[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: draft-narten-ipv6-3177bis-48boundary-02.txt
On 13 jul 2006, at 15.36, Thomas Narten wrote:
Kurt,
3. Other RFC 3177 considerations
RFC3177 suggested that some multihoming approaches (e.g., GSE)
might
benefit from having a fixed /48 boundary. This no longer
appears to
be a consideration. There is no such requirement coming out of
the
IETF multi6 or shim6 efforts.
I think we can take the last sentence out. Mostly as we are seeing
several parallel solutions being pursued and deployed, such as shim6,
PI space and more specific PAs.
So here is the dilemma. I put in this sentence because 3177 argues
that /48 might be useful in the context of multihoming/GSE. A while
back, I asked multi6 whether this was the case and got a "no" back in
response. Pekka (in another thread) complains that this document
doesn't address all the points that 3177 did. So, if I take the above
out, then what can I (usefully) say about the above point?
Well, first of all there will never be a solution out of multi6 (And
the WG is to be closed :-) ), so that is an error. But I am more
worried about only addressing this particular solution while there
are clearly others being deployed. Note that I am not arguing about
the first two sentences quoted above i.e "RFC3177 suggested that some
multihoming approaches (e.g., GSE) might benefit from having a
fixed /48 boundary. This no longer appears to be a consideration.". I
think that should address moving 3177 to historic - no? Maybe you
could add "There are several solutions either being worked on or
deployed for multihoming today, neither of which would benefit from a
fixed /48 boundary".
I don't believe this should go into the document,
Why not? Is it incorrect, or just unnecessary?
See above. I think it exclusive and should be rewritten.
but I will also
offer the observation that in some of the RIR regions the end-site
allocation size have impacts on the LIRs relationship to the RIR -
and this 'rule' has the potential for creating artificial complexity
in that area.
what do you mean by "this rule"? That there is no longer a fixed /48
can be used? If so, I agree, but the RIRs are already heading in that
direction and I view it ias inevitable because we will eventually see
end sites that need more than a /48, in which case we have to figure
out how to count/measure/justify them.
By 'rule' I meant /48 for end-sites so I think we agree.
- kurtis -