| Internet-Draft | Prefer RFC8781 | July 2025 |
| Buraglio, et al. | Expires 8 January 2026 | [Page] |
On networks providing IPv4-IPv6 translation (RFC7915), hosts and other endpoints might need to know the IPv6 prefix(es) used for translation (the NAT64 prefix). While "Discovery of the IPv6 Prefix Used for IPv6 Address Synthesis" (RFC7050) defines a DNS64-based prefix discovery mechanism, more robust methods have been developed since then. This document provides guidelines for NAT64 prefix discovery, recommending more deterministic alternatives like obtaining the NAT64 prefix from Router Advertisement option (RFC8781) over RFC7050 when available.¶
This note is to be removed before publishing as an RFC.¶
The latest revision of this draft can be found at https://github.com/buraglio/draft-nbtjjl-v6ops-prefer8781. Status information for this document may be found at https://datatracker.ietf.org/doc/draft-ietf-v6ops-prefer8781/.¶
Discussion of this document takes place on the v6ops Working Group mailing list (mailto:v6ops@ietf.org), which is archived at https://datatracker.ietf.org/wg/v6ops/about/. Subscribe at https://www.ietf.org/mailman/listinfo/v6ops/.¶
Source for this draft and an issue tracker can be found at https://github.com/buraglio/draft-nbtjjl-v6ops-prefer8781.¶
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 8 January 2026.¶
Copyright (c) 2025 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 Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
Devices translating between IPv4 and IPv6 packet headers [RFC7915] use a NAT64 prefix to map IPv4 addresses into the IPv6 address space, and vice versa. When a network provides NAT64, it is advantageous for endpoints to acquire the network's NAT64 prefixes (PREF64). Discovering the PREF64 enables endpoints to:¶
Implement the customer-side translator (CLAT) function of the 464XLAT architecture [RFC6877];¶
Translate IPv4 literals to IPv6 literals (Section 7.1 of [RFC8305]);¶
Dynamic PREF64 discovery is useful to keep the NAT64 prefix configuration up-to-date, particularly for unmanaged endpoints or endpoints which move between networks. [RFC7050] introduces the first DNS64-based mechanism for PREF64 discovery based on [RFC7051] analysis. However, subsequent methods have been developed to address [RFC7050] limitations.¶
For instance, [RFC8781] defines a Neighbor Discovery [RFC4861] option for Router Advertisements (RAs) to convey PREF64 information to hosts. This approach offers several advantages (Section 3 of [RFC8781]), including fate sharing with other host network configuration parameters.¶
Due to fundamental shortcomings of the [RFC7050] mechanism (Section 3), [RFC8781] is the preferred solution for new deployments. Implementations should strive for consistent PREF64 acquisition methods. The DNS64-based mechanism of [RFC7050] should be employed only when RA-based PREF64 delivery is unavailable, or as a fallback for legacy systems incapable of processing the PREF64 RA Option.¶
DNS64: a mechanism for synthesizing AAAA records from A records, defined in [RFC6147].¶
NAT64: a mechanism for translating IPv6 packets to IPv4 packets and vice versa. The translation is done by translating the packet headers according to the IP/ICMP Translation Algorithm defined in [RFC7915]. NAT64 translators can operate in stateless or stateful mode ([RFC6144]).¶
PREF64 (or Pref64::/n, or NAT64 prefix): An IPv6 prefix used for IPv6 address synthesis and for network addresses and protocols translation from IPv6 clients to IPv4 servers using the algorithm defined in [RFC6052].¶
Router Advertisement (RA): A packet used by Neighbor Discovery [RFC4861] and SLAAC to advertise the presence of the routers, together with other IPv6 configuration information.¶
SLAAC: StateLess Address AutoConfiguration, [RFC4862]¶
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.¶
DNS-based method of discovering the NAT64 prefix introduces some challenges, which make this approach less preferable than latest developed alternatives (such as PREF64 RA Option, [RFC8781]). This section outlines the key issues, associated with [RFC7050], with a focus on those not discussed in [RFC7050] or in the analysis of solutions for hosts to discover NAT64 prefix ([RFC7051]).¶
