer______|23_______|Relay_cannot_find_Server_Address____|
|_ICMPError_____|64_______|Server_unreachable_(ICMP_error)_____|
3.5. Configuration Variables
This section presents a table of client and server configuration
variables and the default or initial values for these variables. The
client-specific variables MAY be configured on the server and MAY be
delivered to the client through the ``DHCP Retransmission Parameter
Extension''carried in a Reply message. This extension is documented
in the ``extensions document'' [2].
Bound, Carney, Perkins Expires 1 November 2000 [Page 8]
Internet Draft DHCP for IPv6 5 May 2000
______________________________________________________________
|_Parameter__________|Default_|Description____________________|
|_MIN_SOL_DELAY______|1_______|MIN_(secs)_to_delay_1st_mesg___|
|_MAX_SOL_DELAY______|5_______|MAX_(secs)_to_delay_1st_mesg___|
|_ADV_MSG_TIMEOUT____|500_____|SOL_Retrans_timer_(msecs)______|
|_ADV_MSG_MAX________|30______|MAX_timer_value_(secs)_________|
|_SOL_MAX_ATTEMPTS___|-1______|MAX_attempts_(-1_=_infinite)___|
|_REP_MSG_TIMEOUT____|250_____|REQ_Retrans_timer_(msecs)______|
|_REQ_MSG_ATTEMPTS___|10______|MAX_Request_attempts___________|
|_REL_MSG_ATTEMPTS___|5_______|MAX_Release_attempts___________|
|_RECREP_MSG_TIMEOUT_|2000____|Retrans_timer_(msecs)__________|
|_REC_MSG_ATTEMPTS___|10______|Reconfigure_attempts___________|
|_REC_REP_MIN________|5_______|Minimum_pause_interval_(secs)__|
|_REC_REP_MAX________|7200____|Maximum_pause_interval_(secs)__|
|_REC_THRESHOLD______|100_____|%_of_required_clients__________|
|_SRVR_PREF_WAIT_____|2_______|Advertise_Collect_timer_(secs)_|
4. Requirements
The keywords MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD,
SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL, when they appear in this
document, are to be interpreted as described in [3].
This document also makes use of internal conceptual variables
to describe protocol behavior and external variables that an
implementation must allow system administrators to change. The
specific variable names, how their values change, and how their
settings influence protocol behavior are provided to demostrate
protocol behavior. An implementation is not required to have them in
the exact form described here, so long as its external behavior is
consistent with that described in this document.
5. Background
Related work in IPv6 that would best serve an implementor to study
is the IPv6 Specification [6], the IPv6 Addressing Architecture [8],
IPv6 Stateless Address Autoconfiguration [15], IPv6 Neighbor
Discovery Processing [12], and Dynamic Updates to DNS [17]. These
specifications enable DHCP to build upon the IPv6 work to provide
both robust stateful autoconfiguration and autoregistration of DNS
Host Names.
The IPv6 Specification provides the base architecture and design of
IPv6. A key point for DHCP implementors to understand is that IPv6
requires that every link in the Internet have an MTU of 1280 octets
or greater (in IPv4 the requirement is 68 octets). This means that
a UDP packet of 536 octets will always pass through an internetwork
Bound, Carney, Perkins Expires 1 November 2000 [Page 9]
Internet Draft DHCP for IPv6 5 May 2000
(less 40 octets for the IPv6 header), as long as there are no IP
options prior to the UDP header in the packet. But, IPv6 does not
support fragmentation at routers, so that fragmentation takes place
end-to-end between hosts. If a DHCP implementation needs to send a
packet greater than 1500 octets it can either fragment the UDP packet
into fragments of 1500 octets or less, or use Path MTU Discovery [10]
to determine the size of the packet that will traverse a network
path.
DHCP clients use Path MTU discovery when they have an address of
sufficient scope to reach the DHCP server. If a DHCP client does not
have such an address, that client MUST fragment its packets if the
resultant message size is greater than the minimum 1280 octets.
Path MTU Discovery for IPv6 is supported for both UDP and TCP and
can cause end-to-end fragmentation when the PMTU changes for a
destination.
The IPv6 Addressing Architecture specification [8] defines the
address scope that can be used in an IPv6 implementation, and the
various configuration architecture guidelines for network designers
of the IPv6 address space. Two advantages of IPv6 are that support
for multicast is required, and nodes can create link-local addresses
during initialization. This means that a client can immediately use
its link-local address and a well-known multicast address to begin
communications to discover neighbors on the link. For instance, a
client can send a Solicit message and locate a server or relay.
IPv6 Stateless Address Autoconfiguration [15] (Addrconf) specifies
procedures by which a node may autoconfigure addresses based on
router advertisements [12], and the use of a valid lifetime to
support renumbering of addresses on the Internet. In addition the
protocol interaction by which a node begins stateless or stateful
autoconfiguration is specified. The DHCP is one vehicle to perform
stateful autoconfiguration. Compatibility with addrconf is a design
requirement of DHCP (see Section 6).
IPv6 Neighbor Discovery [12] is the node discovery protocol in IPv6
which replaces and enhances functions of ARP [13]. To understand
IPv6 and Addrconf it is strongly recommended that implementors
understand IPv6 Neighbor Discovery.
Dynamic Updates to DNS [17] is a specification that supports the
dynamic update of DNS records for both IPv4 and IPv6. DHCP can use
the dynamic updates to DNS to integrate addresses and name space
to not only support autoconfiguration, but also autoregistration
in IPv6. The security model to be used with DHCPv6 should conform
as closely as possible to the authentication model outlined in
RFC2402 [9].
Bound, Carney, Perkins Expires 1 November 2000 [Page 10]
Internet Draft DHCP for IPv6 5 May 2000
6. Design Goals
- DHCP is a mechanism rather than a policy. Network administrators
set their administrative policies through the configuration
parameters they place upon the DHCP servers in the DHCP domain
they're managing. DHCP is simply used to deliver parameters
according to that policy to each of the DHCP clients within the
domain.
- DHCP is compatible with IPv6 stateless autoconf [15].
- DHCP does not require manual configuration of network parameters
on DHCP clients, except in cases where such configuration is
needed for security reasons. A node configuring itself using
DHCP should require no user intervention.
- DHCP does not require a server on each link. To allow for scale
and economy, DHCP must work across DHCP relays.
- DHCP coexists with statically configured, non-participating nodes
and with existing network protocol implementations.
- DHCP clients can operate on a link without IPv6 routers present.
- DHCP will provide the ability to renumber network(s) when
required by network administrators [4].
- A DHCP client can make multiple, different requests for
configuration parameters when necessary from one or more DHCP
servers at any time. DHCP will provide enough information
to enable a DHCP server to keep track of a DHCP client's
configuration state.
- DHCP will contain the appropriate time out and retransmission
mechanisms to efficiently operate in environments with high
latency and low bandwidth characteristics.
7. Non-Goals
This specification explicitly does not cover the following:
- Specification of a DHCP server to server protocol.
- How a DHCP server stores its DHCP data.
- How to manage a DHCP domain or DHCP server.
Bound, Carney, Perkins Expires 1 November 2000 [Page 11]
Internet Draft DHCP for IPv6 5 May 2000
- How a DHCP relay is configured or what sort of information it may
log.
8. Overview
This section provides a general overview of the interaction
between the functional entities of DHCP. The overview is organized
as a series of questions and answers. Details of DHCP such
as message formats and retransmissions are left to sections 9,
10, 11, 12, 14, 15, and 16.
8.1. How does a node know to use DHCP?
An unconfigured node determines that it is to use DHCP for
configuration of an interface by detecting the presence (or absence)
of routers on the link. If router(s) are present, the node examines
router advertisements to determine if DHCP should be used to
configure the interface. If there are no routers present, then
the node MUST use DHCP to configure the interface. Detail on
this process can be found in neighbor discovery [12] and stateless
autoconfiguration [15].
8.2. How does a client find out about DHCP agents?
The client forms a Solicit message, and multicasts it to the
FF02::1:2(All DHCP Agents) address. Server(s) receiving the Solicit
respond with Advertise message(s). If requested in the client's
Solicit message, the Advertise message(s) can include one or more
subnet prefix extensions [2], informing the client of subnet prefixes
for the networks(s) managed by the server(s) on the client's link.
Now that the client knows the IP address(es) of agents(s) on the
link, it can request configuration parameters from servers.
8.3. What if the client and server(s) are on different links?
Use of DHCP in such environments requires one or more DHCP relays
be set up on the client's link, because a client may only have a
link-local address. Relays pick up the Solicit and Request messages
from the client and forward them to some set of servers within the
DHCP domain. A relay will include one of its own addresses (of
sufficient scope) of the interface on the same link as the client.
The relay also includes the subnet prefix length of that address
in the client's messages. Servers receiving the forwarded traffic
use this information to aid in selecting configuration parameters
appropriate to the client's link. The servers also use the relay's
Bound, Carney, Perkins Expires 1 November 2000 [Page 12]
Internet Draft DHCP for IPv6 5 May 2000
address as the destination to forward client-destined messages
for final delivery by the relay. Relays forward client messages
to servers using some combination of the FF05::1:3(All Servers)
site-local multicast address, some other (perhaps a combination)
of site-local multicast addresses set up within the DHCP domain to
include the servers in that domain, or a list of unicast addresses
for servers. The network administrator makes relay configuration
decisions based upon the topological requirements (scope) of the
DHCP domain they are managing. Note that if the DHCP domain spans
more than the site-local scope, then the relays MUST be configured
with global addresses for the client's link so as to be reachable by
servers outside the relays' site-local environment.
