Requirements for IPv6 Customer Edge Routers to Support IPv4 Connectivity as-a-Service

This document defines IPv4 service continuity features over an IPv6-only network, for a residential or small-office router, referred to as an "IPv6 Transition CE Router", in order to establish an industry baseline for transition features to be implemented on such a router. These routers rely upon "Basic Requirements for IPv6 Customer Edge Routers" . The scope of this document is to ensure the IPv4 "service continuity" support, for devices in the LAN side. This ensures that remote IPv4-only services continue to be accessible, from an IPv6-only Internet Service Provider (ISP) access network from both, IPv4-only and IPv6-only applications and devices in the LAN side. These ISP access networks are typically referred to as Wide Area Networks (WANs), even if in some cases they may be metropolitan or regional. presents a simplified view of this architecture.

This document covers a set of IP transition techniques required when ISPs have, or want to have, an IPv6-only access network. This is a common situation when sufficient IPv4 addresses are no longer available for every possible customer and device, causing IPv4 addresses to become prohibitively expensive. This, in turn, may result in service providers provisioning IPv6-only WAN access. At the same time, they need to ensure that both IPv4-only and IPv6-only devices and applications in the customer networks can still reach IPv4-only devices and applications in the Internet. This document specifies the IPv4 service continuity mechanisms to be supported by an IPv6 Transition CE Router, and relevant provisioning or configuration information differences from . This document is not a recommendation for service providers to use any specific transition mechanism. Automatic provisioning of more complex topology than a single router with multiple LAN interfaces may be handled by means of HNCP (Home Networking Control Protocol), which is out of the scope of this document. Since it is impossible to know prior to sale which transition mechanism a device will need over its lifetime, an IPv6 Transition CE Router intended for the retail market MUST support all the IPv4aaS transition mechanisms listed in this document. Service providers who specify feature sets for the IPv6 Transition CE Router, may define a different set of features than those included in this document, for example supporting only some of the transition mechanisms enumerated in this document. A complete description of "Usage Scenarios" and "End-User Network Architecture" is provided in Annexes A and B, respectively, which together with , will facilitate the reader to have a clearer understanding of this document.

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 BCP14 when, and only when, they appear in all capitals, as shown here.

This document uses the same terms as in , with minor clarifications. "IPv4aaS" stands for "IPv4 as-a-Service", meaning transition technologies for delivering IPv4 in IPv6-only connectivity. The term "IPv6 transition Customer Edge Router with IPv4aaS" (shortened as "IPv6 Transition CE Router") is defined as an "IPv6 Customer Edge Router" that provides features for the delivery of IPv4 services over an IPv6-only WAN network, including IPv6-IPv4 communications. The "WAN Interface" term used across this document, defines an IPv6 Transition CE Router attachment to an IPv6-only link used to provide connectivity to a service provider network, including link Internet-layer (or higher layers) "tunnels", such as IPv4-in-IPv6 tunnels.

The IPv6 Transition CE Router MUST comply with (Basic Requirements for IPv6 Customer Edge Routers) and this document adds new requirements, as described in the following sub-sections.

A new LAN requirement is added, which in fact is common in regular IPv6 Transition CE Routers, and it is required by most of the transition mechanisms: The IPv6 Transition CE Router MUST implement a DNS proxy as described in (DNS Proxy Implementation Guidelines).

