Internet-Draft intarea-dhcpv4-route4via6 April 2025
Lamparter & Fiebig Expires 25 October 2025 [Page]
Workgroup:
Internet Area Working Group
Internet-Draft:
draft-equinox-intarea-dhcpv4-route4via6-01
Published:
Intended Status:
Experimental
Expires:
Authors:
D. Lamparter
NetDEF, Inc.
T. Fiebig
MPI-INF

DHCPv4 Option for IPv4 routes with IPv6 nexthops

Abstract

As a result of the shortage of IPv4 addresses, installations are increasingly recovering IPv4 addresses from uses where they are not strictly necessary. One such situation is in establishing next hops for IPv4 routes, replacing this use with IPv6 addresses. This document describes how to provision DHCP-configured hosts with their routes in such a situation.

This draft lives at https://github.com/eqvinox/dhc-route4via6

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.

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."

This Internet-Draft will expire on 25 October 2025.

▲

Table of Contents

1. Introduction

IPv4 is currently (and will likely be for some time) in a situation where IPv4 addresses are in short supply, but services still need to be made available to users that do not yet have IPv6 connectivity. In some cases, even the service side may not have IPv6 support yet. In other cases some aspect of the service precludes using proxy-style service delivery with translation technologies on either or both sides. This leads to a need for fine-grained deployment of IPv4 connectivity with minimum wastage of addresses.

A particularly interesting improvement enabled by the extension described here is the complete removal of IPv4 addresses from first-hop routers acting as DHCPv4/v6 relays, while still providing IPv4 connectivity. In this scenario, the relay (assumed colocated with the router) has no IPv4 address to use to communicate with the client. An almost-working solution for this case is presented by [DHCPv6] with the [DHCP4o6] transport method. Since this mechanism encapsulates IPv4 DHCP messages, all related IPv4 configuration can be carried. However, DHCPv4 does not support a way to encode an IPv6 default gateway or other routes, which is necessary in this case.

If the router and relay are not co-located, the relay may have an IPv4 address while the router does not. In this case, the option described in this document could be carried in a plain IPv4 DHCP message.

Note that the changes described in this document are to DHCPv4, not DHCPv6.

2. Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.

3. Extended static route function

This document defines a DHCPv4 option container with 2 suboptions for use within it. The container can occur multiple times, and the suboptions can each occur zero, one, or multiple times in each container. Each container represents one "fully-filled matrix" of destination prefixes and next-hops, i.e. to be interpreted as a full cartesian combination of the two sets. In detail, this means:

  1. each option container is processed as one unit, building up a list of destination prefixes and next-hops. It is expected that the container will in most cases present one destination prefix with one or two next-hops. However, DHCPv4 clients processing the option MUST support processing multiple destination prefix suboptions. NB/TODO: this implies [DHCP-LONGOPT] is NOT applicable here. Not sure if that's a good idea.
  2. when multiple destination prefixes occur in the same container, all of the next-hops for that container MUST be processed for all of the destination prefixes in that container.
  3. if no destination prefix sub-option occurs in a container, the IPv4 default route (0.0.0.0/0) is implicitly expressed and the container MUST be processed as if the 0.0.0.0/0 destination prefix was explicitly encoded.
  4. if no next-hop sub-option occurs in a container, the "source of DHCP packet" function is encoded as with the unspecified-address (::) next-hop described below. The container MUST be processed as if a :: next-hop were explicitly encoded.
  5. both sub-options may be absent, the rules in the previous two bullet points MUST be applied in combination. This is in fact expected to be a relatively common setup and is intentionally made available with a space-preserving encoding.
  6. the container can be repeated an arbitrary number of times with different suboptions. Any single destination prefix MUST NOT be expressed in more than one container. If encountering duplicate destination prefixes, clients MAY pick one at random, or MAY merge nexthops; but this is considered a network side configuration error. However, note that this refers only to repeating the exact same destination prefix and prefix length. A differing prefix length presents a distinct destination prefix and MUST be processed independently.
  7. next-hops MUST NOT repeat within the same container, but the same next-hop MAY occur in any number of containers and MUST be processed normally in that case.
  8. if multiple next-hops are expressed in the same container, this represents equal-cost multipath routes across the set of next-hops. Clients processing the option MUST install at least one of the listed nexthops but MAY choose any subset at their discretion, e.g. if their capability of installing ECMP is limited to some number of next-hops (or is 1, i.e. no ECMP support.)