Fundamentally, the presence of the NAT64 and the exact value of the prefix used for the translation are network-specific attributes. Therefore, [RFC7050] requires the endpoint discovering the prefix to use the DNS64 resolvers provided by the network. If the device or an application is configured to use other recursive resolvers or runs a local recursive resolver, the corresponding name resolution APIs and libraries are required to recognize 'ipv4only.arpa.' as a special name and give it special treatment. This issue and remediation approach are discussed in [RFC8880]. However, it has been observed that very few [RFC7050] implementations support [RFC8880] requirements for special treatment of 'ipv4only.arpa.'. As a result, configuring such systems and applications to use resolvers other than the one provided by the network breaks the PREF64 discovery, leading to degraded user experience.¶
VPN applications may override the endpoint's DNS configuration, for example, by configuring enterprise DNS servers as the node's recursive resolvers and forcing all name resolution through the VPN. These enterprise DNS servers typically lack DNS64 functionality and therefore cannot provide information about the PREF64 used within the local network. Consequently, this prevents the endpoint from discovering the necessary PREF64, negatively impacting its connectivity on IPv6-only networks.¶
If both the network-provided DNS64 and the endpoint's resolver happen to utilize the WKP, the endpoint might, by pure coincidence, use a PREF64 that's valid for the current network. However, if the endpoint changes its network attachment, it can't detect if the new network lacks NAT64 entirely or uses a different NSP.¶
When using SLAAC, an IPv6 host typically requires a single RA to acquire its network configuration. For IPv6-only endpoints, timely PREF64 discovery is critical, particularly for those performing local DNS64 or NAT64 functions, such as CLAT ([RFC6877]). Until a PREF64 is obtained, the endpoint's IPv4-only applications and communication to IPv4-only destinations are impaired. The mechanism defined in [RFC7050] does not bundle PREF64 information with other network configuration parameters.¶
Section 3 of [RFC7050] states: "The node SHALL cache the replies it receives during the Pref64::/n discovery procedure, and it SHOULD repeat the discovery process ten seconds before the TTL of the Well-Known Name's synthetic AAAA resource record expires." As a result, once a PREF64 is discovered, it will be used until the TTL expired, or until the node disconnects from the network. There is no mechanism for an operator to force the PREF64 rediscovery on the node without disconnecting the node from the network. If the operator needs to change the PREF64 value used in the network, they need to proactively reduce the TTL value returned by the DNS64 server. This method has two significant drawbacks:¶
Many networks utilize external DNS64 servers and therefore have no control over the TTL value, if the PREF64 needs to be changed or withdrawn.¶
The PREF64 changes need to be planned and executed at least TTL seconds in advance. If the operator needs to notify nodes that a particular prefix must not be used (e.g. during a network outage or if the nodes learnt a rogue PREF64 as a result of an attack), it might not be possible without interrupting the network connectivity for the affected nodes.¶
Section 3 of [RFC7050] requires a node to examine all received AAAA resource records to discover one or more PREF64s and to utilize all learned prefixes. However, this approach presents challenges in some multihomed topologies where different DNS64 servers belonging to different ISPs might return different PREF64s. In such cases, it is crucial that traffic destined for synthesized addresses is sent to the correct NAT64 and the source address selected for those flows belongs to the prefix from that ISP's address space. In other words, the node needs to associate each discovered PREF64 with upstream information, including the IPv6 prefix and default gateway. Currently, there is no reliable way for a node to map a DNS64 response (and the prefix learned from it) to a specific upstream in a multihoming scenario. Consequently, the node might inadvertently select an incorrect source address for a given PREF64 and/or send traffic to the incorrect uplink.¶
As discussed in Section 7 of [RFC7050], the DNS-based PREF64 discovery is prone to DNS spoofing attacks. In addition to creating a wider attack surface for IPv6 deployments, [RFC7050] has other security challenges worth noting to justify declaring it legacy.¶
[RFC7050] requires a node's communication channel with a DNS64 server to be a "secure channel" which it defines to mean "a communication channel a node has between itself and a DNS64 server protecting DNS protocol-related messages from interception and tampering." This need is redundant when another communication mechanism of IPv6-related configuration, specifically RAs, can already be defended against tampering, for example by enabling RA-Guard [RFC6105]. Requiring nodes to implement two defense mechanisms when only one is necessary when [RFC8781] is used in place of [RFC7050] creates unnecessary risk.¶
One of the two examples that [RFC7050] defines to qualify a communication channel with a DNS64 server is the use of an "IPsec-based virtual private network (VPN) tunnel". As of the time of this writing, this is not supported as a practice by any common operating system DNS client. While they could, there have also since been multiple mechanisms defined for performing DNS-specific encryption such as those defined in [RFC9499] that would be more appropriately scoped to the applicable DNS traffic. These are also compatible with encrypted DNS advertisement by the network using Discovery of Network-designated Resolvers [RFC9463] that would ensure the clients know in advance that the DNS64 server supported the encryption mechanism.¶
The other example given by [RFC7050] that would allow a communication channel with a DNS64 server to qualify as a "secure channel" is the use of a "link layer utilizing data encryption technologies". As of the time of this writing, most common link layer implementations use data encryption already with no extra effort needed on the part of network nodes. While this appears to be a trivial way to satisfy this requirement, it also renders the requirement meaningless since any node along the path can still read the higher-layer DNS traffic containing the translation prefix. This seems to be at odds with the definition of "secure channel" as explained in Section 2.2 of [RFC7050].¶
Operators deploying NAT64 SHOULD provide PREF64 information in Router Advertisements per [RFC8781].¶
While [RFC8781] support is widely integrated into modern operating systems on mobile endpoints, equipment deployed in mobile network environments often lacks abilities to include the PREF64 Option into RAs. Therefore, the immediate deployment and enablement of PREF64 by mobile operators may not currently be feasible and the recommendations outlined in this document are not presently applicable to mobile network operators. These environments are encouraged to incorporate [RFC8781] when made practical by infrastructure upgrades or software stack feature additions.¶
Endpoints SHOULD attempt to obtain PREF64 information from RAs per [RFC8781] instead of using [RFC7050] method. In the absence of the PREF64 information in RAs, an endpoint MAY choose to fall back to the mechanism defined in RFC7050. This recommendation to prefer the [RFC8781] mechanism over one defined in [RFC7050] is consistent with Section 5.1 of [RFC8781].¶
Transitioning from the [RFC7050] heuristic to using the [RFC8781] approach might require a period where both mechanisms coexist. The duration of such period and feasibility of discontinuing DNS64 support, relying solely on RA-based PREF64 signaling in a given network depends on the endpoint footprint, particularly the presence and number of endpoints running outdated operating systems, which do not support [RFC8781].¶
Migrating away from DNS64-based discovery also reduces dependency on DNS64 in general, thereby eliminating DNSSEC and DNS64 incompatibility concerns (Section 6.2 of [RFC6147]).¶
Obtaining PREF64 information using RAs improves the overall security of an IPv6-only endpoint as it mitigates all attack vectors related to spoofed or rogue DNS response, as discussed in Section 7 of [RFC7050]. Security considerations related to obtaining PREF64 information from RAs are discussed in Section 7 of [RFC8781].¶
This document does not introduce any IANA considerations.¶
The authors would like to thank the following people for their valuable contributions: Mohamed Boucadair, Lorenzo Colitti, Tom Costello, Charles Eckel, Nick Heatley, Gabor Lencse and Peter Schmitt.¶