The main target of this document is the support of IPv6-only WAN access. To enable legacy IPv4 functionality, this document also includes the support of IPv4-only devices and applications in the customers LANs, as well as IPv4-only services on the Internet. Thus, both IPv4-only and the IPv6-only devices in the customer-side LANs of the IPv6 Transition CE Router are able to reach the IPv4-only services. Note that this document is only configuring the IPv4aaS in the IPv6 Transition CE Router itself, and not forwarding such information to devices attached to the LANs, so the WAN configuration, availability of native IPv4 or IPv4aaS, is transparent for them. This document takes no position on simultaneous operation of one or several transition mechanisms and/or native IPv4. In order to seamlessly provide IPv4 service continuity in the customer LANs, and allow automated IPv6 transition mechanism provisioning, the following general transition requirements are defined. General transition requirements: The IPv6 Transition CE Router MUST support the DHCPv6 S46 priority options described in (Unified IPv4-in-IPv6 Softwire Customer Premises Equipment (CPE): A DHCPv6-Based Prioritization Mechanism). The IPv6 Transition CE Router MUST have a GUI and either a CLI or API (or both) to manually enable/disable each of the supported transition mechanisms. If an IPv6 Transition CE Router supports more than one LAN subnet, the IPv6 Transition CE Router MUST allow appropriate subnetting and configuration of the address space among the several interfaces. In some transition mechanisms, this may require differentiating mappings/translations on a per-interface basis. In order to allow the service provider to disable all the transition mechanisms and/or choose the most convenient one, the IPv6 Transition CE Router MUST follow the following configuration steps: Request the relevant configuration options for each supported transition mechanisms, which MUST remain disabled at this step. Following Section 1.4 of , MUST check for a valid match in OPTION_S46_PRIORITY, which allows enabling/disabling a transition mechanism. Keep disabled all the transition mechanisms if no match is found between the priority list and the candidate list, unless a NAT64 prefix has been configured, in which case, 464XLAT MUST be enabled. Because 464XLAT has not DHCPv6 configuration options, it can't be included, at the time being, in the OPTION_S46_PRIORITY. In the future, an update of or a NAT64 DHCPv6 configuration option, may enable it. Meanwhile, if an operator provides 464XLAT, it needs to ensure that OPTION_S46_PRIORITY is not sent for any other transition mechanism to the relevant customers. The following sections describe the requirements for supporting each one of the transition mechanisms. An IPv6 Transition CE Router intended for the retail market MUST support all of them.

464XLAT is a technique to provide IPv4 service over an IPv6-only access network without encapsulation. This architecture assumes a NAT64 (Stateful NAT64: Network Address and Protocol Translation from IPv6 Clients to IPv4 Servers) function deployed at the service provider or a third-party network. The IPv6 Transition CE Router MUST support CLAT functionality if intended for the retail market. If 464XLAT is supported, it MUST be implemented according to . The following IPv6 Transition CE Router requirements also apply: 464XLAT requirements: Unless a dedicated /64 prefix has been acquired, either using DHCPv6-PD (IPv6 Prefix Options for DHCPv6) or by alternative means, the IPv6 Transition CE Router MUST perform IPv4 Network Address Translation (NAT) on IPv4 traffic translated using the CLAT. The IPv6 Transition CE Router SHOULD support IGD-PCP IWF (UPnP Internet Gateway Device - Port Control Protocol Interworking Function). If PCP is implemented, the IPv6 Transition CE Router MUST also implement (DHCP Options for the PCP). Following , if no PCP server is configured, the IPv6 Transition CE Router MAY verify if the default gateway, or the NAT64 is the PCP server. The IPv6 Transition CE Router MUST use plain IPv6 mode (i.e., no IPv4-in-IPv6 encapsulation is used) to send PCP requests to the server. The IPv6 Transition CE Router MUST implement (Discovery of the IPv6 Prefix Used for IPv6 Address Synthesis) in order to discover the PLAT-side translation IPv4 and IPv6 prefix(es)/suffix(es). If PCP is implemented, the IPv6 Transition CE Router MUST follow (Discovering NAT64 IPv6 Prefixes Using the PCP), in order to learn the PLAT-side translation IPv4 and IPv6 prefix(es)/suffix(es) used by an upstream PCP-controlled NAT64 device. The priority for the NAT64 prefix, in case the network provides several choices, MUST be: 1) , 2) . The NAT64 prefix could be discovered by means of only in the case the service provider uses DNS64 . It may be the case that the service provider does not use or does not trust DNS64 because the DNS configuration at the CE (or hosts behind the CE) can be modified by the customer. In that case, the service provider may opt to configure the NAT64 prefix by means of . This can also be used if the service provider uses DNS64 .