8.4. How does a client request configuration parameters from servers?
To request configuration parameters, the client forms a Request
message, and sends it to the server either directly (client has an
address of sufficient scope) or indirectly (through the on-link
relay). The client MAY include a Extension Request Extension [2]
along with other extensions to request specific information from the
server. Note that the client MAY form multiple Request messages
and send each of them to different servers to request potentially
different information (perhaps based upon what was advertised) in
order to satisfy its needs. As a client's needs may change over time
(perhaps based upon an application's requirements), the client may
form additional Request messages to request additional information as
it is needed.
The server(s) respond with Reply messages containing the requested
configuration parameters, which can include status information
regarding the information requested by the client. The Reply MAY
also include additional information, such as a reconfiguration event
multicast group for the client to join to monitor reconfiguration
events, as described in section 8.8.
The receipt of a Reply from a server concludes the basic
request/reply transaction of the protocol.
8.5. What are releasable resources, and when are they used?
A releasable resource is configuration information leased to a client
by a server for some finite period of time. When negotiating for a
releasable resource, the client and server agree upon a finite period
of time the client may use the resource. The client MAY request a
renewal of the lease on the resource at any time. The length of time
of the lease (and whether it is renewable) are server-based policy
tunables. The client MUST stop using the resource when the lease on
Bound, Carney, Perkins Expires 1 November 2000 [Page 13]
Internet Draft DHCP for IPv6 5 May 2000
the resource expires. The server MUST NOT reallocate an assigned
resource before either its lease expires or the client releases the
resource.
See the ``extensions document'' [2] for more information about
releasable resources.
8.6. Can a client release its releasable resources before the lease
expires?
A client forms a Release message, including extensions carrying the
resource(s) to be released. The client sends the Release to the
server which leased the resource(s) to the client initially. If that
server cannot be reached after a certain number of attempts (see
section 3.5), the client can abandon the Release attempt. In this
case, the resource(s) will be reclaimed by the server(s) when the
client's lease(s) expire.
8.7. What if the client determines its releasable resource is already
being used by another client?
If the client determines through a releasable resource-specific
manner that the resource it was assigned by the server is already
in use by another client, the client will form a Release message,
including the extension carrying the in-use resource. The
extension's status field MUST be set to the extension-specific value
reflecting the ``in use'' status of the resource.
For example, if the releasable resource is an IP address, the client
uses Duplicate Address Detection (DAD) to verify that the IP address
is not in use. If the client determines that the IP address is
already in use, it forms a Release message including the IP address
extension containing the appropriate status value and sends it to the
server. See the ``extensions document''for details on the IP address
extension. [2].
8.8. How are clients notified of server configuration changes?
There are two possibilities. Either the clients discover the new
information when they revisit the server(s) to request additional
configuration information / renew the lease on a releasable resource,
or through a server-initiated event known as a reconfigure event.
The reconfiguration feature of DHCP offers network administrators
the opportunity to update configuration information on DHCP clients
whenever necessary. If the information to be updated is not
Bound, Carney, Perkins Expires 1 November 2000 [Page 14]
Internet Draft DHCP for IPv6 5 May 2000
client-specific, the server will form a Reconfigure message and add
the new or changed configuration information to it. The Reconfigure
may be unicast or multicast (to a preassigned multicast address for
this purpose) to one or more client(s) to which the new or updated
information needs to be directed. The client(s) will acknowledge the
receipt of the Reconfigure message by forming a Reconfigure-reply
message and unicasting it to the server. If the configuration
information change is different for each client (e.g. a change in
subnet prefix, perhaps, which would affect the IP address releasable
resource(s)), the server will form a Reconfigure-init message and
unicast / multicast as needed to the client(s). A Reconfigure-init
is a trigger which will cause the client(s) to initiate a standard
Request/Reply exchange with the server in order to acquire the new or
updated resources.
9. Message Formats
All reserved fields in a message MUST be transmitted as zeroes and
ignored by the receiver of the message.
9.1. DHCP Solicit Message Format
A client multicasts a DHCP Solicit message to the FF02::1:2(All DHCP
agents) address over the interface to be configured to locate one or
more servers which are configured to provide configuration parameters
to nodes on the client's link.
Unless otherwise noted, the value of all fields are set by the
client.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| msg-type = 1 |C|P| reserved | prefix-len | solicit-ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| client's link-local address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| relay-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
C If set, the client requests that all servers receiving
the message deallocate the releasable resources (e.g.
IP addresses) associated with the client's binding.
Bound, Carney, Perkins Expires 1 November 2000 [Page 15]
Internet Draft DHCP for IPv6 5 May 2000
P If set, the client requests that all servers receiving
the message SHOULD return a list of subnet prefix
extensions identifying the networks on the client's
link that the server(s) are configured to manage.
reserved 0
prefix-len
An unsigned 7 bit number (0-127) non-zero prefix-len is
the number of leftmost bits of the agent's IPv6 address
which make up the subnet prefix. The prefix-len field
is set by the relay if the relay receives the Solicit
message and forwards it to one or more servers.
solicit-ID
An unsigned 9 bit number (0-511) generated by the
client used to identify this Solicit message.
client's link-local address
The IP link-local address of the client interface
through which the client will issue the Solicit
message.
relay-address
Set by the client to be zero. If received by a relay,
set by the relay to the site-local IP address of the
interface on which the relay received the client's
Solicit message. Note that if the DHCP domain crosses
site boundaries, the relay MUST place a globally-scoped
address in this field.
A client MUST send the Solicit message to the All-DHCP-Agents
multicast group (see section 3.1), setting the relay-address to zero.
9.2. DHCP Advertise Message Format
A server sends an Advertise message in response to a client's
Solicit message. The Advertise message notifies the client of the
server's IP address. If the server is so configured by the network
administrator and the client requests it through the ``P'' bit in
its Solicit message, the server SHOULD add a list of subnet prefix
extensions to the Advertise message to notify the client of the
networks it manages on the client's link.
When the client and server are on different links, the server sends
the Advertise message back through the relay whence the corresponding
Solicit came. The solicit-ID is copied from the client's Solicit
Bound, Carney, Perkins Expires 1 November 2000 [Page 16]
Internet Draft DHCP for IPv6 5 May 2000
Message. The value of all fields in the Advertise message are filled
in by the server and not changed in any way by any intervening relay.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| msg-type = 2 | reserved | solicit-ID | preference |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| client's link-local address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| relay-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| server-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| extensions (variable number and length) ... |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
reserved 0
solicit-ID An unsigned 9 bit number (0-511) used to identify
this Advertise message. Copied from the client's
Solicit message.
preference An octet (unsigned) indicating a server's willingness
to provide service to the client.
client's link-local address
The IP link-local address of the client interface
from which the client issued the Solicit message.
relay-address
The IP address of the relay interface on the same
link as the client. Copied from the client's
Solicit. If the server is on the same link as the
client, then this field MUST be zero.
server-address
The site-local IP address of the server. If the DHCP
domain crosses site boundaries, then this address
MUST be globally-scoped.
extensions See the ``extensions document'' for details [2].
See Sections 14.4 and 15.3 for information about how clients and
servers handle the preference field.
Bound, Carney, Perkins Expires 1 November 2000 [Page 17]
Internet Draft DHCP for IPv6 5 May 2000
9.3. DHCP Request Message Format
A client sends a Request message to request configuration parameters
from a server. It MAY append appropriate extensions [2].
When a client reboots, it often does not have a valid IP address of
sufficient scope for the server to communicate with the client. In
such cases, the client MUST NOT unicast the message to the server
because the server could not return a response to the client. The
client MUST send the message to the server indirectly, by using the
on-link relay. The client MUST fill in the relay address field with
the on-link relay's IP address.
If the Request message is being formed in response to a
Reconfigure-init message from the server, then the transaction ID
used must be copied from the Reconfigure-init.
All fields in the DHCP Request message are entered by the client.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| msg-type = 3 |C|R| reserved | transaction-ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| client's link-local address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| relay-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| server-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| extensions (variable number and length) .... |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
C If set, the client requests the server to remove
all releasable resources associated with the client
binding, except those releasable resources provided as
extensions.
R If set, the client has rebooted and requests that the
server clear any transaction-ID cache entries for the
client.
reserved 0
Bound, Carney, Perkins Expires 1 November 2000 [Page 18]
Internet Draft DHCP for IPv6 5 May 2000
transaction-ID
An unsigned integer identifier used to identify this
request.
client's link-local address
The link-local address of the client interface from
which the client will issue the Request message.
relay-address
The IP address of a relay's interface, copied from an
Advertise message. If the server is on the same link
as the client, then this field MUST BE zero.
server-address
The IP address of the server to which the the client's
Request message is directed, copied from an Advertise
message.
extensions
See the ``extensions document'' [2].
A DHCP client selects the transaction-ID from the range of
1024--65535 used to identify its Request. In contrast, a
transaction-ID from the range of 0--1023is selected by a DHCP server
to identify a Reconfigure-init. In the latter case, the transaction
ID from the Reconfigure-init is copied by the client into its Request
message.
When the client sets the `C' bit and adds extensions documenting
the releasable resources the client wishes to keep, the server is
expected to deallocate all other releasable resources not listed.
