| v6ops | J. Palet Martinez |
| Internet-Draft | The IPv6 Company |
| Intended status: Informational | April 2, 2019 |
| Expires: October 4, 2019 |
NAT64/464XLAT Deployment Guidelines in Operator and Enterprise Networks
draft-ietf-v6ops-nat64-deployment-04
This document describes how NAT64 and 464XLAT can be deployed in an IPv6 network, whether cellular ISP, broadband ISP, or enterprise and the issues to be considered when having an IPv6-only access link, regarding: a) DNS64, b) applications or devices that use literal IPv4 addresses or non-IPv6 compliant APIs, and c) IPv4-only hosts or applications.
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 October 4, 2019.
Copyright (c) 2019 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.
NAT64 ([RFC6146]) describes a stateful IPv6 to IPv4 translation mechanism, which allows IPv6-only hosts to communicate with IPv4-only servers using unicast UDP, TCP or ICMP, by means of IPv4 public addresses sharing (assigned to the translator), among multiple IPv6-only hosts.
The translation of the packet headers is done using the IP/ICMP translation algorithm defined in [RFC7915] and algorithmically translating the IPv4 addresses to IPv6 addresses following [RFC6052].
DNS64 ([RFC6147]), is in charge of the synthesis of AAAA records from the A records, so only works for applications making use of DNS, avoiding changes in both, the IPv6-only hosts and the IPv4-only server, so they can use a NAT64. As discussed in Section 5.5 of [RFC6147], a security-aware and validating host has to perform the DNS64 function locally.
However, the use of NAT64 and/or DNS64 present three issues:
The same issues are true if part of, for example, an enterprise network, is connected to other parts of the same network or third party networks by means of IPv6-only connectivity. This applies to many other similar cases.
According to that, across this document, the use of "operator", "operator network", "service provider", and similar ones, are interchangeable with equivalent cases of enterprise networks (and similar ones). This may be also the case for other "managed end-user networks".
An analysis of stateful IPv6/IPv6 mechanisms is provided in [RFC6889].
This document looks into different possible NAT64 ([RFC6146]) deployment scenarios, including IPv4-IPv6-IPv4 and similar ones, which were not documented by [RFC6144], such as 464XLAT ([RFC6877]), in operator (broadband and cellular) and enterprise networks, and provides guidelines to avoid the above-mentioned issues.
Towards that, this document first looks into the possible NAT64 deployment scenarios (split in "known to work" and "known to work under special conditions"), providing a quick and generic comparison table among them. Then the document describes the issues that an operator need to understand on different matters that will allow to define what is the best approach/scenario for each specific network case. A summary provides some recommendations and decision points and then a clarification of the usage of this document for enterprise networks is provided. Finally, an annex provides an example of a broadband deployment using 464XLAT.
The target deployment scenarios in this document may be covered as well by other IPv4-as-a-Service transition mechanisms. It is out of scope of this document to provide a comparison among them.
[RFC7269] provides additional information about NAT64 deployment options and experiences.
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.
Section 7 of DNS64 ([RFC6147]), provides 3 scenarios, looking at the location of the DNS64. However, since the publication of that document, other possible scenarios and NAT64 use cases need to be considered in actual networks, despite some of them were specifically ruled out of the original NAT64/DNS64 work.
Consequently, the perspective in this document is to broaden those scenarios, including a few new ones. However, in order to be able to reduce the number of possible cases, we work under the assumption that typically, the service provider wants to make sure that all the customers have a service without failures. This means considering the worst possible case:
The document uses a common set of possible "participant entities":
The possible scenarios are split in two general categories:
The scenarios in this category are known to work. Each one may have different pros and cons, and in some cases the trade-offs, maybe acceptable for some operators.
In this scenario, the service provider offers both, the NAT64 and the DNS64 functions.
This is probably the most common scenario, however also may have the implications related the DNSSEC.
This scenario also may fail to solve the issue of literal addresses or non-IPv6 compliant APIs, as well as the issue of IPv4-only hosts or applications inside the IPv6-only access network.
+----------+ +----------+ +----------+
| | | NAT64 | | |
| IPv6 +--------+ + +--------+ IPv4 |
| | | DNS64 | | |
+----------+ +----------+ +----------+
Figure 1: NAT64 with DNS64
A totally equivalent scenario will be if the service provider offers only the DNS64 function, and the NAT64 function is provided by an outsourcing agreement with an external provider. All the considerations in the previous paragraphs of this section are the same for this sub-case.
+----------+
| |
| extNAT64 |
| |
+----+-----+
|
|
+----------+ +----+-----+ +----------+
| | | | | |
| IPv6 +--------+ DNS64 +--------+ IPv4 |
| | | | | |
+----------+ +----------+ +----------+
Figure 2: NAT64 in external service provider
As well, is equivalent to the scenario where the outsourcing agreement with the external provider is to provide both the NAT64 and DNS64 functions. Once more, all the considerations in the previous paragraphs of this section are the same for this sub-case.
+----------+
| extNAT64 |
| + |
| extDNS64 |
+----+-----+
|
+----------+ | +----------+
| | | | |
| IPv6 +-------------+--------------+ IPv4 |
| | | |
+----------+ +----------+
Figure 3: NAT64 and DNS64 in external provider
One more equivalent scenario will be if the service provider offers the NAT64 only, and the DNS64 function is from an external provider with or without a specific agreement among them. This is a scenario already common today, as several "global" service providers provide free DNS/DNS64 services and users often configure manually their DNS. This will only work if both the NAT64 and the DNS64 are using the WKP (Well-Known Prefix) or the same NSP (Network-Specific Prefix). All the considerations in the previous paragraphs of this section are the same for this sub-case.
Of course, if the external DNS64 is agreed with the service provider, then we are in the same case as in the previous ones already depicted in this scenario.
