| Internet-Draft | IPv4 routes with an IPv6 next-hop | June 2021 |
| Chroboczek | Expires 20 December 2021 | [Page] |
This document defines an extension to the Babel routing protocol that allows annoncing routes to an IPv4 prefix with an IPv6 next-hop, which makes it possible for IPv4 traffic to flow through interfaces that have not been assigned an IPv4 address. This document updates RFC 8966.¶
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 20 December 2021.¶
Copyright (c) 2021 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
The role of a routing protocol is to build a routing table, a data structure that maps network prefixes in a given family (IPv4 or IPv6) to next hops, pairs of an outgoing interface and a neighbour's network address, for example:¶
destination next hop
2001:db8:0:1::/64 eth0, fe80::1234:5678
203.0.113.0/24 eth0, 192.0.2.1
¶
When a packet is routed according to a given routing table entry, the forwarding plane typically uses a neighbour discovery protocol (the Neighbour Discovery protocol (ND) [RFC4861] in the case of IPv6, the Address Resolution Protocol (ARP) [RFC0826] in the case of IPv4) to map the next-hop address to a link-layer address (a "MAC address"), which is then used to construct the link-layer frames that encapsulate forwarded packets.¶
It is apparent from the description above that there is no fundamental reason why the destination prefix and the next-hop address should be in the same address family: there is nothing preventing an IPv6 packet from being routed through a next hop with an IPv4 address (in which case the next hop's MAC address will be obtained using ARP), or, conversely, an IPv4 packet from being routed through a next hop with an IPv6 address. (In fact, it is even possible to store link-layer addresses directly in the next-hop entry of the routing table, which is commonly done in networks using the OSI protocol suite).¶
The case of routing IPv4 packets through an IPv6 next hop is particularly interesting, since it makes it possible to build networks that have no IPv4 addresses except at the edges and still provide IPv4 connectivity to edge hosts. In addition, since an IPv6 next hop can use a link-local address that is autonomously configured, the use of such routes enables a mode of operation where the network core has no statically assigned IP addresses of either family, which significantly reduces the amount of manual configuration required.¶
We call a route towards an IPv4 prefix that uses an IPv6 next hop a "v4-via-v6" route. This document describes an extension that allows the Babel routing protocol [RFC8966] to announce v4-via-v6 routes across interfaces that have no IPv4 addresses assigned. Section 3 describes procedures that ensure that all routers can originate ICMPv4 packets, even if they have not been assigned any IPv4 addresses.¶
The extension described in this document is inspired by a previously defined extension to the BGP protocol [RFC5549]. This document updates [RFC8966].¶
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.¶
The Babel protocol fully supports dual-stack operation: all data that represent a neighbour address or a network prefix are tagged by an Address Encoding (AE), a small integer that identifies the address family (IPv4 or IPv6) of the address of prefix, and describes how it is encoded. This extension defines a new AE, called v4-via-v6, which has the same format as the existing AE for IPv4 addresses. This new AE is only allowed in TLVs that carry network prefixes: TLVs that carry a neighbour address use one of the normal encodings for IPv6 addresses.¶
A Babel node that needs to announce an IPv4 route over an interface that has no assigned IPv4 address MAY make a v4-via-v6 announcement. In order to do so, it first establishes an IPv6 next-hop address in the usual manner (either by sending the Babel packet over IPv6, or by including a Next Hop TLV containing an IPv6 address and using AE 2 or 3); it then sends an Update, with AE equal to 4 (v4-via-v6) containing the IPv4 prefix being announced.¶
If the outgoing interface has been assigned an IPv4 address, then, in the interest of maximising compatibility with existing routers, the sender SHOULD prefer an ordinary IPv4 announcement; even in that case, however, it MAY send a v4-via-v6 announcement. A node SHOULD NOT send both ordinary IPv4 and v4-via-v6 announcements for the same prefix over a single interface (if the update is sent to a multicast address) or to a single neighbour (if sent to a unicast address), since doing that provides no benefit while doubling the amount of routing traffic.¶