The server SHOULD examine the included extensions to check whether
the client is still authorized to use them.
9.4. DHCP Reply Message Format
A server sends a Reply message in response to a client's Request
message or Release message.
If a Request message is received which contains a non-zero relay
address field, then the client could not unicast the Request message
to the server and thus had to use a on-link relay. In that case, the
server unicasts the Reply message to the relay address found in the
Request message.
If a Release message is received which contains a non-zero relay
address field, then the client will not have an IP address of
sufficient scope after the Release to receive the Reply message. In
Bound, Carney, Perkins Expires 1 November 2000 [Page 19]
Internet Draft DHCP for IPv6 5 May 2000
this case, the server unicasts the Reply message to the relay address
found in the Release message.
All the fields in the DHCP Reply message are set by the DHCP server.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| msg-type = 4 |R| status | transaction-ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| client's link-local address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| relay-address (if present) |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| extensions (variable number and length) .... |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
R If set, the ``relay-address'' field is present.
status
This 7-bit field contains one of the values in the
errors table in section 3.4.
transaction-ID
Copied from the client's Request or Release.
client's link-local address
Copied from the client's Request or Release message.
relay-address
The IP address of a relay's interface, copied from the
Request or Release message. If the server is on the
same link as the client, then the ``R'' bit is not set
and this field is not present.
extensions
See the ``extensions document'' [2].
9.5. DHCP Release Message Format
A client sends a Release message to a server when it wishes to return
one or more releasable resources to the server which allocated
them. This can occur either because the client no longer needs the
resource(s) or the client has determined through a resource-specific
manner that the resource(s) are already in use by different
Bound, Carney, Perkins Expires 1 November 2000 [Page 20]
Internet Draft DHCP for IPv6 5 May 2000
client(s). The client communicates the reason for the premature
release of the resource in the status field of the resource's
extension. See ``extensions document'' [2] for more details.
When a client sends a Release message, it needs to have a valid IP
address with sufficient scope to allow access by the target server.
If such an address is not available, a relay is used. Only those
releasable resources identified by extensions are released. If no
extensions are included in the Release message, then all releasable
resources associated with the client's binding are to be released.
The values of all fields of the Release message are set by the
client. The DHCP server acknowledges the Release message by sending
a Reply message.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| msg-type = 5 |R| reserved | transaction-ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| client's link-local address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| server-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| X-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| extensions (variable number and length) .... |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
R If set, the ``X-address'' field contains the address of
relay. If not set, the ``X-address'' field contains a
non-local scope client address.
reserved 0
transaction-ID
An unsigned integer identifier used to identify this
Release message.
client's link-local address
The client's link-local address for the interface
from which the client issued the Release message (and
to which the releasable resources are bound at the
server).
Bound, Carney, Perkins Expires 1 November 2000 [Page 21]
Internet Draft DHCP for IPv6 5 May 2000
server-address
The IP address of the server which allocated the
resource.
X-address
If the ``R'' bit is set, the ``X-address'' field
contains the IP address of the relay interface on the
same link as the client. If the ``R'' bit is not set,
this field contains a non-link-local IP address of the
client interface from which the the client issued the
Release message.
extensions See the ``extensions document'' [2].
A client selects the transaction-ID from the range of
1024--65535 used to identify the Release message.
A client MUST NOT specify an IP address in the client-address field
that it is releasing in the extensions field.
9.6. DHCP Reconfigure Message Format
A server sends a Reconfigure message when it wishes to inform one or
more clients of new or updated values for configuration parameters.
The new configuration parameters are carried in the extensions
portion of the Reconfigure message. Note that a Reconfigure message
MUST NOT carry releasable resource extensions.
Reconfigure messages can ONLY be sent to clients which have
established an IP address of sufficient scope as to be directly
reachable by the server.
Clients acknowledge Reconfigure messages with Reconfigure-reply
messages.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| msg-type = 6 | reserved | transaction-ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| server-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| extensions (variable number and length) .... |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
reserved 0
Bound, Carney, Perkins Expires 1 November 2000 [Page 22]
Internet Draft DHCP for IPv6 5 May 2000
transaction-ID
An unsigned integer identifier in the range of
0--1023 chosen by the server to identify this
Reconfigure message.
server-address
The IP address of the DHCP server issuing the
Reconfigure message. MUST be of sufficient scope to be
reachable by all clients.
extensions
See the ``extensions document'' [2].
9.7. DHCP Reconfigure-reply Message Format
A client sends a Reconfigure-reply message to acknowledge receipt of
a Reconfigure message from a server.
A Reconfigure-reply message can only be sent if the client has an IP
address of sufficient scope to contact the server. No interaction
with a relay is possible.
All fields in the DHCP Reconfigure-reply message are entered by the
client.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| msg-type = 7 |r| status | transaction-ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| client's link-local address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| server-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
r reserved (0)
status
This 7-bit field contains one of the values from the
errors table in section 3.4.
transaction-ID
An unsigned integer identifier copied from the server's
Reconfigure message.
Bound, Carney, Perkins Expires 1 November 2000 [Page 23]
Internet Draft DHCP for IPv6 5 May 2000
client's link-local address
The client's link-local address for the interface from
which the client issued the Reconfigure-reply message.
server-address
Copied from the Reconfigure message.
9.8. DHCP Reconfigure-init Message Format
A server sends a Reconfigure-init message when it wishes to notify
one or more clients of new or updated values for configuration
parameters available on the server.
Reconfigure-init messages can ONLY be sent to clients which have
established an IP address of sufficient scope as to be directly
reachable by the server.
A ``Reconfigure-init'' serves as a trigger which will cause the
clients to initiate a Request/Reply exchange with the server in order
to receive the new information.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| msg-type = 8 | reserved | transaction-ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| server-address |
| (16 octets) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| extensions (variable number and length) .... |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
reserved 0
transaction-ID
An unsigned integer identifier in the range of
0--1023 chosen by the server to identify this
Reconfigure-init message.
server-address
The IP address of the DHCP server issuing the
Reconfigure-init message. MUST be of sufficient scope
to be reachable by all clients.
extensions SHOULD only include an ERE and/or authentication
extensions. No configuration information SHOULD be
Bound, Carney, Perkins Expires 1 November 2000 [Page 24]
Internet Draft DHCP for IPv6 5 May 2000
included. See the ``extensions document'' [2] for more
information about extensions.
10. DHCP Server Solicitation and Subnet Prefix Discovery
This section describes how a client locates agents (relays and
servers) and how it can learn about the networks on its link that are
managed by these servers. The behavior of client, server, and relay
implementations is discussed, along with the messages they use.
10.1. Solicit Message Validation
Clients MUST silently discard any received Solicit messages.
Agents MUST discard any received Solicit messages if the ``client's
link-local address'' field does not contain a valid link-local
address.
Servers MUST discard each received Solicit message which meet the
following criteria:
o The ``relay-address'' field does not contain an address of
sufficient scope that is reachable by the server.
o The ``relay-address'' field is non-zero, but prefix-len is zero.
An error message MAY be logged by the agent. The logging of
such messages SHOULD be controlled by an agent implementation
configuration flag.
10.2. Advertise Message Validation
Servers MUST silently discard any received Advertise messages.
Clients MUST discard any Advertise messages that meet any of the
following criteria:
o The ``Solicit-ID'' field value does not match the value the
client used in its Solicit message.
o The ``client's link-local address'' field value does not match
the link-local address of the interface upon which the client
sent the Solicit message.
Relays MUST discard any Advertise messages that meet any of the
following criteria:
Bound, Carney, Perkins Expires 1 November 2000 [Page 25]
Internet Draft DHCP for IPv6 5 May 2000
o The ``relay-address'' field does not contain the relay's address
on the same link as the client.
o The ``client's link-local address'' field does not contain a
valid link-local address.
10.3. Client Behavior
Clients use the Solicit message primarily to discover DHCP servers
configured to serve networks on the link containing the client.
Optionally, the client MAY set the ``P'' bit which has the effect
of requesting that the server return subnet prefix extensions
identifying the networks on the client's link the server is
configured to manage.
10.3.1. Creation and sending of the Solicit message
When creating a Solicit message, the client SHOULD start out with
a buffer initialized with zeroed octets. The client sets the
``msg-type'' field to 1, and places the link-local address of the
interface it wishes to configure in the link-local address field.
If the client is prepared to process multiple Advertise messages
in response to its Solicit message, the client will set the
Solicit-ID field to 1. Every time the client initiates a new server
solicitation attempt (not a retransmission), the client increments
the Solicit-ID by one. If the 9-bit field rolls over to 0, then the
client sets the Solicit-ID to 1. A client which will only accept
the first Advertise message it receives leaves the Solicit-ID field
initialized to zero.
The ``C'' bit of the Solicit message is set by the client when the
client has no cached knowledge of previous DHCP configuration for the
interface. Setting this bit requests that the server release any
information assigned to the client for the networks on the client's
link.
If the client desires to learn of the networks managed by DHCP on
the link its interface is attached to, it sets the ``P'' bit in the
Solicit message.
The client transmits the Solicit message to the FF02::1:2 (All DHCP
Agents) multicast address, destination port 547. The source port
selection can be arbitrary, although it SHOULD be possible using a
client configuration facility to set a specific source port value.