Dual-Stack Lite enables continued support for IPv4 services. Dual-Stack Lite enables a broadband service provider to share IPv4 addresses among customers by combining two well-known technologies: IP in IP (IPv4-in-IPv6) and Network Address Translation (NAT). It is expected that DS-Lite traffic is forwarded over the IPv6 Transition CE Router's native IPv6 WAN interface, and not encapsulated in another tunnel. The IPv6 Transition CE Router MUST implement DS-Lite B4 functionality if intended for the retail market. If DS-Lite is supported, it MUST be implemented according to . The following IPv6 Transition CE Router requirements also apply: DS-Lite requirements: The IPv6 Transition CE Router MUST support configuration of DS-Lite via the DS-Lite DHCPv6 option (DHCPv6 Option for Dual-Stack Lite). The IPv6 Transition CE Router MAY use other mechanisms to configure DS-Lite parameters. Such mechanisms are outside the scope of this document. The IPv6 Transition CE Router SHOULD support IGD-PCP IWF (UPnP Internet Gateway Device - Port Control Protocol Interworking Function). If PCP is implemented, the IPv6 Transition CE Router SHOULD implement (DHCP Options for the PCP). If PCP is implemented and a PCP server is not configured, the IPv6 Transition CE Router MUST assume, by default, that the AFTR is the PCP server. The IPv6 Transition CE Router MUST use plain IPv6 mode (i.e., no IPv4-in-IPv6 encapsulation is used) to send PCP requests to the server. The term "default" above is to be interpreted as pertaining to a configuration as applied by a vendor, prior to the administrator changing it for its initial activation. The IPv6 Transition CE Router MUST NOT perform IPv4 Network Address Translation (NAT) on IPv4 traffic encapsulated using DS-Lite .

lw4o6 specifies an extension to DS-Lite which moves the NAPT function from the DS-Lite tunnel concentrator to the tunnel client located in the IPv6 Transition CE Router, removing the requirement for a CGN (Carrier Grade NAT, AFTR - Address Family Transition Router) function in the tunnel concentrator and reducing the amount of centralized state. The IPv6 Transition CE Router MUST implement lwB4 functionality if intended for the retail market. If DS-Lite is implemented, lw4o6 SHOULD be implemented as well. If lw4o6 is supported, it MUST be implemented according to . The following IPv6 Transition CE Router requirements also apply: lw4o6 requirements: The IPv6 Transition CE Router MUST support configuration of lw4o6 via the lw4o6 DHCPv6 options (DHCPv6 Options for Configuration of Softwire Address and Port-Mapped Clients). The IPv6 Transition CE Router MAY use other mechanisms to configure lw4o6 parameters. Such mechanisms are outside the scope of this document. The IPv6 Transition CE Router MUST support the DHCPv4-over-DHCPv6 (DHCP 4o6) transport described in (DHCPv4-over-DHCPv6 Transport). The IPv6 Transition CE Router MAY support Dynamic Allocation of Shared IPv4 Addresses as described in (Dynamic Allocation of Shared IPv4 Addresses).

MAP-E is a mechanism for transporting IPv4 packets across an IPv6 network using IP encapsulation, including an algorithmic mechanism for mapping between IPv6 and IPv4 addresses. The IPv6 Transition CE Router MUST support MAP-E CE functionality if intended for the retail market. If MAP-E is supported, it MUST be implemented according to . The following IPv6 Transition CE Router requirements also apply: MAP-E requirements: The IPv6 Transition CE Router MUST support configuration of MAP-E via the MAP-E DHCPv6 options (DHCPv6 Options for Configuration of Softwire Address and Port-Mapped Clients). The IPv6 Transition CE Router MAY use other mechanisms to configure MAP-E parameters. Such mechanisms are outside the scope of this document. The IPv6 Transition CE Router MAY support Dynamic Allocation of Shared IPv4 Addresses as described in (Dynamic Allocation of Shared IPv4 Addresses).