+----------+
| |
| extDNS64 |
| |
+----+-----+
|
|
+----------+ +----+-----+ +----------+
| | | | | |
| IPv6 +--------+ NAT64 +--------+ IPv4 |
| | | | | |
+----------+ +----------+ +----------+
Figure 4: NAT64; DNS64 by external provider
464XLAT ([RFC6877]) describes an architecture that provides IPv4 connectivity across a network, or part of it, when it is only natively transporting IPv6.
In order to do that, 464XLAT ([RFC6877]) relies on the combination of existing protocols:
Note that even in the 464XLAT ([RFC6877]) terminology, the provider-side translator is referred as PLAT, for simplicity and uniformity, in this document is always referred as NAT64.
In this scenario the service provider deploys 464XLAT with DNS64.
As a consequence, the DNSSEC issues remain, unless the host is doing the address synthesis.
464XLAT ([RFC6877]) is a very simple approach to cope with the major NAT64+DNS64 drawback: Not working with applications or devices that use literal IPv4 addresses or non-IPv6 compliant APIs.
464XLAT ([RFC6877]) has been used initially mainly in IPv6-only cellular networks. By supporting CLAT, the end-user device applications can access IPv4-only end-networks/applications, despite those applications or devices use literal IPv4 addresses or non-IPv6 compliant APIs.
In addition to that, in the same example of the cellular network above, if the User Equipment (UE) provides tethering, other devices behind it will be presented with a traditional NAT44, in addition to the native IPv6 support, so clearly it allows IPv4-only hosts inside the IPv6-only access network.
Furthermore, as indicated in [RFC6877], 464XLAT can be used in broadband IPv6 network architectures, by implementing the CLAT functionality at the CE.
In order to understand all the communication possibilities, let's assume the following representation of two dual-stack peers:
+-------+ .-----. .-----.
| | / \ / \
.-----. | Res./ | / IPv6- \ .-----. / IPv4- \
/ Local \ | SOHO +--( only )---( NAT64 )---( only )
/ \ | | \ Internet/\ `-----´ \ Internet/
( Dual- )--+ IPv6 | \ / \ / \ /
\ Stack / | CE | `--+--´ \ .-----. / `--+--´
\ Peer / | with | | \ / Remote\/ |
`-----´ | CLAT | +---+----+ / \ +---+----+
| | |DNS/IPv6| ( Dual- ) |DNS/IPv4|
+-------+ | with | \ Stack / +--------+
| DNS64 | \ Peer /
+--------+ `-----´
Representation of 464XLAT among two peers with DNS64
The possible communication paths, among the IPv4/IPv6 stacks of both peers, in this case, are:
The following figures show different choices for placing the different elements.
+----------+ +----------+ +----------+
| IPv6 | | NAT64 | | |
| + +--------+ + +--------+ IPv4 |
| CLAT | | DNS64 | | |
+----------+ +----------+ +----------+
Figure 5: 464XLAT with DNS64
An equivalent scenario will be if the service provider offers only the DNS64 function, and the NAT64 function is provided by an outsourcing agreement with an external provider. All the considerations in the previous paragraphs of this section are the same for this sub-case.
+----------+
| |
| extNAT64 |
| |
+----+-----+
|
|
+----------+ +----+-----+ +----------+
| IPv6 | | | | |
| + +--------+ DNS64 +--------+ IPv4 |
| CLAT | | | | |
+----------+ +----------+ +----------+
Figure 6: 464XLAT with DNS64; NAT64 in external provider
As well, is equivalent to the scenario where the outsourcing agreement with the external provider is to provide both the NAT64 and DNS64 functions. Once more, all the considerations in the previous paragraphs of this section are the same for this sub-case.
+----------+
| extNAT64 |
| + |
| extDNS64 |
+----+-----+
|
+----------+ | +----------+
| IPv6 | | | |
| + +-------------+--------------+ IPv4 |
| CLAT | | |
+----------+ +----------+
Figure 7: 464XLAT with DNS64; NAT64 and DNS64 in external provider
The major advantage of this scenario, using 464XLAT without DNS64, is that the service provider ensures that DNSSEC is never broken, even in case the user modifies the DNS configuration.
In this scenario, as in the previous one, there are no issues related to IPv4-only hosts (or IPv4-only applications) inside the IPv6-only access network, neither to the usage of IPv4 literals or non-IPv6 compliant APIs.
Let's assume the representation of two dual-stack peers as in the previous case:
+-------+ .-----. .-----.
| | / \ / \
.-----. | Res./ | / IPv6- \ .-----. / IPv4- \
/ Local \ | SOHO +--( only )---( NAT64 )---( only )
/ \ | | \ Internet/\ `-----´ \ Internet/
( Dual- )--+ IPv6 | \ / \ / \ /
\ Stack / | CE | `--+--´ \ .-----. / `--+--´
\ Peer / | with | | \ / Remote\/ |
`-----´ | CLAT | +---+----+ / \ +---+----+
| | |DNS/IPv6| ( Dual- ) |DNS/IPv4|
+-------+ +--------+ \ Stack / +--------+
\ Peer /
`-----´
Representation of 464XLAT among two peers without DNS64
The possible communication paths, among the IPv4/IPv6 stacks of both peers, in this case, are:
It needs to be noticed that this scenario works while the local hosts/applications are dual-stack (which is the current situation), because the connectivity from a local-IPv6 to a remote-IPv4 is not possible without an AAAA synthesis. This aspect is important only when in the LANs behind the CLAT there are IPv6-only hosts and they need to communicate with remote IPv4-only hosts. However, doesn't look a sensible approach from an Operating System or application vendor perspective, to provide IPv6-only support unless, similarly to c. above, there is certainty of peers supporting IPv6 as well.