Bound, Carney, Perkins Expires 1 November 2000 [Page 26]
Internet Draft DHCP for IPv6 5 May 2000
10.3.2. Time out and retransmission of Solicit Messages
The client's first Solicit message on the interface MUST be delayed
by a random amount of time between the interval of MIN_SOL_DELAY and
MAX_SOL_DELAY. This random delay desynchronizes clients which start
at the same time (e.g., after a power outage).
The client waits ADV_MSG_TIMEOUT, collecting Advertise messages.
If no Advertise messages are received, the client retransmits
the Solicit, and doubles the ADV_MSG_TIMEOUT value. This process
continues until either one or more Advertise messages are received or
ADV_MSG_TIMEOUT reaches the ADV_MSG_MAX value. Thereafter, Solicits
are retransmitted every ADV_MSG_MAX until SOL_MAX_ATTEMPTS have been
made, at which time the client stops trying to DHCP configure the
interface. An event external to DHCP is required to restart the DHCP
configuration process.
Default and initial values for MIN_SOL_DELAY, MAX_SOL_DELAY,
ADV_MSG_TIMEOUT, AND ADV_MSG_MAX are documented in section 3.5.
10.3.3. Receipt of Advertise messages
Upon receipt of one or more validated Advertise messages, the client
selects one or more Advertise messages based upon the following
criteria.
- Those Advertise messages with the highest server preference
value (see section 14.4) are preferred over all other Advertise
messages.
- Within a group of Advertise messages with the same server
preference value, a client MAY select those servers whose
Advertise messages advertise information of interest to
the client. For example, one server may be advertising the
availability of IP addresses on networks which have an address
scope of interest to the client.
Once a client has selected Advertise message(s), the client will
typically store information about each server, such as relay address
and prefix length, server preference value, networks advertised,
when the advertisement was received, and so on. Depending on the
requirements of the client's invoking user, the client MAY initiate a
configuration exchange with the server(s) immediately, or MAY defer
this exchange until later.
Bound, Carney, Perkins Expires 1 November 2000 [Page 27]
Internet Draft DHCP for IPv6 5 May 2000
10.4. Relay Behavior
For this discussion, the Relay is assumed to have been configured
with some list of server destination addresses, which may be unicast,
the FF05::1:3 (All DHCP Servers) multicast address, or some other
multicast address selected by the network administrator.
10.4.1. Relaying of Solicit messages
When a Relay receives a valid Solicit message, it places the IP
address of the interface upon which it received the Solicit message
in the ``relay-address'' field of the Solicit. The Relay also places
the number of bits of that make up the subnet prefix for this address
in the ``prefix-len'' field of the Solicit.
The Relay then forwards this Solicit to the list of server
destination addresses that it has been configured with.
10.4.2. Relaying of Advertise messages
When a Relay receives a valid Advertise message, it unicasts the
message to the link-local address found in the ``client's link-local
address'' field by way of the appropriate network interface.
10.5. Server Behavior
For this discussion, the Server is assumed to have been configured in
an implementation specific manner. This configuration is assumed to
contain all network topology information for the DHCP domain, as well
as any necessary authentication information.
10.5.1. Receipt of Solicit messages
Upon the receipt of a valid Solicit message, the server first
identifies the client's location within the DHCP domain. If the
``relay-address'' and / or ``prefix-len'' fields of the Solicit are
zeroed, then the client is attached to the same link as the server.
If these fields are non-zero, then the client exists on the same link
as the network identified by these two fields.
If administrative policy permits the server to respond to a client on
that link, the server will generate and send an Advertise message to
the client.
Bound, Carney, Perkins Expires 1 November 2000 [Page 28]
Internet Draft DHCP for IPv6 5 May 2000
10.5.2. Creation and sending of Advertise messages
When creating an Advertise message, the server SHOULD start out
with a buffer initialized with zeroed octets. The server sets the
``msg-type'' field to 2 and copies the values of the following fields
from the client's Solicit to the Advertise message:
o solicit-ID
o client's link-local address
o relay-address
The server places one of its IP addresses (determined through
administrator setting) in the ``server-address'' field of the
Advertise message. The server initializes the ``preference''
field from its configuration information. See section 15.3 for a
description of server preference.
If the client requests subnet prefix extensions (by setting the ``P''
bit in its Solicit) and the server implements and is configured to
provide prefix extensions, the server will generate and insert a
subnet prefix extension for each network on the client's link it is
configured to manage.
If the ``relay-address'' field of the Advertise message is zero, then
the server unicasts the Advertise message directly to the client
using the ``client's link-local address'' field value as destination
address. If the ``relay-address'' field is non-zero, then the server
unicasts the Advertise message directly to the relay using the
``relay-address'' field value as the destination address.
11. DHCP Client-Initiated Configuration Exchange
A client initiates a configuration exchange with one or more servers
it has found through DHCP server solicitation whenever requested to
do so by the application layer in order to acquire configuration
information of interest.
11.1. Request Message Validation
Clients MUST silently discard any received Request messages.
Agents MUST discard any Request messages in which the ``client's
link-local address'' field does not contain a valid link-local
address.
Bound, Carney, Perkins Expires 1 November 2000 [Page 29]
Internet Draft DHCP for IPv6 5 May 2000
Relays MUST discard any received Request messages in which the
``relay-address'' field value does not match any of the relay's
addresses.
Servers MUST discard any received Request message which meets any of
the following criteria:
o The ``server-address'' field value does not match any of the
server's addresses.
o If the ``relay-address'' field is set, and that field's value
does not contain an address of sufficient scope as to be
reachable by the server.
o The ``extensions'' field contains an authentication extension,
and the server cannot successfully authenticate the client.
11.2. Reply Message Validation
Servers MUST silently discard any received Reply messages.
Clients MUST discard any Reply message that meets any of the
following criteria:
o The ``transaction-ID'' field value does not match the value the
client used in its Request or Release message.
o The ``client's link-local address'' field value does not match
the link-local address of the interface upon which the client
sent in its Request or Release message.
o The Reply message contains an authentication extension, and the
client's attempt to authenticate the message fails.
Relays MUST discard any Reply message that meets any of the following
criteria:
o The ``R'' bit isn't set.
o The ``relay-address'' field value does not contain the relay's
address on the same link as the client.
o The ``client's link-local address'' field value does not contain
a valid link-local address.
Bound, Carney, Perkins Expires 1 November 2000 [Page 30]
Internet Draft DHCP for IPv6 5 May 2000
11.3. Release Message Validation
Clients MUST silently discard any received Release messages.
Agents MUST discard any Release message that meets any of the
following criteria:
o The ``transaction-ID'' field contains a value not in the
1024--65535 range.
o The ``client's link-local address'' field does not contain a
valid link-local address.
Relays MUST discard any received Release message that meets any of
the following criteria:
o The ``R'' bit is not set.
o The ``X-address'' field value does not match any of the relay's
addresses.
Servers MUST discard any received Release message which meets any of
the following criteria:
o The ``X-address'' field does not contain an address of sufficient
scope as to be reachable by the server.
o The ``extensions'' field contains an authentication extension,
and the server cannot successfully authenticate the client.
11.4. Client Behavior
A client will generate one or more Request messages when prompted by
the application layer in order to acquire configuration information.
A client may initiate such an exchange automatically in order to
acquire the necessary network parameters to communicate with nodes
off-link. The client uses the server and relay address information
from previous Advertise message(s) for use in delivering Request
message(s). Note that a client may request configuration information
from one or more servers at any time.
A client uses the Release message in the management of releasable
resources when:
o The client has determined through a resource-specific manner
that the resource assigned by the server is already in use by a
different client.
Bound, Carney, Perkins Expires 1 November 2000 [Page 31]
Internet Draft DHCP for IPv6 5 May 2000
o The client has been instructed to release the resource prior to
the lease expiration time since it is no longer needed.
11.4.1. Creation and sending of Request messages
When creating a Request message, the client SHOULD start out with
a buffer initialized with zeroed octets. The client sets the
``msg-type'' field to 3, and places the link-local address of the
interface it wishes to associate with the configuration information
with in the ``client's link-local address'' field.
Unless the Request message is created in response to a
Reconfigure-init message, the client generates a transaction
ID in the range of 1024--65535 and inserts this value in the
``transaction-ID'' field.
The client places the address of the destination server in the
``server-address'' field.
If the client is not on the same link as the destination
server, the client places the appropriate relay's address in the
``relay-address'' field.
If the client is acquiring configuration information on the interface
for the first time, the client SHOULD set the ``C'' bit in the
header. How the client determines if this is the first configuration
attempt on the interface is implementation-specific. A client may
implement a cache of configuration information on a per-interface
basis; if that cache does not exist, that client would set the
``C'' bit. Clients which do not implement caching of per-interface
configuration information MUST always set the ``C'', and include
any extensions carrying releasable resources received from earlier
configuration exchanges in the extensions field of the Request.
If the client has determined through an implementation-specific
manner that the client implementation itself has restarted, it MUST
set the ``R'' bit in the header. After the first successful exchange
with the server, the client MUST NOT set the ``R'' bit in subsequent
Request messages.
Client considerations for extensions are now considered (see the
``extensions document'', [2] for more details).
If the client already has an IP address of sufficient scope to
directly reach the server, then the client SHOULD unicast the Request
to the server. Otherwise, if the server is off-link, the client
unicasts the Request message to the appropriate relay.