MAP-T is a mechanism similar to MAP-E, differing from it in that MAP-T uses IPv4-IPv6 translation, instead of encapsulation, as the form of IPv6 domain transport. The IPv6 Transition CE Router MUST support MAP-T CE functionality if intended for the retail market. If MAP-T is supported, it MUST be implemented according to . The following IPv6 Transition CE Router requirements also apply: MAP-T requirements: The IPv6 Transition CE Router MUST support configuration of MAP-T via the MAP-T DHCPv6 options (DHCPv6 Options for Configuration of Softwire Address and Port-Mapped Clients). The IPv6 Transition CE Router MAY use other mechanisms to configure MAP-T parameters. Such mechanisms are outside the scope of this document. The IPv6 Transition CE Router MAY support Dynamic Allocation of Shared IPv4 Addresses as described in (Dynamic Allocation of Shared IPv4 Addresses).

Existing IPv4 deployments support IPv4 multicast for services such as IPTV. In the transition phase, it is expected that multicast services will still be provided using IPv4 to the customer LANs. If the IPv6 Transition CE Router supports delivery of IPv4 multicast services, then it MUST support (Delivery of IPv4 Multicast Services to IPv4 Clients over an IPv6 Multicast Network) and (DHCPv6 Option for IPv4-Embedded Multicast and Unicast IPv6 Prefixes).

If the UPnP WANIPConnection:2 service is enabled on a CE router, but cannot be associated with an IPv4 interface established by an IPv4aaS mechanism or cannot determine which ports are available, an AddPortMapping() or AddAnyPortMapping() action MUST be rejected with error code 729 "ConflictWithOtherMechanisms". Port availability could be determined through PCP or access to a configured port set (if the IPv4aaS mechanism limits the available ports). An AddPortMapping() request for a port that is not available MUST result in "ConflictInMappingEntry". An AddAnyPortMapping() request for a port that is not available SHOULD result in a successful mapping with an alternative "NewReservedPort" value from within the configured port set range, or as assigned by PCP as per , Section 5.6.1. Note that IGD:1 and its WANIPConnection:1 service have been deprecated by OCF (Open Connectivity Foundation).

This document doesn't include support for 6rd , because it is an IPv6-in-IPv4 tunneling. Regarding DS-LITE , this document includes slightly different requirements, related to the support of PCP , IGD-PCP IWF and the prioritization of the transition mechanisms, including dual-stack.

At the time of this writing, one of the apparent main issues for vendors to include new functionalities, such as support for new transition mechanisms, is the lack of space in the flash (or equivalent) memory. However, it has been confirmed from existing open source implementations (OpenWRT/LEDE, Linux, VPP, others), that adding the support for the new transition mechanisms, requires around 10-12 Kbytes, because most of the code base is shared among several transition mechanisms, which are already supported by . A single data plane is common to all them, which typically means, in popular CEs already in the market , the new required code is only about 0.15% of the total existing code size. In general, the new requirements don't have extra cost in terms of RAM memory, nor other hardware requirements such as more powerful CPUs, if compared to the cost of NAT44 code, so existing hardware should be able to support all them with minimal impact. The other issue seems to be the cost of developing the code for those new functionalities. However, at the time of writing this document, it has been confirmed that there are several open source versions of the required code for supporting all the new transition mechanisms, and several vendors already have implementations and provide it to ISPs, so the development cost is negligible, and only integration and testing cost may become an issue. Finally, in some cases, operators supporting several transition mechanisms may need to consider training costs for staff in all the techniques for their operation and management, even if this is not directly caused by supporting this document, but because the business decisions behind that.

The IPv6 Transition CE Router must comply with the Security Considerations as stated in , as well as those stated by each transition mechanism implemented by the IPv6 Transition CE Router. As described in and Security Consideration sections, there are generic DHCP security issues, which in the case of this document means that malicious nodes may alter the priority of the transition mechanisms. Access network architecture for securing DHCP within the access network is out of scope of this document. Securing DHCP in the LAN is also not in scope. DHCP packets MUST NOT be forwarded between LAN and WAN interfaces of an IPv6 Transition CE router.

This document does not have any new specific IANA considerations.

