Re: Question on anycast IID range(s)

[Fernando Gont <[e-mail]>]

Hi, Brian,

On 2/1/19 16:57, Brian E Carpenter wrote:
>> My question was ill-posed *and* contained a typo.  I'm really trying to
>> figure out
>> which range(s) of IIDs RFC 2526 is trying to reserve for anycast use.  I
>> now think
>> the answer is fdff:ffff:ffff:ff80-fdff:ffff:ffff:ffff based on RFC 5453 and
>> https://www.iana.org/assignments/ipv6-interface-ids/ipv6-interface-ids.xhtml
> 
> If I take RFC2526 literally, ffff:ffff:ffff:ff80-ffff:ffff:ffff:ffff
> is also reserved, for IIDs not in modified EUI-64 format. 

Indeed.



> But RFC7217 doesn't mention RFC2526, which might be a bug.

My recollection of writing this specific bit is that we were referencing
some RFC (certainly *not* this one), and then ended up pointing to the
IANA registry. IMO, the place to fix this is in the IANA registry,
possible with a note that such registration only applies to addresseses
not required to have an IID of 64 bits.

It would seem to me that we probably should have listed RFC2526 in the
"Updates" of RFC8064. Or, probably better, RFC7136 could/should have
updated RFC2526, too.

Essentially, RFC7136 says that IIDs have no internal structure (yeah, I
don't need to tell *you* ;-) ), and RFC7217/RFC8064 got rid of the specs
that might have created the expectation (in practical terms) for IIDs to
have some kind of internal structure.



> BTW, RFC2526 is unmotivated. Why do we *need* a convention for
> anycast IIDs?

I assume the motivation is actually to have "well known anycast
addresses". For that, you need:

* An IID range from which you'll pick them -- well, strictly speaking,
you don't, but if you're going to be selecting them from one specific
range, then *avoiding* them is easier and requires less code in things
like RFC7217.

* A registry where you list such well-known addresses.


RFC2526 seems to do a bit of both, but also gets into the convention for
the bits of IIDs, since it predates RFC7136/RFC7212/RFC8064.


I'd say we should probably do a clean-up of the relevant specs...
because at the time of this writing, e.g. all RFCs updated by RFC8064
still contain old text talking about Modified EUI-64 and such....


Thanks!

Cheers,
-- 
Fernando Gont
SI6 Networks
e-mail: [e-mail]
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492