Bound, Carney, Perkins Expires 1 November 2000 [Page 32]
Internet Draft DHCP for IPv6 5 May 2000
11.4.2. Time out and retransmission of Request Messages
The client waits REP_MSG_TIMEOUT milliseconds, collecting
Reply messages. If no Reply messages are received, the client
retransmits the Request with the same transaction-ID, and doubles
the REP_MSG_TIMEOUT value, and waits again. The client continues
this process until a Reply is received or REQUEST_MSG_ATTEMPTS
unsuccessful attempts have been made, at which time the client MUST
abort the configuration attempt. The client SHOULD report the abort
status to the application layer.
Default and initial values for REP_MSG_TIMEOUT and REQ_MSG_ATTEMPTS
are documented in section 3.5.
11.4.3. Receipt of Reply message in response to a Request
Upon the receipt of a valid Reply message, the client extracts the
configuration information contained in the Reply. If the ``status''
field contains a non-zero value, the client reports the error status
to the application layer.
If the extensions field contains one or more ``Reconfigure Multicast
Address'' extensions (see ``extensions document'', ``Reconfigure
Multicast Address Extension'' section [2]), the client MUST join
these multicast groups, and MUST monitor the UDP 546 port for
Reconfigure or Reconfigure-init messages on the networks configured
by DHCP.
If the configuration information returned in the Reply contains
releasable resources, then the client MUST take over lease management
of the resource. A client MUST NOT request releasable resources
unless it is prepared to appropriately manage the resource lease.
11.4.4. Creation and sending of Release messages
When creating a Release message, the client SHOULD start out with
a buffer initialized with zeroed octets. The client sets the
``msg-type'' field to 5, and places the link-local address of the
interface the configuration information it wishes to release is
associated with in the ``client's link-local address'' field.
The client generates a transaction ID in the range of
1024--65535 and inserts this value in the ``transaction-ID''
field.
The client includes extensions containing the releasable resources it
is releasing in the ``extensions'' field. The appropriate ``status''
Bound, Carney, Perkins Expires 1 November 2000 [Page 33]
Internet Draft DHCP for IPv6 5 May 2000
field in the extensions MUST be set to indicate the reason for the
release.
The client places the IP address of the server who allocated the
resource(s) in the ``server-address'' field.
If the client will have an appropriately scoped IP address after the
release transaction is completed, the client clears the ``R'' bit
and places this address in the ``X-address'' field. If the client
will not have an appropriately scoped IP address after the release
transaction is completed, the client sets the ``R'' bit and places
the address of the appropriate relay in the ``X-address'' field.
If the client is configured to use authentication, the client
generates the appropriate authentication extension, and adds this
extension to the ``extensions'' field. Note that the authentication
extension MUST be the last extension in the ``extensions''
field. See the ``extension document'' for more details about the
authentication extension [2].
If the ``R'' bit is set, then the client MUST unicast the Release
to the relay indicated in the ``X-address'' field. Otherwise, the
client unicasts the Release message directly to the server indicated
in the ``server-address'' field.
11.4.5. Time out and retransmission of Release Messages
The client waits REP_MSG_TIMEOUT milliseconds, collecting Reply
messages. If no Reply messages are received, the client retransmits
the Release, and doubles the REP_MSG_TIMEOUT value, and waits again.
The client continues this process until a Reply is received or
REL_MSG_ATTEMPTS unsuccessful attempts have been made, at which
time the client SHOULD abort the release attempt. The client
SHOULD return the abort status to the application, if an application
initiated the release.
Default and initial values for REP_MSG_TIMEOUT and REL_MSG_ATTEMPTS
are documented in section 3.5.
Note that if the client fails to release the resource, the resource
will be reclaimed by the server when the lease associated with it
expires.
Bound, Carney, Perkins Expires 1 November 2000 [Page 34]
Internet Draft DHCP for IPv6 5 May 2000
11.4.6. Receipt of Reply message in response to a Release
Upon receipt of a valid Reply message, the client can consider the
Release event successful, and SHOULD return the successful status to
the application layer, if an application initiated the release.
11.5. Relay Behavior
11.5.1. Relaying of Request or Release messages
When a Relay receives a valid Request or Release message, it forwards
it to the IP address found in the ``server-address'' field of the
message.
11.6. Server Behavior
For this discussion, the Server is assumed to have been configured
in an implementation specific manner with configuration of interest
to clients. Such configuration information MAY contain releasable
resources such as IP addresses.
11.6.1. Receipt of Request messages
Upon the receipt of a valid Request message from a client the server
can respond to, (implementation-specific administrative policy
satisfied) the server scans the extensions field.
If the client has set the ``C'' bit, the server MUST release all
releasable resources currently associated with the client's binding
that do not appear in the ``extensions'' field.
If the client has set the ``R'' bit, the server MUST delete any
transaction-ID cache entries it is maintaining for this client, if
the server implements such a cache.
Server considerations for extensions are now evaluated (see the
``extensions document'', [2] for more details).
If the configuration information to be returned to the client
includes releasable resources, the server checks if a binding
already exists for the client. If so, the server examines the
data records within the binding to determine if the client's
Request is a retransmission of an earlier Request or a new Request.
Releasable resource identifiers are stored within the binding with
the transaction-ID used by the client to request the resource's
assignment. If the transaction-ID's match, this is a retransmission
Bound, Carney, Perkins Expires 1 November 2000 [Page 35]
Internet Draft DHCP for IPv6 5 May 2000
and the server simply return the contents of the client's binding
which satisfy its request. If the transaction-ID's do not match,
the server records the additional resources it is assigning in the
existing binding with the new Request's transaction-ID.
If the client does not have an existing binding, the server creates a
binding for the client and records the resources it is assigning in
this binding along with the transaction-ID from the client's Request.
The server then constructs a Reply message and sends it to the
client.
11.6.2. Receipt of Release messages
Upon the receipt of a valid Release message, the server performs a
lookup to find the client's binding. If the binding is found, the
server examines the binding to see if the resource(s) identified by
the client in the Release message's extensions field are in fact
assigned to the client. If they are, the server deletes these
resources from the client's binding, making them available to other
clients.
The server then generates a Reply message. If a binding was
found and the resources presented to the server were deleted from
the client's binding, the server sets the ``status'' field to
``Success''. If no binding is found, the server sets the ``status''
field to ``NoBinding''(section 3.4).
11.6.3. Creation and sending of Reply messages
When creating a Reply message, the server SHOULD start out with
a buffer initialized with zeroed octets. The server sets the
``msg-type'' field to 4 and copies the values of the following fields
from the client's Request or Release to the Reply message:
o transaction-ID
o client's link-local address
o If the client's message is a Request with a non-zero
``relay-address'' field value, the server sets the ``R'' bit in
the Reply and copies the ``relay-address'' field value from the
Request to the Reply. If the client's message is a Release with
the ``R'' bit set, the server sets the ``R'' bit in the Reply and
sets the ``relay-agent'' field to the contents of the Release's
X-address field.
Bound, Carney, Perkins Expires 1 November 2000 [Page 36]
Internet Draft DHCP for IPv6 5 May 2000
The server sets the ``status'' field appropriately (see the table
in section 3.4) based upon the results of processing the client's
request.
If configured to do so, a server will include ``Reconfigure Multicast
Address'' extensions (see ``extensions document'', ``Reconfigure
Multicast Address Extension'' [2]), in Reply messages sent in
response to a Request, informing the client of one or more multicast
groups it should join to facilitate the receipt of Reconfigure or
Reconfigure-init messages.
If the DHCP domain is using authentication, the server will generate
an authentication extension with the appropriate settings and add
that extension as the last extension in the ``extensions'' field of
the Reply message.
If the ``relay-address'' field of the Reply message is zero, then the
server unicasts the Reply directly to the client using the ``client's
link-local address'' field value as destination address. If the
``relay-address'' field is non-zero, then the server unicasts the
Reply directly to the relay using the ``relay-address'' field value
as the destination address.
If the server implements a transaction-ID cache, the server would add
an entry for the client to this cache.
12. DHCP Server-Initiated Configuration Exchange
A server initiates a configuration exchange on behalf of the
administrator of the DHCP domain. An administrator may initiate such
an exchange when new networks are added to the domain or existing
networks are to be renumbered. Other examples include changes in
the location of directory servers, addition of new services such as
printing, and availability of new software (system or application).
12.1. Reconfigure Message Validation
Agents MUST silently discard any received Reconfigure messages.
Clients MUST discard any Reconfigure message that meets any of the
following criteria:
o The ``transaction-ID'' field value is not within
the 0--1023 range.
o The Reconfigure message contains an authentication extension, and
the client's attempt to authenticate the message fails.
Bound, Carney, Perkins Expires 1 November 2000 [Page 37]
Internet Draft DHCP for IPv6 5 May 2000
12.2. Reconfigure-reply Message Validation
Clients and Relays MUST silently discard any received
Reconfigure-reply messages.
Servers MUST discard any Reconfigure-reply message that meets any of
the following criteria:
o The ``transaction-ID'' field value is not that same value the
server used in its Reconfigure message.
o The ``server-address'' field value does not match the value the
server placed in its Reconfigure message.
12.3. Reconfigure-init Message Validation
Agents MUST silently discard any received Reconfigure-init messages.
Clients MUST discard any Reconfigure-init messages that meets any of
the following criteria:
o The ``transaction-ID'' field value is not within
the 0--1023 range.
o The Reconfigure-init message contains an authentication
extension, and the client's attempt to authenticate the message
fails.
