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

Re: tunneling [Was: Agenda for Vienna]



On woensdag, mei 21, 2003, at 10:23 Europe/Amsterdam, Tim Chown wrote:

Didn't we kinda sorta agree that loc/id is our collective favorite
architecture and that the rest is either flawed, should be worked on
elsewhere or no interest in working on it as a group?

So if you were drawing out a multihoming roadmap, where would loc/id
be on that roadmap?
I don't really want to use the words "long term" as that might make people think there is no point in working on it / waiting for it. But it should solve the problem rather definitively, so in that sense it is long term. I'm not entirely sure how long it takes from draft to implementation but I think a couple of years should be doable with some tail wind so in that sense intermediate term. After that another few years before general adoption. I don't think any solution that needs changes on both ends can do much better.

The question then is whether we need to create any swamp in the interim
for large enterprises for which Christian's approach is out of scope and
where loc/id (in the host or at the edge) is too far off.
I hear some people mention giving /32s to large enterprises. I'm not going to bore you with rants about why this is morally reprehensible but rather observe that if we give the big boys a portable prefix they won't have any reason to support loc/id, and the single homed crowd doesn't either because they don't need it to talk to the big multihomed sites, so that leaves the small multihomers without anyone to talk to multihomedly. We have to all be in the same boat to get this off the ground.

Might we for example agree what the multi6 group can do now to assist
multihoming for new IPv6 deployments, and then "split off" long term
approaches (either within the group or out to IRTF if long-term really
is just that)?   Should multi6 focus on what can be done now, or what
can be done in 3+ years?
Well, if you really want to start multihoming quickly you should do geo. :-)

Changing protocols takes time. The correlation between the amount of change and the amount of time is well below 1. Since there isn't any IPv6 deployment to speak of and there are much bigger fish yet to be fried before that will happen (where is my IPv6 root and TLD delegations?) I think we can afford to take the time to get it right. Or rather, we can't afford not to.

If we can steer clear of stuff like what the next generation namespace should look like I see no reason to involve the IRTF.