The following figures show different choices for placing the different elements.
+----------+ +----------+ +----------+
| IPv6 | | | | |
| + +--------+ NAT64 +--------+ IPv4 |
| CLAT | | | | |
+----------+ +----------+ +----------+
Figure 8: 464XLAT without DNS64
This is equivalent to the scenario where there is an outsourcing agreement with an external provider for the NAT64 function. All the considerations in the previous paragraphs of this section are the same for this sub-case.
+----------+
| |
| extNAT64 |
| |
+----+-----+
|
+----------+ | +----------+
| IPv6 | | | |
| + +-------------+--------------+ IPv4 |
| CLAT | | |
+----------+ +----------+
Figure 9: 464XLAT without DNS64; NAT64 in external provider
The scenarios in this category are known to not work unless significant effort is devoted to solve the issues, or are intended to solve problems across "closed" networks, instead of as a general Internet access usage. In addition to the different pros, cons and trade-offs, which may be acceptable for some operators, they have implementation difficulties, as they are beyond the original expectations of the NAT64/DNS64 original intent.
In this scenario, the service provider offers a NAT64, however there is no DNS64 function support.
As a consequence, an IPv6 host in the IPv6-only access network, will not be able to detect the presence of DNS64 by means of [RFC7050], neither learning the IPv6 prefix to be used for the NAT64.
This can be sorted out as indicated in Section 4.1.1.
However, despite that, because the lack of the DNS64 function, the IPv6 host will not be able to obtain AAAA synthesized records, so the NAT64 becomes useless.
An exception to this "useless" scenario will be manually configure mappings between the A records of each of the IPv4-only remote hosts and the corresponding AAAA records, with the WKP (Well-Known Prefix) or NSP (Network-Specific Prefix) used by the service provider NAT64, as if they were synthesized by a DNS64.
This mapping could be done by several means, typically at the authoritative DNS server, or at the service provider resolvers by means of DNS RPZ (Response Policy Zones). The latest, may have implications in DNSSEC, if the zone is signed. Also, if the service provider is using an NSP, having the mapping at the authoritative server, will mean that may create troubles to other parties trying to use different NSP or the WKP, unless multiple DNS "views" are also being used at the authoritative servers.
Generally, the mappings alternative, will only make sense if a few set of IPv4-only remote hosts need to be accessed by a single network or reduced set of them, which support IPv6-only in the access, with some kind of mutual agreement for using this procedure, so it doesn't care if they become a trouble for other parties across Internet ("closed services").
In any case, this scenario doesn't solve the issue of literal addresses or non-IPv6 compliant APIs, neither it solves the problem of IPv4-only hosts within that IPv6-only access network.
+----------+ +----------+ +----------+
| | | | | |
| IPv6 +--------+ NAT64 +--------+ IPv4 |
| | | | | |
+----------+ +----------+ +----------+
Figure 10: NAT64 without DNS64
In this scenario, the service provider offers the NAT64, but not the DNS64. However, the IPv6 hosts have a built-in DNS64 function.
This may become common if the DNS64 function is implemented in all the IPv6 hosts/stacks, which is not the actual situation. At this way, the DNSSEC validation is performed on the A record, and then the host can use the DNS64 function so to be able to use the NAT64, without any DNSSEC issues.
This scenario fails to solve the issue of literal addresses or non-IPv6 compliant APIs, unless the IPv6 hosts also supports Happy Eyeballs v2 ([RFC8305], Section 7.1), which may solve that issue.
However, this scenario still fails to solve the problem of IPv4-only hosts or applications inside the IPv6-only access network.
+----------+ +----------+ +----------+
| IPv6 | | | | |
| + +--------+ NAT64 +--------+ IPv4 |
| DNS64 | | | | |
+----------+ +----------+ +----------+
Figure 11: NAT64; DNS64 in IPv6 hosts
In this scenario, the service provider offers the NAT64 only. The remote IPv4-only network offers the DNS64 function.
This is not common, and looks like doesn't make too much sense that a remote network, not deploying IPv6, is providing a DNS64 function and as in the case of the scenario depicted in Section 3.2.1, it will only work if both sides are using the WKP or the same NSP so, the same considerations apply. It can be also tuned to behave as in Section 3.1.1
This scenario still fails to solve the issue of literal addresses or non-IPv6 compliant APIs.
This scenario also fails to solve the problem of IPv4-only hosts or applications inside the IPv6-only access network.
+----------+ +----------+ +----------+
| | | | | IPv4 |
| IPv6 +--------+ NAT64 +--------+ + |
| | | | | DNS64 |
+----------+ +----------+ +----------+
Figure 12: NAT64; DNS64 in the IPv4-only
This section compares the different scenarios, including the possible variations (each one represented in the precedent sections by a different figure), looking at the following parameters:
In the next table, the columns represent each of the scenario from the previous sections, by the Figure number. The possible values are:
Needs to be noted that in some cases "countermeasures", alternative or special configurations, may be available for the items designated as "bad", so this comparison is making a generic case, as a quick comparison guide. In some cases, a "bad" item is not necessarily a negative aspect, all it depends on the specific needs/characteristics of the network where the deployment will take place. For instance, in a network which has only IPv6-only hosts and apps using only DNS and IPv6-compliant APIs, there is no impact using only NAT64 and DNS64, but if the hosts may validate DNSSEC, that item is still relevant.
