Re: Thoughts for the second session in Vienna

[Iljitsch van Beijnum <iljitsch@muada.com>]

On dinsdag, mei 20, 2003, at 09:34 Europe/Amsterdam, Kurt Erik 
Lindqvist wrote:

> I would like to see what we do on the mailinglist for the architecture 
> before Vienna. I think that before the second session, it would be 
> good to try and list
> a) What we agree on in terms of architecture (I can go and dig through 
> the archives and try and make a list)
> b) Items of architecture that we think we need to discuss further 
> (Eriks and Iljitsch list might be a good start)
I would be especially interested in a good "pro" argument about tunnel 
headers. As I've said before, I don't think they're useful, but many 
people keep mentioning them. It would be good if we can reach a 
conclusion here.

Also, I suspect the mapping discovery/negotiation and the actual 
failover will be the hard parts in practice. The identifier<->locator 
stuff itself isn't that complicated once we agree on the best way to do 
this.

> If we manage to get some form of consensus on the above, we need to 
> decide on what do use it for, but I think that is best to spare some 
> 15 minutes at the end of the second session to discuss.
Maybe it would be a good idea to get together in a smaller group and 
see if we can hammer out something workable. A little whiteboard can go 
a long way. We could do this between the first and second session and 
then see if there is consensus on the result in the second session.