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

["Manfredi, Albert E" <[e-mail]>]

-----Original Message-----
From: ipv6 [mailto:[e-mail]] On Behalf Of Ola Thoresen

> I am talking about the idea of deducing the prefix length for
> _Link_Local_ addresses from the RAs. If the link local address is
> randomly generated, which is the case in major OSes today, there is
> no way to ensure that the LL you created for one > /64 prefix
> matches the LL of a second router which also announces a prefix >
> /64 on the same link.

If two routers on a link offer prefixes of the same length to a given host, then the simple truncation algorithm I alluded to would generate the same IID to be used in both cases. For instance, "truncate the upper N bits of that random 128-bit value," where N is the prefix length.

If those two routers offered different prefix lengths, then clearly the IIDs used would have to be of different lengths. Is there a problem? I don’t see a problem. You should be able to have multiple addresses assigned to an interface with different lengths of prefix and IID, with no problem.

> Link local needs to be /64 to ensure that you can reach all possible
> routers on the link - unless you are going to add multiple LL
> addresses to an interface as well.

I don’t get that. If you start with the premise that all IIDs are to be 64 bits wide, THEN you can create a simple rule that the LLA IID and any SLAAC IID will end up being identical. But there's no other reason to assume that this is a requirement. It just follows as a consequence of that initial premise, which was based on the idea of EUI-64.

Leaving aside SLAAC itself, IPv4 is perfectly capable of dealing with multiple IP addresses assigned to an interface, with a different address mask per address. This creates no problems, in my experience.

Bert