| Standards Track |
| Experimental |
| Informational |
| Drafts |
| Deprecated |
6rd is a mechanism to facilitate IPv6 rapid deployment across IPv4 infrastructures of Internet service providers (ISPs).
It is derived from 6to4, a preexisting mechanism to transfer IPv6 packets over the IPv4 network, with the significant change that it operates entirely within the end-user's ISP's network, thus avoiding the major architectural problems inherent in the original design of 6to4. The name 6rd is a reference to both the rapid deployments of IPv6 it makes possible and, informally, the initials (RD) of its inventor, Rémi Després. A description of 6rd principles and of how they were first used by Free is published in RFC 5569.[1]
The detailed 6rd specification prepared for standardization in the IETF is available as RFC 5969.[2]
Contents |
In November 2007 Rémi Després[3]—who in the 1970s one of the creators of the Transpac data network in France—proposed to Free, the second largest ISP in France, to use the 6rd mechanism he had invented to rapidly deploy IPv6. While Free had until then no short term plan to offer IPv6 service, Rani Assaf, the CTO of Free, immediately decided to implement the solution. Five weeks later [4], with due marketing approval and operational validation, the press release [5] announcing that IPv6 was available to Free's customers was issued.
The first draft describing the 6rd mechanism and Free's deployment was submitted to IETF on 9 February 2008[6] After improvements, it was published on 24 January 2010 as informational RFC 5569.[1]
In March 2010 a Working Group of the IETF approved that its latest draft on 6rd should become, after some more modifications, a standards-track RFC. In August 2010 the standards-track RFC 5969.[2] was published. In October 2010 Comcast made 6rd software for home gateway devices available via open source for free.[7]
6to4 works by relaying traffic between native IPv6 and IPv4 using relay servers which have a common IPv6 prefix, but there is no guarantee that all native IPv6 hosts have a working route toward such a relay. Because of this a 6to4 host is not guaranteed to be reachable by all native IPv6 hosts. Besides, the operator of a 6to4 relay has no control of which hosts use it, which limits its incentive to maintain a good quality of service as traffic grows. With 6rd, this model is changed by making each ISP use one of its own IPv6 prefixes instead of the special 2002::/16 prefix standardized for 6to4.[8] With this, a provider is guaranteed that its 6rd hosts will be reachable from all native IPv6 addresses and, because there is no relay server to be used outside of its own control, it keeps full responsibility for the quality of service experienced by its own customers
Because 6rd relays can only be used by a limited set of hosts that are all under the control of the same administrative entity, it also reduces the scope for traffic anonymization attacks such as those possible with 6to4.[2]
The simplest 6rd deployment, which uses 32 bits of IPv6 address space to map the entire IPv4 address space, consumes more address space than typical with IPv6 natively supported in all ISP routers. This can be mitigated by omitting redundant parts of the IPv4 address space, and in some cases by deploying multiple 6rd domains.[1][2]
The default allocation of IPv6 space by an RIR is a 32-bit prefix. Since it takes 32 bits to map an IPv4 address with 6rd, this implies that an ISP would only be able to allocate 64-bit IPv6 prefixes to its customers if it were to use entire IPv4 addresses. 6rd, however, allows any redundant part of an IPv4 address to be discarded: For example, if the IPv4 addresses an ISP issues to its customers all share the same first eighteen bits, a 6rd prefix only need include the remaining fourteen bits. Without this flexibility, Free originally assigned 64-bit IPv6 prefixes to its customers but was able to assign them shorter prefixes once it obtained a larger allocation of IPv6 space (a 26-bit prefix) from the RIPE NCC.