+----------------+---+---+---+---+---+---+---+---+---+----+----+----+
| Item / Figure | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 |
+----------------+---+---+---+---+---+---+---+---+---+----+----+----+
| DNSSEC | - | - | - | - | - | - | - | + | + | + | + | + |
+----------------+---+---+---+---+---+---+---+---+---+----+----+----+
| Literal/APIs | - | - | - | - | + | + | + | + | + | - | - | - |
+----------------+---+---+---+---+---+---+---+---+---+----+----+----+
| IPv4-only | - | - | - | - | + | + | + | + | + | - | - | - |
+----------------+---+---+---+---+---+---+---+---+---+----+----+----+
| Foreign DNS | - | - | - | - | + | + | + | + | + | - | + | - |
+----------------+---+---+---+---+---+---+---+---+---+----+----+----+
Figure 13: Scenario Comparison
As a general conclusion, we should note that if the network must support applications using literals, non-IPv6-compliant APIs, or IPv4-only hosts or applications, only the scenarios with 464XLAT, or equivalent built-in local address synthesis features, will provide a solution. Further to that, those scenarios will also keep working if the user changes the DNS setup. Clearly also, depending on if DNS64 is used or not, DNSSEC may be broken for those hosts doing DNSSEC validation.
This section reviews the different issues that an operator needs to consider towards a NAT64/464XLAT deployment, as they may bring to decision points about how to approach that deployment.
As indicated in Section 8 of [RFC6147] (DNS64, Security Considerations), because DNS64 modifies DNS answers and DNSSEC is designed to detect such modifications, DNS64 can break DNSSEC.
If a device connected to an IPv6-only WAN queries for a domain name in a signed zone, by means of a recursive name server that supports DNS64, and the result is a synthesized AAAA record, and the recursive name server is configured to perform DNSSEC validation and has a valid chain of trust to the zone in question, it will cryptographically validate the negative response from the authoritative name server. This is the expected DNS64 behavior: The recursive name server actually lies to the client device. However, in most of the cases, the client will not notice it, because generally, they don't perform validation themselves and instead, rely on the recursive name servers.
A validating DNS64 resolver in fact, increase the confidence on the synthetic AAAA, as it has validated that a non-synthetic AAAA for sure, doesn't exists. However, if the client device is NAT64-oblivious (most common case) and performs DNSSEC validation on the AAAA record, it will fail as it is a synthesized record.
The best possible scenario from DNSSEC point of view is when the client requests the DNS64 server to perform the DNSSEC validation (by setting the DO bit to 1 and the CD bit to 0). In this case, the DNS64 server validates the data thus tampering may only happen inside the DNS64 server (which is considered as a trusted part, thus its likelihood is low) or between the DNS64 server and the client. All other parts of the system (including transmission and caching) are protected by DNSSEC ([Threat-DNS64]).
Similarly, if the client querying the recursive name server is another name server configured to use it as a forwarder, and is performing DNSSEC validation, it will also fail on any synthesized AAAA record.
All those considerations are extensively covered in Sections 3, 5.5 and 6.2 of [RFC6147].
The ideal solution to avoid DNSSEC issues, will be that all the signed zones also provide IPv6 connectivity, together with the corresponding AAAA records, which is out of the control of the operator needing to deploy NAT64. This has been proposed already in [I-D.bp-v6ops-ipv6-ready-dns-dnssec].
An alternative solution, which was the one considered while developing [RFC6147], is that validators will be DNS64-aware, so could perform the necessary discovery and do their own synthesis. That was done under the expectation that it was sufficiently early in the validator-deployment curve that it would be ok to break certain DNSSEC assumptions for networks who were really stuck in a NAT64/DNS64-needing world.
As already indicated, the scenarios in the previous section, are in fact somehow simplified, looking at the worst possible case (or saying it in a different way: "trying to look for the most perfect approach"), because breaking DNSSEC will not happen if the end-host is not doing validation. Previous data seems to indicate that the figures of DNSSEC actually broken by using DNS64 will be around 1.7% ([About-DNS64]) of the cases. So, a decision point for the operator must depend on "do I really care for that percentage of cases and the impact in my helpdesk or can I provide alternative solutions for them?". Some possible solutions may be taken, as depicted in the next sections.
The ideal solution will be to avoid using DNS64, but as already indicated this is not possible in all the scenarios.
However, not having a DNS64, means that is not possible to heuristically discover the NAT64 ([RFC7050]) and consequently, an IPv6 host in the IPv6-only access network, will not be able to detect the presence of the NAT64, neither to learn the IPv6 prefix to be used for it, unless it is configured by alternative means.
The learning of the IPv6 prefix could be solved by means of adding the relevant AAAA records to the ipv4only.arpa. zone of the service provider recursive servers, i.e., if using the WKP (64:ff9b::/96):
ipv4only.arpa. SOA . . 0 0 0 0 0 ipv4only.arpa. NS . ipv4only.arpa. AAAA 64:ff9b::192.0.0.170 ipv4only.arpa. AAAA 64:ff9b::192.0.0.171 ipv4only.arpa. A 192.0.0.170 ipv4only.arpa. A 192.0.0.171
An alternative option to the above, is the use of DNS RPZ (Response Policy Zones).
One more alternative, only valid in environments with PCP support (for both the hosts or CEs and for the service provider network), to follow [RFC7225] (Discovering NAT64 IPv6 Prefixes using PCP).
Other alternatives may be available in the future, such as Router Advertising ([I-D.ietf-6man-ra-pref64]) or DHCPv6 options.
It may be convenient to support at the same time several of the approaches described, in order to ensure that clients with different ways to configure the NAT64 prefix, successfully obtain it. This is also convenient even if DNS64 is being used.
In general, DNS servers with DNS64 function, by default, will not synthesize AAAA responses if the DNSSEC OK (DO) flag was set in the query. In this case, as only an A record is available, it means that the CLAT will take the responsibility, as in the case of literal IPv4 addresses, to keep that traffic flow end-to-end as IPv4, so DNSSEC is not broken. However, this will not work if a CLAT is not present as the hosts will not be able to use IPv4 (scenarios without 464XLAT).