3.1. Applicable next-hop behavior

Outlined in [IANA-IPv6], not all IPv6 addresses are valid for use when encoded as next-hop and some have specific functionality [IANA-IPv6-SPECIAL] attached to them as follows:

  1. the unspecified-address nexthop (which is also implied by the absence of any nexthop suboption) indicates that the routes in the container should use the DHCP packet's source address as nexthop. When [DHCP4o6] is in use, hosts MUST retrieve the IPv6 source address of the DHCPv6 packet carrying the DHCPV4-RESPONSE message. TODO: does it really make sense to support IPv4 here? Maybe only allow this with DHCP4o6?
  2. the Discard-Only Address block (0100::/64) [DISCARD] MAY be used to express unreachable destinations, in particular if only limited but not global IPv4 connectivity is available. If this is used, it MUST be the only next-hop suboption in one container, clients SHOULD ignore the entire container if this is not respected. If a client is unable to mark destinations as unreachable in its routing table, it MAY ignore the container and SHOULD indicate a client configuration issue in its administrative interfaces.
  3. any unicast IPv6 address MAY be used as next-hop. This specifically also covers link-local addresses, which the client MUST support and MUST associate with the link that it has received the DHCP packets on.
  4. TODO: is ::ffff:192.0.2.123 an IPv4 nexthop? Is this worth supporting explicitly, and then saying that the other static route / default gateway options should be ignored?

  5. the following types/ranges of addresses are invalid and MUST NOT be used; no client behavior is specified if any are present in a container:

    • the loopback address (::1) TODO: express other directly-connected IPv4 hosts with this?
    • any multicast address (ff00::/8)
    • any address with a reserved allocation

3.2. Applicable destination prefix behavior

Some IPv4 prefixes, due to their function given in [IANA-IPv4], do not make sense to use with this option. DHCPv4 servers MUST NOT encode and DHCPv4 clients MUST ignore the following prefixes as well as any more-specific prefixes within them:

  • 0.0.0.0/8 (note that 0.0.0.0/0 is less specific than this, and thus valid)
  • 127.0.0.0/8
  • 224.0.0.0/4
  • 255.255.255.255/32

Behavior for 240.0.0.0/4 is outside the scope of this document.

4. Expected host behavior

The option described in this document is intended to be implemented on hosts supporting IPv4 routes with IPv6 nexthops as described in [v4overv6]. Hosts that do not support the behavior described there MUST NOT request and MUST ignore the option described in this document.

Hosts that support [v4overv6] behavior and acquire their configuration from [DHCP] SHOULD implement the option described here.

4.1. Singular address assignment

While not limited to this case, this option is expected and intended to be used with assigning a singular IPv4 address to a DHCPv4 client. This implies that the Subnet Mask option defined in [DHCP-OPT] will have the value 255.255.255.255.

DHCPv4 clients implementing the option described in this document MUST process such a Subnet Mask option value as assigning a single address. There is no network or broadcast address for this "single-sized" pseudo-subnet. No IPv4 addresses are expressed to be on-link for the purposes of [ARP] (though they MAY become so due to additional, e.g. local configuration assigning additional addresses to the interface.)

Whether the address is bound to the interface or host (strong vs. weak host model), and whether to perform or skip [DADv4] for the address is beyond the scope of this document.

4.2. Overlapping routes from other sources

[RFC3442] documents a mechanism to communicate a set of routes and their nexthops over DHCP. The original DHCP "router" option (code 3) may communicate a default router. If either of these options is used, the routes communicated may overlap.

To get consistent and unsurprising behavior, this document places the follwing expectations on the host: TODO: redundant paragraph/merge with text above, needs some merging/editing.

  • Routes that describe distinct destination prefixes MUST be handled independently. This includes routes that differ only in prefix length. As a result, the routing table MAY contain a mix of IPv4 routes with IPv4 nexthops as well as IPv6 nexthops. Standard longest prefix match behavior MUST be observed.
  • If routes with the same destination prefix are described both with previously existing methods as well as the options documented here, the route described by the latter MUST be used and the routes with IPv4 nexthops MUST be discarded. This notably includes "unreachable" routes described here; a route with an IPv4 nexthop for such a destination MUST still be discarded.
  • Multiple routes for the same destination prefix with different nexthops of the same address family SHOULD be combined into a single route for equal-cost multipath behavior, if the host supports this. If ECMP routes are not supported, the host MUST deterministically choose one of the routes. This MAY be done by using the first or last option as seen in DHCP packet order, or by choosing the numerically lowest or highest nexthop.

4.3. Default route

The default route is expressed here as a route for 0.0.0.0/0, which is also implied by the absence of any destination prefix suboption. There is no distinct special encoding for a default gateway, any nexthop for 0.0.0.0/0 MUST be treated as if it were a default gateway.

4.4. Routes clashing with the connected subnet

(only applicable if NOT assigning a single IPv4 address as /32) TODO: determine what behavior is reasonable here. (The client is likely to be given a /32 subnet mask anyway.)

5. DHCP Option encoding 

 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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|  Type (TBA1)  |     Length    |          Suboptions           |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+                               |
:                              ...                              :
Type
TBA1 (field defined in [DHCP-OPT])
Length
as defined in [DHCP-OPT]
Suboptions
zero or more suboptions as defined below

5.1. Destination prefix 

 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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|    Type = 1   |     Length    | R | prefixlen |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|                    IPv4 prefix (0 - 4 octets)                 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Type
1
Length
as defined in [DHCP-OPT]
R
Reserved bits. MUST be sent as zero, MUST be ignored on receipt.
prefixlen
IPv4 prefix length, integer value from 0 to 32 (inclusive)
IPv4 prefix

The route's destination prefix, encoded in as few bytes necessary for the given prefixlen value, i.e. calculate length as ceil(prefixlen / 8). TODO: maybe just ditch minimum length encoding, make it match other DHCPv4 bits?