Updates with infinite metric are retractions: they indicate that a previously announced route is no longer available. Retractions do not require a next hop, and there is therefore no difference between v4-via-v6 retractions and ordinary retractions. A node MAY send IPv4 retractions only, or it MAY send v4-via-v6 retractions on interfaces that have not been assigned an IPv4 address.¶
Upon reception of an Update TLV with AE equal to 4 (v4-via-v6) and finite metric, a Babel node computes the IPv6 next hop, as described in Section 4.6.9 of [RFC8966]. If no IPv6 next hop exists, then the Update MUST be silently ignored. If an IPv6 next hop exists, then the node MAY acquire the route being announced, as described in Section 3.5.3 of [RFC8966]; the parameters of the route are as follows:¶
An Update TLV with a v4-via-v6 AE and metric equal to infinity is a retraction: it announces that a previously available route is being retracted. In that case, no next hop is necessary, and the retraction is treated as described in Section 4.6.9 of [RFC8966].¶
As usual, a node MAY ignore the update, e.g., due to filtering (Appendix C of [RFC8966]). If a node cannot install v4-via-v6 routes, e.g., due to hardware or software limitations, then routes to an IPv4 prefix with an IPv6 next hop MUST NOT be selected, as described in Section 3.5.3 of [RFC8966].¶
Prefix and seqno requests are used to request an update for a given prefix. Since they are not related to a specific next hop, there is no semantic difference between IPv4 and v4-via-v6 requests. Therefore, a node SHOULD NOT send requests of either kind with the AE field being set to 4 (v4-via-v6); instead, it SHOULD request IPv4 updates by sending requests with the AE field being set to 1 (IPv4).¶
When receiving requests, AEs 1 (IPv4) and 4 (v4-via-v6) MUST be treated in the same manner: the receiver processes the request as described in Section 3.8 of [RFC8966]. If an Update is sent, then it MAY be sent with AE 1 or 4, as described in Section 2.1 above, irrespective of which AE was used in the request.¶
When receiving a request with AE 0 (wildcard), the receiver SHOULD send a full route dump, as described in Section 3.8.1.1 of [RFC8966]. Any IPv4 routes contained in the route dump MAY use either AE 1 (IPv4) or AE 4 (v4-via-v6), as described in Section 2.1 above.¶
The only other TLVs defined by [RFC8966] that carry an AE field are Next Hop and TLV. Next Hop and IHU TLVs MUST NOT carry the AE 4 (v4-via-v6).¶
The Internet Control Message Protocol (ICMPv4, or simply ICMP) [RFC0792] is a protocol related to IPv4 that is primarily used to carry diagnostic and debugging information. ICMPv4 packets may be originated by end hosts (e.g., the "destination unreachable, port unreachable" ICMPv4 packet), but they may also be originated by intermediate routers (e.g., most other kinds of "destination unreachable" packets).¶
Some protocols deployed in the Internet rely on ICMPv4 packets sent by intermediate routers. Most notably, path MTU Discovery (PMTUd) [RFC1191] is an algorithm executed by end hosts to discover the maximum packet size that a route is able to carry. While there exist variants of PMTUd that are purely end-to-end [RFC4821], the variant most commonly deployed in the Internet has a hard dependency on ICMPv4 packets originated by intermediate routers: if intermediate routers are unable to send ICMPv4 packets, PMTUd may lead to persistent blackholing of IPv4 traffic.¶
Due to this kind of dependency, every Babel router that is able to forward IPv4 traffic MUST be able originate ICMPv4 traffic. Since the extension described in this document enables routers to forward IPv4 traffic received over an interface that has not been assigned an IPv4 address, a router implementing this extension MUST be able to originate ICMPv4 packets even when the outgoing interface has not been assigned an IPv4 address.¶