Thanks to Mikael Abrahamsson, Fred Baker, Mohamed Boucadair, Brian Carpenter, Ian Farrer, Lee Howard, Richard Patterson, Barbara Stark, Ole Troan, James Woodyatt, Lorenzo Colitti and Alejandro D'Egidio, for their review and comments in this and/or previous versions of this document, as well as to the Last Call reviewers by the Ops-dir (Dan Romascanu), Sec-dir (Christian Huitema), Rtg-dir (Daniele Ceccarelli), Tsv-art (Martin Stiemerling), Gen-art (Matthew Miller) and IESG (Alissa Cooper, Benjamin Kaduk, Suresh Krishnan, Ben Campbell, Spencer Dawkins, Mirja Kuhlewind, and Adam Roach).

The situation previously described, where there is ongoing IPv6 deployment and lack of IPv4 addresses, is not happening at the same pace in every country, and even within every country, for every ISP. For different technical, financial, commercial/marketing and socio-economic reasons, each network is transitioning at their own pace; the global transition timings cannot be reliably estimated. Different studies (for example ) also show that the IPv6 deployment is a changing situation. In a single country, not all operators will necessarily provide IPv6 support. Consumers may also switch ISPs, and use the same IPv6 Transition CE Router with either an ISP that provides IPv4-only or an ISP that provides IPv6 with IPv4aaS. So, to cover all those evolving situations, an IPv6 Transition CE Router is required, at least from the perspective of the transition support. Moreover, because some services will remain IPv4-only for an undetermined time, and some service providers will remain IPv4-only for an undetermined period of time, IPv4 will be needed for an undetermined period of time. There will be a need for CEs with support "IPv4 as-a-Service" for an undetermined period of time. This document, based on those premises, ensures that the IPv6 Transition CE Router allows the continued transition from networks that today may provide access with dual-stack or IPv6-in-IPv4, as described in , and as an "extension" to it, evolve to an IPv6-only access with IPv4-as-a-Service. Considering that situation and different possible usage cases, the IPv6 Transition CE Router described in this document is expected to be used typically, in residential/household, Small Office/Home Office (SOHO) and Small/Medium Enterprise (SME). Common usage is any kind of Internet access (web, email, streaming, online gaming, etc.) and even more advanced requirements including inbound connections (IP cameras, web, DNS, email, VPN, etc.). The above is not intended to be a comprehensive list of all the possible usage cases, just an overall view. In fact, combinations of the above usages are also possible, as well as situations where the same CE is used at different times in different scenarios or even different with services providers that may use a different transition mechanism. The mechanisms for allowing inbound connections are "naturally" available in any IPv6 router when using GUA (IPv6 Global Unicast Addresses), unless they are blocked by firewall rules, which may require some manual configuration. However, in the case of IPv4aaS, because the usage of private addresses and NAT and even depending on the specific transition mechanism, inbound connections typically require some degree of more complex manual configuration such as setting up a DMZ, virtual servers, or port/protocol forwarding. In general, IPv4 CE Routers already provide a GUI, CLI or API to manually configure them, or the possibility to setup the CE in bridge mode, so another CE behind it takes care of that. The requirements for that support are out of the scope of this document. It is not relevant who provides the IPv6 Transition CE Router. In most of the cases it is the service provider, and in fact they are responsible, typically, of provisioning/managing at least the WAN side. Commonly, the user has access to configure the LAN interfaces, firewall, DMZ, and many other features. However, in many cases, the user must supply or may replace the IPv6 Transition CE Router. This underscores the importance of the IPv6 Transition CE Routers fulfilling the same requirements defined in this document. The IPv6 Transition CE Router described in this document is not intended for usage in other scenarios such as large Enterprises, Data Centers, Content Providers, etc. So even if the documented requirements meet their needs, they may have additional requirements, which are out of the scope of this document.