Valid values are described in Section 3.2.

If the Length field indicates additional data past the IPv4 prefix value, clients MUST ignore it. Future documents MAY introduce other behavior here and servers MUST NOT send any such data until such a point.

5.2. Next-hop 

 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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|    Type = 2   |     Length    |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|                  IPv6 addresses (n*16 octets)                 |
:                              ...                              :
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Type
2
Length
as defined in [DHCP-OPT], a multiple of 16.
IPv6 addresses

One or more IPv6 addresses specifying the nexthop for this route. Refer to Section 3.1 for valid values and associated behavior. TODO: is this a list, or do we just repeat the suboption? In theory, this could have sub-suboptions, but unlikely to need?

6. Security Considerations

TBD

7. Privacy Considerations

TBD

8. IANA Considerations

A codepoint from the "BOOTP Vendor Extensions and DHCP Options" registry is requested for use with the container option described in Section 5. Editor note: 2 places of TBA1

A registry is requested to be created for the sub-options in the option above. TBD: proper wording for this, and fill in values 1 & 2

9. References

9.1. Normative References

[RFC2119]
Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, , <https://www.rfc-editor.org/info/rfc2119>.
[RFC8174]
Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, , <https://www.rfc-editor.org/info/rfc8174>.
[DHCP]
Droms, R., "Dynamic Host Configuration Protocol", RFC 2131, DOI 10.17487/RFC2131, , <https://www.rfc-editor.org/rfc/rfc2131>.
[DHCP-OPT]
Alexander, S. and R. Droms, "DHCP Options and BOOTP Vendor Extensions", RFC 2132, DOI 10.17487/RFC2132, , <https://www.rfc-editor.org/rfc/rfc2132>.
[DHCP-LONGOPT]
Lemon, T. and S. Cheshire, "Encoding Long Options in the Dynamic Host Configuration Protocol (DHCPv4)", RFC 3396, DOI 10.17487/RFC3396, , <https://www.rfc-editor.org/rfc/rfc3396>.
[v4overv6]
Chroboczek, J., Kumari, W. A., and T. Høiland-Jørgensen, "IPv4 routes with an IPv6 next hop", Work in Progress, Internet-Draft, draft-chroboczek-intarea-v4-via-v6-03, , <https://datatracker.ietf.org/doc/html/draft-chroboczek-intarea-v4-via-v6-03>.
[DISCARD]
Hilliard, N. and D. Freedman, "A Discard Prefix for IPv6", RFC 6666, DOI 10.17487/RFC6666, , <https://www.rfc-editor.org/rfc/rfc6666>.
[IANA-IPv4]
IANA, "IPv4 Address Space Registry", <https://www.iana.org/assignments/ipv4-address-space/>.
[IANA-IPv6]
IANA, "Internet Protocol Version 6 Address Space", <https://www.iana.org/assignments/ipv6-address-space/>.
[IANA-IPv6-SPECIAL]
IANA, "IPv6 Special-Purpose Address Registry", <https://www.iana.org/assignments/iana-ipv6-special-registry/>.

9.2. Informative References

[RFC3442]
Lemon, T., Cheshire, S., and B. Volz, "The Classless Static Route Option for Dynamic Host Configuration Protocol (DHCP) version 4", RFC 3442, DOI 10.17487/RFC3442, , <https://www.rfc-editor.org/info/rfc3442>.
[DHCPv6]
Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., Richardson, M., Jiang, S., Lemon, T., and T. Winters, "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", RFC 8415, DOI 10.17487/RFC8415, , <https://www.rfc-editor.org/info/rfc8415>.
[DHCP4o6]
Sun, Q., Cui, Y., Siodelski, M., Krishnan, S., and I. Farrer, "DHCPv4-over-DHCPv6 (DHCP 4o6) Transport", RFC 7341, DOI 10.17487/RFC7341, , <https://www.rfc-editor.org/info/rfc7341>.
[ARP]
Plummer, D., "An Ethernet Address Resolution Protocol: Or Converting Network Protocol Addresses to 48.bit Ethernet Address for Transmission on Ethernet Hardware", STD 37, RFC 826, DOI 10.17487/RFC0826, , <https://www.rfc-editor.org/rfc/rfc826>.
[DADv4]
Cheshire, S., "IPv4 Address Conflict Detection", RFC 5227, DOI 10.17487/RFC5227, , <https://www.rfc-editor.org/rfc/rfc5227>.

Acknowledgements

The authors would like to acknowledge and thank Tomek Mrugalski for very extensive comments, and in particular pointing out the proper way to use DHCP options.

Comments and feedback has been received and appreciated from Ole Troan.

Example encoded options

TBD: outdated examples removed, will be re-added

Revision history (TO BE REMOVED)

Authors' Addresses

David 'equinox' Lamparter
NetDEF, Inc.
San Jose,
United States of America
Tobias Fiebig
Max-Planck-Institut fuer Informatik
Campus E14
66123 Saarbruecken
Germany
Phone: +49 681 9325 3527