There are various ways to meet this requirement, and choosing between them is left to the implementation. For example, if a router has an interface that has been assigned an IPv4 address, or if an IPv4 address has been assigned to the router itself (to the "loopback interface"), then that IPv4 address may be "borrowed" to serve as the source of originated ICMPv4 packets. If no IPv4 address is available, a router may choose a source address from a prefix known to be unused, for example from a suitably chosen private address range [RFC1918]. If no more suitable address is available, then a router MAY use the IPv4 dummy address 192.0.0.8 as the source address of the IMCPv4 packets that it sends. Note however that using the same address on multiple routers may hamper debugging and fault isolation, e.g., when using the "traceroute" utility.¶
This extension defines the v4-via-v6 AE, whose value is 4. This AE is solely used to tag network prefixes, and MUST NOT be used to tag neighbour addresses, e.g. in Next Hop or IHU TLVs.¶
This extension defines no new TLVs or sub-TLVs.¶
Network prefixes tagged with AE 4 (v4-via-v6) MUST be encoded and decoded just like prefixes tagged with AE 1 (IPv4), as described in Section 4.3.1 of [RFC8966].¶
A new compression state for AE 4 (v4-via-v6) distinct from that of AE 1 (IPv4) is introduced, and MUST be used for address compression of prefixes tagged with AE 4, as described in Section 4.6.9 of [RFC8966]¶
The following TLVs MAY be tagged with AE 4 (v4-via-v6):¶
As AE 4 (v4-via-v6) is suitable only for network prefixes, IHU (Type = 5) and Next-Hop (Type = 7) TLVs MUST NOT be tagged with AE 4. Such (incorrect) TLVs MUST be ignored upon reception.¶
An Update (Type = 8) TLV with AE 4 is constructed as described in Section 4.6.9 of [RFC8966] for AE 1 (IPv4), with the following specificities:¶
When tagged with the AE 4, Route Request and Seqno Request updates MUST be constructed and decoded as described in Section 4.6 of [RFC8966], and the network prefixes contained within them decoded as described in Section 4.1 above.¶
This protocol extension adds no new TLVs or sub-TLVs.¶
This protocol extension uses a new AE. As discussed in Appendix D of [RFC8966] and specified in the same document, implementations that do not understand the present extension will silently ignore the various TLVs that use this new AE. As a result, incompatible versions will ignore v4-via-v6 routes. They will also ignore requests with AE 4, which, as stated in Section 2.3, are NOT RECOMMENDED.¶
Using a new AE introduces a new compression state, used to parse the network prefixes. As this compression state is separate from other AEs' states, it will not interfere with the compression state of unextended nodes.¶
This extension reuses the next-hop state from AEs 2 and 3 (IPv6), but makes no changes to the way in which it is updated, and therefore causes no compatibility issues.¶
As mentioned in Section 2.1, ordinary IPv4 announcements are preferred to v4-via-v6 announcements when the outgoing interface has an assigned IPv4 address; doing otherwise would prevent routers that do not implement this extension from learning the route being announced.¶
IANA has allocated value 4 in the "Babel Address Encodings" registry as follows:¶
| AE | Name | Reference |
|---|---|---|
| 4 | v4-via-v6 | (this document) |
The extension defined in this document does not fundamentally change the security properties of the Babel protocol. However, by allowing IPv4 routes to be propagated across routers that have not been assigned IPv4 addresses, it might invalidate the assumptions made by network administrators, which could conceivably lead to security issues.¶
For example, if an island of IPv4-only hosts is separated from the IPv4 Internet by routers that have not been assigned IPv4 addresses, a network administrator might reasonably assume that the IPv4-only hosts are unreachable from the IPv4 Internet. This assumption is broken if the intermediary routers implement the extension described in this document, which might expose the IPv4-only hosts to traffic from the IPv4 Internet. If this is undesirable, the flow of IPv4 traffic must be restricted by the use of suitable filtering rules (Appendix C of [RFC8966]) together with matching packet filters in the data plane.¶
This protocol extension was originally designed, described and implemented in collaboration with Theophile Bastian. Margaret Cullen pointed out the issues with ICMP and helped coin the phrase "v4-via-v6". The author is also indebted to Donald Eastlake, Toke Hoiland-Jorgensen, and David Schinazi.¶