IETF Lifts /48 Recommendation for End Sites
By stretch | Wednesday, March 30, 2011 at 3:42 a.m. UTC
RFC 6177, IPv6 Address Assignment to End Sites, was published this week. It seeks to reverse the recommendation originally put forth in RFC 3177 regarding the assignment of /48 IPv6 prefixes to all end sites. The effort behind the RFC seems to have stemmed primarily from observations that assigning a "one-size-fits-all" prefix to medium-sized organizations and residential users alike is excessively wasteful in the case of the latter.
RFC 6177 has obsoleted its predecessor, updating or negating the three core recommendations as such:
1) It is no longer recommended that /128s be given out. While there may be some cases where assigning only a single address may be justified, a site, by definition, implies multiple subnets and multiple devices.
2) RFC 3177 specifically recommended using prefix lengths of /48, /64, and /128. Specifying a small number of fixed boundaries has raised concerns that implementations and operational practices might become "hard-coded" to recognize only those fixed boundaries (i.e., a return to "classful addressing"). The actual intention has always been that there be no hard-coded boundaries within addresses, and that Classless Inter-Domain Routing (CIDR) continues to apply to all bits of the routing prefixes.
3) This document moves away from the previous recommendation that a single default assignment size (e.g., a /48) makes sense for all end sites in the general case. End sites come in different shapes and sizes, and a one-size-fits-all approach is not necessary or appropriate.
In summary, the IETF has decided that deciding on prefix sizes for end sites is "an issue for the operational community." However, section five of the RFC does offer some guidance well wroth reading for those writing internal IPv6 address allocation policies.
About the Author
Jeremy Stretch is a network engineer living in the Raleigh-Durham, North Carolina area. He is known for his blog and cheat sheets here at Packet Life. You can reach him by email or follow him on Twitter.
March 30, 2011 at 12:50 p.m. UTC
here nice best practices for IPv6 addressing plan which could be helpful
March 30, 2011 at 7:31 p.m. UTC
I think their suggestion of /56 is definitely reasonable. However I am concerned that some consumer ISP's might attempt to give out much smaller ranges of addresses to their customers (either through incompetence or as a way to differentiate "business" plans from "consumer" ones), which could end up making it difficult for people to segment their networks into /64 subnets. (And thus force people to look at using NAT again).
March 31, 2011 at 12:42 a.m. UTC
I can't help it but to see this as a bad move.
My concerns are the same as Olver's. This makes me doubt about the goal of promoting full end-to-end connectivity. For a residential user, a /56 might be ok, but for a small business with a simple Internet connectivity plan it might not guarantee scalability on the long term, which could imply renumbering.
March 31, 2011 at 6:00 p.m. UTC
I fail to see the concern that you both speak of ever occurring in giving businesses each a /64 bit address block out of the entire 128 bit address scope.
Figuring a /64 as a standard even you would have 18446744073709551616 networks with 18446744073709551616 addresses. I just can't see exhausting that address space so easily even excluding NAT. They are talking with /56 of 4722366482869645213696 addresses which is pretty much 256 times more than the /64 addressing.
If I'm misunderstanding what you are implying please forgive me.
March 31, 2011 at 7:38 p.m. UTC
@Mike: The way IPv6 works, a /64 prefix is generally considered a single network; it's beneficial to not use longer prefixes on multiaccess links so that autoconfiguration functionality is preserved. A /56 is equivalent to providing 256 prefixes, and a /64 65,536 prefixes.
April 1, 2011 at 11:06 a.m. UTC
I went to a ripe meeting when they were talking about IPv6 allocations. They were quite happy to recommend giving out /48's to end users. I failed to see the reason for this.
What on earth does a user need 65k subnets for?
Granted, even if you give every user in the world a /48 you've still got plently. But isn't engineering all about efficiency? We've got trillions, so lets get everyone a few million each.
A /56 seems more than enough for a home user. Even for a small business. However I might consider giving the business a /48
April 1, 2011 at 12:56 p.m. UTC
Yep, I totally disregarded the /64 boundary in any event there are a ton of addresses and like you stated either 256 or 65k different subnets with those addresses available..