12.4. Server Behavior
For this discussion, the server is assumed to have a
implementation-specific interface by which an administrator
may initiate a reconfiguration event with some set of clients.
There are two methods of initiating a reconfiguration event. Each
has its advantages:
Reconfigure with payload
This method uses the Reconfigure message. Items
to be changed are included as extensions in the
``extensions'' field. This method MUST NOT be used
to reconfigure releasable resources. Examples of
information which can be reconfigured using this
method are DNS domain and servers, NTP servers, other
name service parameters. The server generates and
sends the Reconfigure message; clients respond with
Reconfigure-reply messages.
Bound, Carney, Perkins Expires 1 November 2000 [Page 38]
Internet Draft DHCP for IPv6 5 May 2000
Reconfigure Trigger
This method uses the Reconfigure-init message. When
a client receives a Reconfigure-init message, it
initiates a Request/Reply exchange with the server.
Any kind of resource can be reconfigured using this
method, including releasable resources. An example
of an releasable resource is an IP address.
A server can send Reconfigure and Reconfigure-init messages only to
those clients who have an address of sufficient scope to be reachable
by the server. Thus, those clients who have not requested an IP
address and are off-link cannot be reconfigured by the server.
Before initiating a reconfigure process, the server SHOULD be
configured with a REC_THRESHOLD threshold value which represents
the percentage of clients successfully reconfigured before the
reconfigure process is considered a success. See section 3.5 for the
default setting of REC_THRESHOLD. Note that the server MUST be able
to determine the set of clients that should receive the reconfigure,
in order to determine when the reconfigure process is complete.
12.4.1. Creation and sending of Reconfigure messages
When creating a Reconfigure message, the server SHOULD start out
with a buffer initialized with zeroed octets. The server sets the
``msg-type'' field to 6. The server generates a transaction-ID
from the 0--1023 range and inserts it in the ``transaction-ID''
field. The server places its address (of appropriate scope) in the
``server-address'' field.
The server then generates extensions for the non-releasable resources
to be changed and places them in the ``extensions'' field.
If the DHCP domain is using authentication, the server will generate
an authentication extension with the appropriate settings and add
that extension as the last extension in the ``extensions'' field of
the Reconfigure message.
The server multicasts the Reconfigure message to one or more
Reconfigure Multicast Addresses previously sent as extensions to the
clients. Note that a server MAY unicast Reconfigure message(s) to
specific clients by walking its list of bindings to determine the
unicast address(es) of the clients. Whether or not the Reconfigure
is multicast or unicast is an implementation detail.
A server waits for Reconfigure-reply messages from clients confirming
that they have received the Reconfigure.
Bound, Carney, Perkins Expires 1 November 2000 [Page 39]
Internet Draft DHCP for IPv6 5 May 2000
12.4.2. Time out and retransmission of Reconfigure messages
The server waits RECREP_MSG_TIMEOUT milliseconds, collecting
Reconfigure-reply messages. If all the expected Reconfigure-reply
messages are received, then the reconfigure process is successful.
If some or all of the expected Reconfigure-reply messages are not
received, then the server retransmits the Reconfigure, and doubles
the RECREP_MSG_TIMEOUT value, and waits again. The server continues
this process until all Reconfigure-reply messages are received or
REC_MSG_ATTEMPTS unsuccessful attempts have been made, at which time
the server SHOULD abort the reconfigure process. The server SHOULD
log the result of the reconfigure process.
Default and initial values for RECREP_MSG_TIMEOUT and
REC_MSG_ATTEMPTS are documented in section 3.5.
12.4.3. Receipt of Reconfigure-reply messages
Upon receipt of a valid Reconfigure-reply message, the server
removes that client from the list of clients it is expecting a
Reconfigure-reply message from.
12.4.4. Creation and sending of Reconfigure-init messages
When creating a Reconfigure-init message, the server SHOULD start
out with a buffer initialized with zeroed octets. The server sets
the ``msg-type'' field to 8. The server generates a transaction-ID
from the 0--1023 range and inserts it in the ``transaction-ID''
field. The server places its address (of appropriate scope) in the
``server-address'' field.
The server MAY generate an ERE extension to inform the client of what
information has been changed or new information that has been added.
If the DHCP domain is using authentication, the server will generate
an authentication extension with the appropriate settings and add
that extension as the last extension in the ``extensions'' field of
the Reconfigure-init message.
Typically, the server will not provide more than an ERE and / or
Authentication extension, since it will provide the new configuration
information as part of the Request/Reply transaction triggered by the
Reconfigure-init message.
The server multicasts the Reconfigure-init message to one or more
Reconfigure Multicast Addresses previously sent as extensions
to the clients. Note that a server MAY unicast Reconfigure-init
Bound, Carney, Perkins Expires 1 November 2000 [Page 40]
Internet Draft DHCP for IPv6 5 May 2000
message(s) to specific clients by walking its list of bindings to
determine the unicast address(es) of the clients. Whether or not the
Reconfigure-init is multicast or unicast is an implementation detail.
A server waits for a Request message from each client confirming that
they have received the Reconfigure-init and are thus initiating a
Request/Reply transaction with the server. The server can determine
that a Request message is in response to a Reconfigure-init because
the transaction-ID in the Request will be the same value as was used
in the Reconfigure-init message.
12.4.5. Time out and retransmission of Reconfigure-init messages
The server uses the same algorithm and configuration values for
sending Reconfigure-init messages as it does with Reconfigure
messages. See Section 12.4.2 for this algorithm.
12.4.6. Receipt of Request messages
Upon receipt of a valid Request message with the same transaction-ID
as the Reconfigure-init messages it sent, the server removes that
client from the list of clients it is expecting to initiate a
Request/Reply transaction.
The server generates and sends Reply message(s) to the client as
described in section 11.6.3, including in the ``extension'' field
new values for configuration parameters. If the extensions include
releasable resources, the server will include two extensions for each
resource - one with the original values with the lease times set to
zero, and another with new values and lease times. Note that the
server can terminate the client's ability to use a resource simply by
including only the first extension value.
12.5. Client Behavior
A client MUST always monitor UDP port 546 for Reconfigure and
Reconfigure-init messages on interfaces upon which it has acquired
DHCP parameters. Since the results of a reconfiguration event may
affect application layer programs, the client SHOULD log these
events, and MAY notify these programs of the change through an
implementation-specific interface.
Bound, Carney, Perkins Expires 1 November 2000 [Page 41]
Internet Draft DHCP for IPv6 5 May 2000
12.5.1. Receipt of Reconfigure messages
Upon receipt of a valid Reconfigure message, the client extracts
the configuration parameters contained in the ``extensions''
field, and notifies the application layer that new values for these
parameters are available. The client then generates and sends a
Reconfigure-reply message to the server.
12.5.2. Creation and sending of Reconfigure-reply messages
When creating a Reconfigure-reply message, the client SHOULD start
out with a buffer initialized with zeroed octets. The client sets
the ``msg-type'' field to 7, and places the link-local address of
the interface upon which it received the Reconfigure message in
the ``client's link-local address'' field. The client copies the
values of the following fields from the Reconfigure message to the
Reconfigure-reply message:
o transaction-ID
o server-address
The client sets the ``status'' field appropriately (see the table
in section 3.4) based upon the results of processing the server's
reconfigure-reply.
The client places the address of the destination server in the
``server-address'' field.
If the client is configured to use authentication, the client
generates the appropriate authentication extension, and adds this
extension to the ``extensions'' field. Note that the authentication
extension MUST be the last extension in the ``extensions'' field.
The client delays the sending of the Reconfigure-reply by some
random value selected in the range of REC_REP_MIN and REC_REP_MAX
seconds. This delay helps reduce the load on the server generated by
processing large numbers of Reconfigure-reply messages.
Default and initial values for REC_REP_MIN and REC_REP_MAX are
documented in section 3.5.
The client unicasts the Reconfigure-reply to the address identified
in the ``server-address'' field. Sending the Reconfigure-reply
completes the reconfiguration process for the client.
Bound, Carney, Perkins Expires 1 November 2000 [Page 42]
Internet Draft DHCP for IPv6 5 May 2000
12.5.3. Receipt of Reconfigure-init messages
Upon receipt of a valid Reconfigure-init message, the client
initiates a Request/Reply transaction with the server.
12.5.4. Creation and sending of Request messages
When responding to a Reconfigure-init, the client creates and
sends the Request message in exactly the same manner as outlined in
section 11.4.1 with the following differences:
transaction-ID
The client copies the transaction-ID from the
Reconfigure-init message into the Request message.
Pause before sending Request
The client pauses before sending the Request for
a random value within the range REC_REP_MIN and
REC_REP_MAX seconds, as outlined in section 12.5.2.
12.5.5. Time out and retransmission of Request messages
The client uses the same variables and retransmission algorithm as it
does with Request messages generated as part of a client-initiated
configuration exchange. See section 11.4.2 for details.
12.5.6. Receipt of Reply messages
Upon the receipt of a valid Reply message, the client extracts
the contents of the ``extension'' field, and sets (or resets)
configuration parameters appropriately. If the configuration
parameters changed were requested by the application layer, the
client notifies the application layer of the changes using an
implementation-specific interface. If the resources changed are
releasable, the client makes the appropriate adjustments to its
management of the leases of these resources.
13. Using DHCP for network renumbering
An administrator can use DHCP to renumber links within her DHCP
domain through two techniques, passive renumbering and active
renumbering.