According to the descriptions in the preceding sections, an end-user network will likely support both IPv4 and IPv6. It is not expected that an end-user will change their existing network topology with the introduction of IPv6. There are some differences in how IPv6 works and is provisioned; these differences have implications for the network architecture. A typical IPv4 end-user network consists of a "plug and play" router with NAT functionality and a single link upstream, connected to the service provider network. From the perspective of an "IPv4 user" behind an IPv6 transition Customer Edge Router with IPv4aaS, this doesn't change. However, while a typical IPv4 NAT deployment by default blocks all incoming connections and may allow opening of ports using a Universal Plug and Play Internet Gateway Device (UPnP IGD) or some other firewall control protocol, in the case of an IPv6-only access and IPv4aaS, that may not be feasible depending on specific transition mechanism details. PCP (Port Control Protocol, ) may be an alternative solution. Another consequence of using IPv4 private address space in the end-user network is that it provides stable addressing; that is, it doesn't change, even when you change service providers, and the addresses are always usable even when the WAN interface is down or the customer edge router has not yet been provisioned. In the case of an IPv6-only access, private IPv4 addresses are also available if the IPv4aaS transition mechanism keeps running the NAT interface towards the LAN side when the WAN interface is down. More advanced routers support dynamic routing (which learns routes from other routers), and advanced end-users can build arbitrary, complex networks using manual configuration of address prefixes combined with a dynamic routing protocol. Once again, this is true for both IPv4 and IPv6. In general, the end-user network architecture for IPv6 should provide equivalent or better capabilities and functionality than the current IPv4 architecture. The end-user network is a stub network, in the sense that is not providing transit to other external networks. However, HNCP allows support for automatic provisioning of downstream routers. illustrates the model topology for the end-user network.

This architecture describes the: Basic capabilities of the IPv6 Transition CE Router Provisioning of the WAN interface connecting to the service provider Provisioning of the LAN interfaces The IPv6 Transition CE Router may be manually configured in an arbitrary topology with a dynamic routing protocol or using HNCP . Automatic provisioning and configuration is described for a single IPv6 Transition CE Router only.

Section to be removed by RFC Editor. Significant updates are: ID-Nits: IANA section. ID-Nits: RFC7084 reference removed from Abstract. This document no longer updates RFC7084. UPnP section reworded. "CE Router" changed to "IPv6 Transition CE Router". Reduced text in Annex A.

Section to be removed by RFC Editor. Significant updates are: TRANS requirements reworked in order to increase operator control and allow gradual transitioning from dual-stack to IPv6-only on specific customers. New TRANS requirement so all the supported transition mechanisms are disabled by default, in order to facilitate the operator management. New TRANS requirement in order to allow turning on/off each transition mechanism by the user. Clarification on how to obtain multiple /64 for 464XLAT. S46 priority update to RFC8026 for including 464XLAT and related changes in several sections.

Section to be removed by RFC Editor. Significant updates are: RFC8026 update removed, not needed with new approach. TRANS and 464XLAT requirements reworded in order to match new approach to allow operator control on each/all the transition mechanisms. Added text in 464XLAT to clarify the usage.

Section to be removed by RFC Editor. Significant updates are: Several editorial changes across the document, specially TRANS requirements. DNS proxy MUST instead of SHOULD.

Section to be removed by RFC Editor. Significant updates are: Removed G-1. Added support for draft-pref64folks-6man-ra-pref64. General text clarifications.

Section to be removed by RFC Editor. Significant updates are: Reworded and shorter UPnP section and new informative reference. New general transition requirement in case multiple public IPv4 prefixes are provided, so to run multiple instances according to each specific transition mechanism. General text clarifications.

Section to be removed by RFC Editor. Significant updates are: Removed reference and text related to pref64folks-6man-ra-pref64. General text clarifications.

Section to be removed by RFC Editor. Significant updates are: Added text to UPnP section.

Section to be removed by RFC Editor. Significant updates are: Editorial edits.

Section to be removed by RFC Editor. Significant updates are: Changes related to suggestions by Ops-dir, Sec-dir, Rtg-dir, Tsv-art and Gen-art, as well as comments from IESG review. IANA section removed as a consequence of the removal of the inclusion of 464XLAT in the RFC8026 priority mechanism.