If the DO flag is set and the client device performs DNSSEC validation, and the Checking Disabled (CD) flag is set for a query, as the DNS64 recursive server will not synthesize AAAA responses, the client could perform the DNSSEC validation with the A record and then may query the network for a NAT64 prefix ([RFC7050], [RFC7225], [I-D.ietf-6man-ra-pref64] or other methods) in order to synthesize the AAAA ([RFC6052]). This allow the client device to avoid using the CLAT and still use NAT64 even with DNSSEC.
If the end-host is IPv4-only, this will not work if a CLAT is not present (scenarios without 464XLAT), unless the client is able to locally perform the address synthesis.
Some devices/OSs may implement, instead of CLAT, a similar function by using Bump-in-the-Host ([RFC6535]), implemented as part of Happy Eyeballs v2 (Section 7.1 of [RFC8305]). In this case, the considerations in the above paragraphs are also applicable.
If a CE includes CLAT support and also a DNS proxy, as indicated in Section 6.4 of [RFC6877], the CE could behave as a stub validator on behalf of the client devices, following the same approach described in the precedent Section 4.1.3. So, the DNS proxy actually lies to the client devices, which in most of the cases will not notice it unless they perform validation themselves. Again, this allow the client devices to avoid using the CLAT and still use NAT64 with DNSSEC.
Once more, this will not work without a CLAT (scenarios without 464XLAT).
In cases of dual-stack clients, stub resolvers should send the AAAA queries before the A ones. So, such clients, if DNS64 is enabled, will never get A records, even for IPv4-only servers, and they may be in the path before the NAT64 and accessible by IPv4. If DNSSEC is being used for all those flows, specific addresses or prefixes can be left-out the DNS64 synthesis by means of ACLs.
Once more, this will not work without a CLAT (scenarios without 464XLAT).
If there are well-known specific IPv4 addresses or prefixes using DNSSEC, they can be mapped-out of the DNS64 synthesis.
Even if this is not related to DNSSEC, this "mapping-out" feature is actually, quite commonly used to ensure that [RFC1918] addresses (for example used by LAN servers) are not synthesized to AAAA.
Once more, this will not work without a CLAT (scenarios without 464XLAT).
When a client device, using a name server configured to perform DNS64, tries to reverse-map a synthesized IPv6 address, the name server responds with a CNAME record pointing the domain name used to reverse-map the synthesized IPv6 address (the one under ip6.arpa), to the domain name corresponding to the embedded IPv4 address (under in-addr.arpa).
This is the expected behavior, so no issues to be considered regarding DNS reverse mapping.
In the case the client device is IPv6-only (either because the stack or application is IPv6-only, or because it is connected via an IPv6-only LAN) and the remote server is IPv4-only (either because the stack is IPv4-only, or because it is connected via an IPv4-only LAN), only NAT64 combined with DNS64 will be able to provide access among both. Because DNS64 is then required, DNSSEC validation will be only possible if the recursive name server is validating the negative response from the authoritative name server and the client is not performing validation.
Note that is not expected at this stage of the transition, that applications or operating systems are IPv6-only, and it will not be a sensible decision for a developer to work on that direction, unless it is clear that the deployment scenario allows it. On the other hand, an end-user or enterprise network may decide to run IPv6-only in the LANs, but in case there is any chance for applications to be IPv6-only, the operating system may be responsible either for doing a local address synthesis, or alternatively, setting up some kind of on-demand VPN (IPv4-in-IPv6), which need to be supported by that network. This may become very common in enterprise networks, where "Unique IPv6 Prefix per Host" [RFC8273].
However, when the client device is dual-stack and/or connected in a dual-stack LAN by means of a CLAT (or has the built-in CLAT), DNS64 is an option.
Note that the extra translation, when DNS64 is not used, takes place at the CLAT, which means no extra overhead for the operator, and no perceptible impact for a CE in a broadband network, while it may have some impact in a battery powered device. This cost for a battery powered device, is possibly comparable to the cost when the device is doing a local address synthesis (see Section 7.1 of [RFC8305]).
When clients, in a service provider network, use DNS servers from other networks, for example manually configured by users, they may support or not DNS64, so the considerations in Section 4.3 will apply as well.
Even in the case that the external DNS supports DNS64 function, we may be in the situation of providing incorrect configurations parameters, for example un-matching WKP or NSP, or a case such the one described in Section 3.2.3.
A similar situation may happen in case of split DNS scenarios, for example, when using a VPN that forces all the DNS queries thru the VPN, ignoring the DNS64.
Having a CLAT, even if using an external DNS without DNS64, ensures that everything will work, so the CLAT must be considered as an advantage against user configuration errors.
However, it needs to be reinforced, that if there is not a CLAT (scenarios without 464XLAT), an external DNS without DNS64 support, will disallow any access to IPv4-only networks, and will not guarantee DNSSEC, so will behave as in the Section 3.2.1.
If clients use mechanisms for DNS privacy, such as DNS over TLS ([RFC7858]), DNS over DTLS ([RFC8094]), DNS queries over HTTPS ([RFC8484]) or DNS over QUIC ([I-D.huitema-quic-dnsoquic]), as they may provide different results to the same query, it must be expected equivalent effects as described in Section 4.4.
As already indicated in precedent sections, the successful use of the DNS64 is not guaranteed when networks or hosts can use "split-DNS" (also called Split Horizon), private DNS. Section 4. of [RFC6950], analyses this case. This a very common situation when using VPNs.
[RFC6052] (IPv6 Addressing of IPv4/IPv6 Translators), Section 3, discusses some considerations which are useful to decide if an operator should use the WKP or an NSP.
