On 13 January 2017 at 11:49, Randy Bush <randy@xxxxxxx> wrote: >>> to be clear, i have no problem with iids being 64-bit. my issue is with >>> unicast globals being classful in 2.4.4. >> Randy I take your point, but this supposed conflict isn't new, it's not >> introduced in 4291bis, it goes back to RFC3513. > > i know; and i have pushed back every cm of the way. it took years to > get the other classful insanity, tls/nla, removed. the old cidr war > continues. this last bit of classfulness (excuse the word) too will > pass. > >> Do you have a suggestion how to change this within the context of >> advancing this to Internet Standard? > > yes. simply remove the mandatory requirement for classful global > unicast addresses. > I think people are combining together in this thread two separate things: - the addressing structure - how addresses are processed by routers when the router is forwarding a packet. In IPv6 they're separate things, in classful IPv4 they weren't (if I recall correctly). I think unicast IPv6 could only be described as classful if the structure of the address dictated how processing of addresses for forwarding occurred. (You could describe the difference between IPv6 unicast and multicast forwarding to be classful as forwarding is different for those two types of addresses.) BCP198/RFC7608, "IPv6 Prefix Length Recommendation for Forwarding" makes it very clear that forwarding is to occur based on the longest match rule of an IPv6 address, with no consideration of what the structure of the address happens to be for the purposes of device configuration or anything else. "Abstract IPv6 prefix length, as in IPv4, is a parameter conveyed and used in IPv6 routing and forwarding processes in accordance with the Classless Inter-domain Routing (CIDR) architecture. The length of an IPv6 prefix may be any number from zero to 128, although subnets using stateless address autoconfiguration (SLAAC) for address allocation conventionally use a /64 prefix. Hardware and software implementations of routing and forwarding should therefore impose no rules on prefix length, but implement longest-match-first on prefixes of any valid length." Regards, Mark.