Bad routes update

[Richard Draves]

>   First off, you're going to have a very hard time arguing 
> that the new /28s
> shouldn't be seen.

Quite right, I was forgetting the new /28s. So make that /28s.
(If we moved away from test addresses to real TLAs, then it would be /16s.)

> Second, there *must* be some provision for sites to
> multi-home without being a full pTLA, or pTLA allocations 
> will explode as will
> the allocations of pTLAs to people who shouldn't have them.

Yes, we spent two hours on this in Oslo - look for the minutes. But the
basic idea is to trade more scalable backbone routing for greater complexity
at the network periphery doing source & destination address selection. An
RFC 2260-style approach is another part of the proposed solution.

Rich