Taking in consideration that discussion and other issues, we can summarize the possible decision points as:
A hosts or application using literal IPv4 addresses or older APIs, behind a network with IPv6-only access, will not work unless a CLAT (or equivalent function) is present.
A possible alternative approach is described as part of Happy Eyeballs v2 Section 7.1 ([RFC8305]). When HEv2 is not supported, one more alternative is using Bump-in-the-Host ([RFC6535]), and then a DNS64 function.
Those alternatives will solve the problem for and end-hosts, however, if that end-hosts is providing "tethering" or an equivalent service to others hosts, that need to be considered as well. In other words, in a case of a cellular network, it resolves the issue for the UE itself, but may be not the case for hosts behind it.
Otherwise, 464XLAT is the only valid approach to resolve this issue.
An IPv4-only hosts or application behind a network with IPv6-only access, will not work unless a CLAT is present. 464XLAT is the only valid approach to resolve this issue.
As described in Section 6.3 of [RFC6877] (IPv6 Prefix Handling), if the CLAT can be configured with a dedicated /64 prefix for the NAT46 translation, then it will be possible to do a more efficient stateless translation.
However, if this dedicated prefix is not available, the CLAT will need to do a stateful translation, for example performing stateful NAT44 for all the IPv4 LAN packets, so they appear as coming from a single IPv4 address, and then in turn, stateless translated to a single IPv6 address.
One possible setup, in order to maximize the CLAT performance, is to configure the dedicated translation prefix. This can be easily achieved automatically, if the broadband CE or end-user device is able to obtain a shorter prefix by means of DHCPv6-PD ([RFC8415]), or other alternatives, so the CE can use a specific /64 for the translation. This is also possible when broadband is provided by a cellular access.
The above recommendation is often not possible for cellular networks, when connecting smartphones (as UEs), as generally they don't use DHCPv6-PD ([RFC8415]) an instead a single /64 is provided for each PDP context and use /64 prefix sharing ([RFC6877]). So, in this case, the UEs typically have a build-in CLAT client, which is doing a stateful NAT44 before the stateless NAT46.
Explicit Address Mappings for Stateless IP/ICMP Translation [RFC7757] provides a way to configure explicit mappings between IPv4 and IPv6 prefixes of any length. When this is used, for example in a CLAT, it may provide a simple mechanism in order to avoid traffic flows between IPv4-only nodes or applications and dual-stack destinations to be translated twice (NAT46 and NAT64), by creating mapping entries with the GUA of the IPv6-reachable destination. This optimization of the NAT64 usage is very useful in many scenarios, including CDNs and caches, as described in [I-D.palet-v6ops-464xlat-opt-cdn-caches].
It can be argued that none of the possible transition mechanisms is perfect, and somehow, we may consider that actually this is a good thing as a way to push for the IPv6 deployment, or otherwise, it may be further delayed, with clear undesirable effects for the global Internet.
However, for an operator, being in business means minimizing the adverse transition effects, and provide the most perfect one, reasonably balanced with cost (CAPEX/OPEX) and at the same time, looking for a valid long-term vision.
NAT64/464XLAT has demonstrated to be a valid choice in several scenarios (IPv6-IPv4 and IPv4-IPv6-IPv4), with hundreds of millions of users, offering different choices of deployment, depending on each network case, needs and requirements.
Depending on those requirements, DNS64 may be a required function, while in other cases the adverse effects may be counterproductive. Similarly, in some cases NAT64, together with DNS64, may be a valid solution, when for sure there is no need to support hosts or applications which are IPv4-only (Section 4.8 and Section 4.9). However, in other cases the limitations they have, may suggest the operator to look into 464XLAT as a more complete solution.
Service providers willing to deploy NAT64, need to take into account the considerations of this document in order to better decide what is more appropriate for their own specific case.
In the case of broadband managed networks (CE provided or suggested/supported by the operator), in order to fully support the actual user needs (IPv4-only devices and applications, usage of literals and old APIs), they should consider the 464XLAT scenario and in that case, must support the customer-side translator (CLAT).
If the operator offers DNS services, in order to increase performance by reducing the double translation for all the IPv4 traffic, they may support DNS64 and avoid, as much as possible, breaking DNSSEC. In this case, if the DNS service is offering DNSSEC validation, then it must be in such way that it is aware of the DNS64. This is considered the simpler and safer approach, and may be combined as well with the other possible recommendations described in this document:
This "increased performance" approach has the disadvantage of potentially breaking DNSSEC for a small percentage of validating end-hosts versus the small impact of a double translation taking place in the CE. If CE performance is not an issue, which is the most frequent case, then a much safer approach is to not use DNS64 at all, and consequently, ensure that all the IPv4 traffic is translated at the CLAT (Section 4.3).
If DNS64 is not used, at least one of the alternatives described in Section 4.1.1, must be followed in order to learn he NAT64 prefix.
The operator needs to consider that if the user can modify the DNS configuration (which most probably is impossible to avoid), and instead of configuring a DNS64 choose an external regular DNS (non-DNS64), a scenario with only NAT64 will not work with any IPv4-only remote host, while it will continue working in the case of 464XLAT (Section 4.4). Same effects are to be expected if DNS privacy protocols are being used by customers (Section 4.5), as well as in the case of Split DNS (Section 4.6).
Similar considerations need to be taken regarding the usage of a NAT64 Well-Known-Prefix (WKP) vs Network-Specific Prefix (NSP) (Section 4.7), in the sense of, if using DNS64, they must match and, if the user can change the DNS configuration, they will, most probably, not match. If there is a CLAT and the users chosen DNS is not a DNS64, the network will keep working of other means of learning the NAT64 are available.
As described in Section 4.10 in broadband networks, it is recommended that CEs supporting CLAT, supports DHCPv6-PD ([RFC8415]), or alternative means for configuring a shorter prefix, and internally reserve one /64 for the stateless NAT46 translation. The operator must ensure that the customers get allocated prefixes shorter than /64 in order to support this optimization. One way or the other, this is not impacting the performance of the operator network.
