[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Another rebel meeting at 3.30, was: Re: My impressions of the Sunday meetings....




Second, almost everyone agreed that the final solution will need to be
or involve a new routing paradigm in one way or the other.
Yes. However, I'm also getting a sense that most people agree we can't
wait this long. Christian took some steps down an incremental steps path

Which is also what I concluded. We will need to do this in steps. If this is based on a host solution, my suggestion of PI space or something else remains to be seen. If we go for a solution that would require host or router implementation we still have at least two years to go. I don't think we have even that.

Another interesting development is that the v6ops wg just reached
consensus that it should work on non-routable/for-local-use-only address
provider independent address space that should be used as a replacement
for site local addresses for most uses. And immediately people started
yelling this address space should be routable after all.
I assume you mean the IPv6 WG? My opinion is that we should use global addresses everywhere but we "lost"...

However, I am really worried that we keep mixing a routing scalability issue with an addressing issue. I am not sure these needs to be related at all.

There is a lot more to be said on this issue but it belongs in the IPv6 WG not here.

So lets discuss this in the meeting this afternoon. As I'm out of
contact with the other "rebels" I'm not going to suggest a full agenda
right now, but this should definately be on it.

Well, I would prefer to keep the discussions around the document on the mailinglist (which of course doesn't stop us from discussing it anyway) as there have no announced multi6 wg meeting and to start "semi-offical" discussions with this short notice does not seem fair.

One way going forward that I prefer is to ask IANA delegate a
_temporary_ PI space prefix - with the _clear notion that this address
space will be called back_. This space would be allocated as "RIR
ALLOCATED PA" in RIPE terminology. This would give us a jump-start and
let some enterprises start out with trying multihoming, and IPv6 in
general. It will also give us more operational experience and perhaps
give us answers to some of my questions above. The additional prefixes
created is something we can handle (with the current ~250 non-6bone
prefixes I would say the problem is rather on the contrary), and with
clear guidelines on when and how the addresses will be called back this
should not be to hard to accept by the enterprises.
There are more ways to skin this cat. Now I haven't been a great fan of
what's happening with the requirements in this wg, but for this specific
purpose it might be useful to draw up some requirements.
Again, I think we need a requirements document that actually gives us directions on the solution in terms of capabilities and scalability as well listing the lessons currently learnt (which I think the current document is closer to). Then as said above and in other emails, I think we need to buy time and to me the easiest way to do this would be to go with what we have and that we know. I haven't seen anyone do a real analysis on the impact routing table growth with doing controlled address allocations.

- kurtis -