Bound, Carney, Perkins Expires 1 November 2000 [Page 43]
Internet Draft DHCP for IPv6 5 May 2000
13.1. Passive Renumbering
The administrator can configure her servers to return relatively
short preferred and valid lifetimes for the IP addresses she
makes available to clients. When she determines that she'd like
to renumber a network, she configures her servers through an
implementation-specific manner to disallow the extension of the IP
address lifetimes on the original network, and adds the new network
configuration data to the server's database.
The clients on the original network will fail to acquire lifetime
extensions on their IP addresses, and will request and acquire
IP addresses from the new network when the valid lifetime of the
original IP addresses approaches expiration.
When the lifetimes for all of the IP addresses on the original
network expire, the network can be considered renumbered.
13.2. Active Renumbering
The administrator can force the renumbering of networks in her DHCP
domain by using the reconfigure feature of DHCP. She instructs her
servers of the network renumbering through an implementation-specific
interface. The servers in the domain will generate Reconfigure-init
messages, which will cause the clients to initiate a Request/Reply
transaction with the server. The servers will include two IP address
extensions for each IP address being changed. The first will contain
the original IP address, with the preferred and valid lifetimes set
to zero. The second will contain the new IP address, with non-zero
preferred and valid lifetimes.
A server implementation MAY permit the administrator to set the
original IP address lifetimes to some small value greater than zero,
to allow applications running on the client to orderly transfer to
the new network over time.
14. DHCP Client Implementator Notes
This section provides helpful information for the client implementor
regarding their implementations. The text described here is not part
of the protocol, but rather a discussion of implementation features
we feel the implementor should consider during implementation.
Bound, Carney, Perkins Expires 1 November 2000 [Page 44]
Internet Draft DHCP for IPv6 5 May 2000
14.1. Primary Interface
Since configuration parameters acquired through DHCP can be
interface-specific or more general, the client implementor SHOULD
provide a mechanism by which the client implementation can be
configured to specify which interface is the primary interface. The
client SHOULD always query the DHCP data associated with the primary
interface for non-interface specific configuration parameters. An
implementation MAY implement a list of interfaces which would be
scanned in order to satisfy the general request. In either case, the
first interface scanned is considered the primary interface.
By allowing the specification of a primary interface, the client
implementor identifies which interface is authoritative for
non-interface specific parameters, which prevents configuration
information ambiguity within the client implementation.
14.2. Advertise Message and Configuration Parameter Caching
If the hardware the client is running on permits it, the implementor
SHOULD provide a cache for Advertise messages and a cache of
configuration parameters received through DHCP. Providing these
caches prevents unnecessary DHCP traffic and the subsequent load
this generates on the servers. The implementor SHOULD provide a
configuration knob for setting the amount of time the cache(s) are
valid.
14.3. Time out and retransmission variables
Note that the client time out and retransmission variables outlined
in section 3.5 can be configured on the server and sent to the client
through the use of the ``DHCP Retransmission Parameter Extension'',
which is documented in the ``extensions document'' [2]. A client
implementation SHOULD be able to reset these variables using the
values from this extension.
14.4. Server Preference
A client MUST wait for SRVR_PREF_WAIT seconds after sending a DHCP
Solicit message to collect Advertise messages and compare their
preferences (see section 15.3), unless it receives an Advertise
message with a preference of 255. If the client receives an
Advertise message with a preference of 255, then the client MAY act
immediately on that Advertise without waiting for any more additional
Advertise messages.
Bound, Carney, Perkins Expires 1 November 2000 [Page 45]
Internet Draft DHCP for IPv6 5 May 2000
15. DHCP Server Implementator Notes
This section provides helpful information for the server implementor.
15.1. Client Bindings
A server implementation can use the client's link-local address
and the subnet prefix specification from which the client sent its
Request message(s) as an index for finding configuration parameters
assigned to the client. While it isn't critical to keep track
of which clients were given information (resources) that isn't
releasable, it IS critical for the server to keep track of which
client it has assigned releasable resources. The server MUST
include the transaction-ID from the client's Request along with
the releasable resource identifier(s) within the binding. This is
done so that the server can detect whether a client Request is a
retransmission of an earlier Request or an entirely new Request.
The server should periodically scan its bindings for releasable
resources whose leases have expired. When the server finds expired
resource assignments, it MUST delete these assignments, thereby
making these resources available to other clients.
The client bindings MUST be stored in non-volatile storage.
The server implementation should provide policy knobs to control
whether or not the lease on a releasable resource is renewable, and
by how long.
15.2. Reconfigure Considerations
A server implementation MUST provide an interface to the
administrator for initiating reconfigure events.
A server implementation may provide a mechanism for allowing the
specification of how many clients comprise a reconfigure multicast
group. This enables the administrator to control the hit a server
takes when a reconfigure event occurs.
15.3. Server Preference
The server implementation SHOULD allow the setting of a server
preference value by the administrator. The server preference
variable is an unsigned single octet value (0--255), with the lowest
preference being 0 and the highest 255. Clients will choose higher
preference servers over those with lower preference values. If you
Bound, Carney, Perkins Expires 1 November 2000 [Page 46]
Internet Draft DHCP for IPv6 5 May 2000
don't choose to implement this feature in your server, you MUST set
the server preference field to 0 in the Advertise messages generated
by your server.
15.4. Request Message Transaction-ID Cache
In order to improve performance, a server implementation MAY include
an in memory transaction-ID cache. This cache is indexed by client
binding and transaction-ID, and enables the server to quickly
determine whether a Request is a retransmission or a new Request
without the cost of a database lookup. If an implementor chooses to
implement this cache, then they SHOULD provide a configuration knob
to tune the lifetime of the cache entries.
16. DHCP Relay Implementator Notes
A relay implementation SHOULD allow the specification of a list of
destination addresses for Solicit messages. This list MAY contain
any mixture of unicast addresses and multicast addresses.
If a relay receives an ICMP message in response to a DHCP message it
has forwarded, it SHOULD log this event.
17. Open Issues for Working Group Discussion
This section contains some items for discussion by the working group.
17.1. Trade-offs: Optional fields in DHCP messages
You'll notice that the message formats have changed. In particular,
some of the optional fields are now required. This will increase the
size of DHCP messages in some cases, consuming network bandwidth and
memory on the DHCP client (an issue for small devices such as PDAs).
The changes were made for the following reasons:
o Fields that were used most of the time were made required.
o Some fields that were optional were either made required or added
to messages which previously didn't have them. This was done for
robustness reasons (receivers can validate that the message is
for them, and in the case of clients, know which interface the
message is intended for).
o Simplicity.
Bound, Carney, Perkins Expires 1 November 2000 [Page 47]
Internet Draft DHCP for IPv6 5 May 2000
Please look at the messages as they are now defined, and let us know
your opinion.
17.2. Use DHCPv4 authentication or the current DHCPv6 method?
Now that the DHCPv4 authentication draft is in last call, should
we use the technique described in that document to provide
authentication for DHCPv6, or should we continue with the
authentication technique currently documented in the extensions
draft?
17.3. The Reconfigure Message and Subnet Prefix Extensions
The drafts currently specify that Releasable resources (such as an IP
address) can only be reconfigured using the Reconfigure-init trigger
message. This was done for simplicity (enables clients to perform
DAD on the new address and return the appropriate result to the
server) using the same mechanism as a standard Request/Reply/Release
exchange. This method also makes no assumptions about the
charactistics of the releasable resource.
However, for IP addresses with interface IDs, one could send out
two IP address extensions, one for the old prefix and one for the
new, and cause clients to change the prefix and thus renumber over
time. This scheme avoids the added DHCP Request traffic - clients
acknowledge with a Reconfigure-reply message.
17.4. ``R'' bit in Request message not needed?
Now that the transaction-ID is stored along with the releasable
resource identifier in a client's binding, the transaction-ID cache
becomes an optional feature of the DHCP server implementation, not a
requirement of the protocol. Should we do away with the ``R'' bit?
18. Security Considerations
Clients and servers often have to authenticate the messages they
exchange. For instance, a server may wish to be certain that a
Request originated from the client identified by the <link-local
address, subnet-prefix> fields included within the Request message
header. Conversely, it is quite often essential for a client to
be certain that the configuration parameters and addresses it has
received were sent to it by an authoritative server. Similarly, a
server should only accept a Release message which seems to be from
one of its clients, if it has some assurance that the client actually
Bound, Carney, Perkins Expires 1 November 2000 [Page 48]
Internet Draft DHCP for IPv6 5 May 2000
did transmit the Release message. Again, a client might wish to only
accept Reconfigure or Reconfigure-init messages that are certain to
have originated from a server with authority to issue them.
The IPv6 Authentication Header can provide security for DHCPv6
messages when both endpoints have a suitable IP address. However,
a client often has only a link-local address, and such an address
is not sufficient for a server which is off-link. In those
circumstances the relay is involved, so that the DHCP message MUST
have the relay's address in the IP destination address field, even
though the client aims to deliver the message to the server. The
DHCP Client-Server Authentication Extension [2] is intended to be
used in these circumstances.
Note that, if a client receives a DHCP message which fails
authentication, it should continue to wait for another message which
might be correctly authenticated just as if the failed message had
never arrived; however, receiving such failed messages SHOULD be
logged.
19. Year 2000 considerations
Since all times are relative to the current time of the transaction,
there is no problem within the DHCPv6 protocol related to any
hardcoded dates or two-digit representation of the current year.