Operators may follow Section 7 of [RFC6877] (Deployment Considerations), for suggestions in order to take advantage of traffic engineering requirements.
In the case of cellular networks, the considerations regarding DNSSEC may appear as out-of-scope, because UEs OSs, commonly don't support DNSSEC, however applications running on them may do, or it may be an OS "built-in" support in the future. Moreover, if those devices offer tethering, other client devices behind the UE, may be doing the validation, hence the relevance of a proper DNSSEC support by the operator network.
Furthermore, cellular networks supporting 464XLAT ([RFC6877]) and "Discovery of the IPv6 Prefix Used for IPv6 Address Synthesis" ([RFC7050]), allow a progressive IPv6 deployment, with a single APN supporting all types of PDP context (IPv4, IPv6, IPv4v6), in such way that the network is able to automatically serve every possible combinations of UEs.
If the operator chooses to provide validation for the DNS64 prefix discovery, it must follow the advice from Section 3.1. of [RFC7050] (Validation of Discovered Pref64::/n).
One last consideration, is that many networks may have a mix of different complex scenarios at the same time, for example, customers requiring 464XLAT, others not requiring it, customers requiring DNS64, others not, etc. In general, the different issues and the approaches described in this document can be implemented at the same time for different customers or parts of the network. That mix of approaches don't present any problem or incompatibility, as they work well together, being just a matter of appropriate and differentiated provisioning.
In an ideal world will, we could safely use DNS64, if the approach proposed in [I-D.bp-v6ops-ipv6-ready-dns-dnssec] is followed, avoiding the cases where DNSSEC may be broken. However, this will not solve the issues related to DNS Privacy and Split DNS.
The only 100% safe solution, which also resolves all the issues, will be, in addition to having a CLAT, not using a DNS64 but instead making sure that the hosts have a built-in address synthesis feature. Operators could manage to use the CLAT, however the built-in address synthesis feature is out of their control, and can only be resolved by operating system vendors.
Whenever feasible, using EAMT ([RFC7757]) as indicated in Section 4.11, provides a very relevant optimization, avoiding double-translations.
The recommendations of this document can be used as well in enterprise networks, campus and other similar scenarios (including managed end-user networks).
This include scenarios where the NAT64 (and/or DNS64) are under the control of that network (or can be configured manually according to that network specific requirements), and for whatever reasons, there is a need to provide "IPv6-only access" to any part of that network or it is IPv6-only connected to third party networks.
An example of that is the IETF meetings network itself, where a NAT64 and DNS64 are provided, presenting in this case the same issues as per Section 3.1.1. If there is a CLAT in the IETF network, then there is no need to use DNS64 and it falls under the considerations of Section 3.1.3. Both scenarios have been tested and verified already in the IETF network itself.
Next figures are only meant to represent a few of the possible scenarios, not pretending to be the only feasible ones.
The following figure provides an example of and IPv6-only enterprise network connected with dual-stack to Internet and using local NAT64 and DNS64.
+----------------------------------+
| Enterprise Network |
| +----------+ +----------+ | +----------+
| | IPv6 | | NAT64 | | | IPv4 |
| | only +--------+ + | +-------+ + |
| | LANs | | DNS64 | | | IPv6 |
| +----------+ +----------+ | +----------+
+----------------------------------+
Figure 14: IPv6-only enterprise with NAT64 and DNS64
The following figure provides an example of dual-stack (DS) enterprise network connected with dual-stack (DS) to Internet and using CLAT, without DNS64.
+----------------------------------+
| Enterprise Network |
| +----------+ +----------+ | +----------+
| | IPv6 | | | | | IPv4 |
| | + +--------+ NAT64 | +-------+ + |
| | CLAT | | | | | IPv6 |
| +----------+ +----------+ | +----------+
+----------------------------------+
Figure 15: DS enterprise with CLAT, DS Internet, without DNS64
Finally, the following figure provides an example of an IPv6-only provider with NAT64, and a dual-stack (DS) enterprise network by means of their own CLAT, without DNS64.
+----------------------------------+
| Enterprise Network |
| +----------+ +----------+ | +----------+
| | IPv6 | | | | IPv6 | |
| | + +--------+ CLAT | +--------+ NAT64 |
| | IPv4 | | | | only | |
| +----------+ +----------+ | +----------+
+----------------------------------+
Figure 16: DS enterprise with CLAT, IPv6-only Internet, without DNS64
This document does not have any new specific security considerations.
This document does not have any new specific IANA considerations.
Note: This section is assuming that https://www.rfc-editor.org/errata/eid5152 is resolved, otherwise, this section may include the required text to resolve the issue.
The author would like to acknowledge the inputs of Gabor Lencse, Andrew Sullivan, Lee Howard, Barbara Stark, Fred Baker, Mohamed Boucadair and TBD ...
Conversations with Marcelo Bagnulo, one of the co-authors of NAT64 and DNS64, as well as several emails in mailing lists from Mark Andrews, have been very useful for this work.
Christian Huitema inspired working in this document by suggesting that DNS64 should never be used, during a discussion regarding the deployment of CLAT in the IETF network.
This section summarizes how an operator may deploy an IPv6-only network for residential/SOHO customers, supporting IPv6 inbound connections, and IPv4-as-a-Service (IPv4aaS) by using 464XLAT.
Note that an equivalent setup could also be provided for enterprise customers. In case they need to support IPv4 inbound connections, several mechanisms, depending on specific customer needs, allow that.
