Re: draft-bourbaki-6man-classless-ipv6-00

[Fred Baker <[e-mail]>]

> On Jun 10, 2017, at 1:34 PM, David Farmer <[e-mail]> wrote:
> 
> I've been thinking about this for a while and I think the only way to resolve this conflict is to stop using "subnet prefix" in reference to IPv6 and only use it in reference to IPv4. 

Very much agree. That's the reason RFC 7608 calls it a "prefix length". Prefixes, of course, are a routing concept, and where you are in the network affects your perception of the prefix length immensely. RIRs are allocated prefixes by the IANA and in turn sub-allocate them to ISPs and PI prefix holders. ISPs in turn allocate prefixes to customers, and those customers and PI prefix holders end up at some point with LAN subnet prefixes or prefixes on p2p links, and hosts think in terms of /128. There may be other prefix lengths along the way.

The only place it makes sense to me to talk about a specific length for an IID is with SLAAC; if a network is allocating addresses manually or using DHCP, they are in control and can do whatever they like - I dare you to stop them. But the IID is the part that is never a prefix - the last N bits of a /128, and by convention 64 bits.