20. IANA Considerations
This document defines message types 1--8 to be received by UDP at
port numbers 546 and 547. Additional message types may be defined in
the future.
Section 3.1 lists several multicast addresses used by DHCP.
This document also defines several status codes that are to
be returned with the Reply and Reconfigure-reply messages (see
sections 9.4 and 9.7). The non-zero values for these status codes
which are currently specified are shown in the table in section 3.4.
There is a DHCPv6 extension [2] which allows clients and servers to
exchange values for some of the timing and retransmission parameters
defined in section 3.5. Adding new parameters in the future would
require extending the values by which the parameters are indicated in
the DHCP extension. Since there needs to be a list kept, the default
values for each parameter should also be stored as part of the list.
Bound, Carney, Perkins Expires 1 November 2000 [Page 49]
Internet Draft DHCP for IPv6 5 May 2000
All of these protocol elements may be specified to assume new values
at some point in the future. New values should be approved by the
process of IETF Consensus [11].
21. Acknowledgements
Thanks to the DHC Working Group for their time and input into the
specification. Ralph Droms and Thomas Narten have had a major
role in shaping the continued improvement of the protocol by their
careful reviews. Many thanks to Matt Crawford, Erik Nordmark, Gerald
Maguire, and Mike Carney for their studied review as part of the
Last Call process. Thanks also for the consistent input, ideas, and
review by (in alphabetical order) Brian Carpenter, Jack McCann, Yakov
Rekhter, Matt Thomas, Sue Thomson, and Phil Wells.
Thanks to Steve Deering and Bob Hinden, who have consistently
taken the time to discuss the more complex parts of the IPv6
specifications.
A. Comparison between DHCPv4 and DHCPv6
This appendix is provided for readers who will find it useful to see
a model and architecture comparison between DHCPv4 [7, 1] and DHCPv6.
There are three key reasons for the differences:
o IPv6 inherently supports a new model and architecture for
communications and autoconfiguration of addresses.
o DHCPv6 benefits from the new IPv6 features.
o New features were added to support the expected evolution and
the existence of more complicated Internet network service
requirements.
IPv6 Architecture/Model Changes:
o The link-local address permits a node to have an address
immediately when the node boots, which means all clients have a
source IP address at all times to locate an on-link server or
relay.
o The need for BOOTP compatibility and the broadcast flag have been
removed.
o Multicast and address scoping in IPv6 permit the design of
discovery packets that would inherently define their range by the
multicast address for the function required.
Bound, Carney, Perkins Expires 1 November 2000 [Page 50]
Internet Draft DHCP for IPv6 5 May 2000
o Stateful autoconfiguration has to coexist and integrate with
stateless autoconfiguration supporting Duplicate Address
Detection and the two IPv6 lifetimes, to facilitate the dynamic
renumbering of addresses and the management of those addresses.
o Multiple addresses per interface are inherently supported in
IPv6.
o Some DHCPv4 options are unnecessary now because the configuration
parameters are either obtained through IPv6 Neighbor Discovery or
the Service Location protocol [16].
DHCPv6 Architecture/Model Changes:
o The message type is the first byte in the packet.
o IPv6 Address allocations are now handled in a message extension
as opposed to the message header.
o Client/Server bindings are now mandatory and take advantage of
the client's link-local address to always permit communications
either directly from an on-link server, or from a off-link server
through an on-link relay.
o Servers are discovered by a client Solicit, followed by a server
Advertise message
o The client will know if the server is on-link or off-link.
o The on-link relay may locate off-link server addresses from
system configuration or by the use of a site-wide multicast
packet.
o ACKs and NAKs are not used.
o The server assumes the client receives its responses unless it
receives a retransmission of the same client request. This
permits recovery in the case where the network has faulted.
o Clients can issue multiple, unrelated Request messages to the
same or different servers.
o The function of DHCPINFORM is inherent in the new packet design;
a client can request configuration parameters other than IPv6
addresses in the optional extension headers.
o Clients MUST listen to their UDP port for the new Reconfigure
message from servers.
Bound, Carney, Perkins Expires 1 November 2000 [Page 51]
Internet Draft DHCP for IPv6 5 May 2000
o New extensions have been defined.
With the changes just enumerated, we can support new user features,
including
o Configuration of Dynamic Updates to DNS
o Address deprecation, for dynamic renumbering.
o Relays can be preconfigured with server addresses, or use of
multicast.
o Authentication
o Clients can ask for multiple IP addresses.
o Addresses can be reclaimed using the Reconfigure-init message.
o Integration between stateless and stateful address
autoconfiguration.
o Enabling relays to locate off-link servers.
B. Full Copyright Statement
Copyright (C) The Internet Society (2000). 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
Bound, Carney, Perkins Expires 1 November 2000 [Page 52]
Internet Draft DHCP for IPv6 5 May 2000
HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
References
[1] S. Alexander and R. Droms. DHCP Options and BOOTP Vendor
Extensions. Request for Comments (Draft Standard) 2132,
Internet Engineering Task Force, March 1997.
[2] J. Bound, M. Carney, and C. Perkins. Extensions for the Dynamic
Host Configuration Protocol for IPv6.
draft-ietf-dhc-dhcpv6ext-12.txt, May 2000. (work in progress).
[3] S. Bradner. Key words for use in RFCs to Indicate Requirement
Levels. Request for Comments (Best Current Practice) 2119,
Internet Engineering Task Force, March 1997.
[4] S. Bradner and A. Mankin. The Recommendation for the IP Next
Generation Protocol. Request for Comments (Proposed Standard)
1752, Internet Engineering Task Force, January 1995.
[5] W. J. Croft and J. Gilmore. Bootstrap Protocol. Request for
Comments 951, Internet Engineering Task Force, September 1985.
[6] S. Deering and R. Hinden. Internet Protocol, Version 6 (IPv6)
Specification. Request for Comments (Draft Standard) 2460,
Internet Engineering Task Force, December 1998.
[7] R. Droms. Dynamic Host Configuration Protocol. Request for
Comments (Draft Standard) 2131, Internet Engineering Task Force,
March 1997.
[8] R. Hinden and S. Deering. IP Version 6 Addressing Architecture.
Request for Comments (Proposed Standard) 2373, Internet
Engineering Task Force, July 1998.
[9] S. Kent and R. Atkinson. IP Authentication Header. Request for
Comments (Proposed Standard) 2402, Internet Engineering Task
Force, November 1998.
[10] J. McCann, S. Deering, and J. Mogul. Path MTU Discovery for
IP version 6. Request for Comments (Proposed Standard) 1981,
Internet Engineering Task Force, August 1996.
[11] T. Narten and H. Alvestrand. Guidelines for Writing an IANA
Considerations Section in RFCs. Request for Comments (Best
Current Practice) 2434, Internet Engineering Task Force, October
1998.
Bound, Carney, Perkins Expires 1 November 2000 [Page 53]
Internet Draft DHCP for IPv6 5 May 2000
[12] T. Narten, E. Nordmark, and W. Simpson. Neighbor Discovery for
IP Version 6 (IPv6). Request for Comments (Draft Standard)
2461, Internet Engineering Task Force, December 1998.
[13] D. C. Plummer. Ethernet Address Resolution Protocol: Or
converting network protocol addresses to 48.bit Ethernet address
for transmission on Ethernet hardware. Request for Comments
(Standard) 826, Internet Engineering Task Force, November 1982.
[14] J. Postel. User Datagram Protocol. Request for Comments
(Standard) 768, Internet Engineering Task Force, August 1980.
[15] S. Thomson and T. Narten. IPv6 Stateless Address
Autoconfiguration. Request for Comments (Draft Standard) 2462,
Internet Engineering Task Force, December 1998.
[16] J. Veizades, E. Guttman, C. Perkins, and S. Kaplan. Service
Location Protocol. Request for Comments (Proposed Standard)
2165, Internet Engineering Task Force, June 1997.
[17] P. Vixie, Ed., S. Thomson, Y. Rekhter, and J. Bound. Dynamic
Updates in the Domain Name System (DNS UPDATE). Request for
Comments (Proposed Standard) 2136, Internet Engineering Task
Force, April 1997.
Bound, Carney, Perkins Expires 1 November 2000 [Page 54]
Internet Draft DHCP for IPv6 5 May 2000
Chair's Address
The working group can be contacted via the current chair:
Ralph Droms
Computer Science Department
323 Dana Engineering
Bucknell University
Lewisburg, PA 17837
Phone: (570) 577-1145
E-mail: droms@bucknell.edu
Author's Address
Questions about this memo can be directed to:
Jim Bound
Compaq Computer Corporation
Mail Stop: ZK03-3/U14
110 Spitbrook Road
Nashua, NH 03062
USA
Phone: +1-603-884-0400
Email: bound@zk3.dec.com
Mike Carney
Sun Microsystems, Inc
Mail Stop: UMPK17-202
901 San Antonio Road
Palo Alto, CA 94303-4900
USA
Phone: +1-650-786-4171
Email: mwc@eng.sun.com
Charles E. Perkins
Communications Systems Lab
Nokia Research Center
313 Fairchild Drive
Mountain View, California 94043
USA
Phone: +1-650 625-2986
EMail: charliep@iprg.nokia.com
Fax: +1 650 625-2502
Bound, Carney, Perkins Expires 1 November 2000 [Page 55]