Conceptually, most of the operator network could be IPv6-only (represented in the next pictures as "IPv6-only Internet"). This part of the network connects the IPv6-only subscribers (by means of IPv6-only access links), to the IPv6 upstream providers, as well as to the IPv4-Internet by means of the NAT64 (PLAT in the 464XLAT terminology).
The traffic flow from and back to the CE to services available in the IPv6 Internet (or even dual-stack remote services, when IPv6 is being used), is purely native IPv6 traffic, so no special considerations about it.
Looking at the picture from the DNS perspective, there are remote networks with are IPv4-only, and typically will have only IPv4 DNS (DNS/IPv4), or at least will be seen as that from the CE perspective. At the operator side, the DNS, as seen from the CE, is only IPv6 (DNS/IPv6) and has also a DNS64 function.
In the customer LANs side, there is actually one network, which of course could be split in different segments, and the most common setup will be those segments being dual-stack (global IPv6 addresses and [RFC1918] for IPv4, as usual in any regular residential/SOHO IPv4 network today). In the figure it is represented as tree segments, just to show that the three possible setups are valid (IPv6-only, IPv4-only and dual-stack).
.-----. +-------+ .-----. .-----.
/ IPv6- \ | | / \ / \
( only )--+ Res./ | / IPv6- \ .-----. / IPv4- \
\ LANs / | SOHO +--( only )--( NAT64 )--( only )
`-----´ | | \ Internet/ `-----´ \ Internet/
.-----. | IPv6 | \ / \ /
/ IPv4- \ | CE | `--+--´ `--+--´
( only )--+ with | | |
\ LANs / | CLAT | +---+----+ +---+----+
`-----´ | | |DNS/IPv6| |DNS/IPv4|
.-----. +---+---+ | with | +--------+
/ Dual- \ | | DNS64 |
( Stack )------| +--------+
\ LANs /
`-----´
Figure 17: CE setup with built-in CLAT with DNS64
In addition to the regular CE setup, which will be typically access-technology dependent, the steps for the CLAT configuration can be summarized as:
A more detailed configuration approach is described in [I-D.ietf-v6ops-transition-ipv4aas].
The operator network needs to ensure that the correct responses are provided for the discovery of the PLAT prefix, as well as it is highly recommended follows [RIPE-690], in order to ensure that multiple /64s are available including the one needed for the NAT46 stateless translation.
The operator needs to understand other issues, described across this document, in order to take the relevant decisions. For example, if several NAT64 are needed in the context of scalability/high-availability, an NSP should be considered (Section 4.7).
More complex scenarios are possible, for example, if a network offers multiple NAT64 prefixes, destination-based NAT64 prefixes, etc.
If the operator decides not to provide DNS64, then this setup turns into the one in the following Figure. This will be also the setup that, if the user has changed the DNS and consequently is not using the operator DNS64, "it will be seen" from the perspective of the CE.
.-----. +-------+ .-----. .-----.
/ IPv6- \ | | / \ / \
( only )--+ Res./ | / IPv6- \ .-----. / IPv4- \
\ LANs / | SOHO +--( only )--( NAT64 )--( only )
`-----´ | | \ Internet/ `-----´ \ Internet/
.-----. | IPv6 | \ / \ /
/ IPv4- \ | CE | `--+--´ `--+--´
( only )--+ with | | |
\ LANs / | CLAT | +---+----+ +---+----+
`-----´ | | |DNS/IPv6| |DNS/IPv4|
.-----. +---+---+ +--------+ +--------+
/ Dual- \ |
( Stack )------|
\ LANs /
`-----´
Figure 18: CE setup with built-in CLAT without DNS64
In this case, the discovery of the PLAT prefix need to be arranged as indicated in Section 4.1.1.
In this case, the CE doesn't have a built-in CLAT, or the customer can choose to setup the IPv6 operator-managed CE in bridge mode (and optionally use its own external router), or for example, there is an access technology that requires some kind of media converter (ONT for FTTH, CableModem for DOCSIS, etc.), the complete setup will look as in the next figure. Obviously, there will be some intermediate configuration steps for the bridge, depending on the specific access technology/protocols, which should not modify the steps already described in the previous cases for the CLAT configuration.
+-------+ .-----. .-----.
| | / \ / \
| Res./ | / IPv6- \ .-----. / IPv4- \
| SOHO +--( only )--( NAT64 )--( only )
| | \ Internet/ `-----´ \ Internet/
| IPv6 | \ / \ /
| CE | `--+--´ `--+--´
| Bridge| | |
| | +---+----+ +---+----+
| | |DNS/IPv6| |DNS/IPv4|
+---+---+ +--------+ +--------+
|
.-----. +---+---+
/ IPv6- \ | |
( only )--+ IPv6 |
\ LANs / | Router|
`-----´ | |
.-----. | with |
/ IPv4- \ | CLAT |
( only )--+ |
\ LANs / | |
`-----´ | |
.-----. +---+---+
/ Dual- \ |
( Stack )------|
\ LANs /
`-----´
Figure 19: CE setup with bridged CLAT without DNS64
It should be avoided that several routers (i.e., the operator provided CE and a downstream user provided router) enable simultaneously routing and/or CLAT, in order to avoid multiple NAT44 and NAT46 levels, as well as ensuring the correct operation of multiple IPv6 subnets. In those cases, it is suggested the use of HNCP ([RFC8375]).
Note that the procedure described here for the CE setup, can be simplified if the CE follows [I-D.ietf-v6ops-transition-ipv4aas].
In addition to the regular set of features for a CE, a CLAT CE implementation requires support of:
There are several OpenSource implementations of CLAT, such as:
Several documents provide references to benchmarking, for example in the case of DNS64, [DNS64-Benchm].
Section to be removed after WGLC. Significant updates are:
Section to be removed after WGLC. Significant updates are:
Section to be removed after WGLC. Significant updates are: