On 2008-12-24 16:20, William Herrin wrote:
> Draft #6 of the Summary of Routing Architectures discussed in the RRG
> is now available at:
>
> http://bill.herrin.us/network/rrgarchitectures.html
Grossly oversimplifying, strategy B is unworkable for IPv4
and very close to the plan of record for IPv6.
Robin has been eloquent about the problems with this strategy,
and I think that for legacy hosts those arguments are undeniable.
But people don't seem to have really absorbed that IPv6 stacks,
and applications updated to use them, already work with multiprefix
PA address assignments. This is a form of B. The negative aspect
is that transport sessions will fail when there's an address change.
There are measures that can correct that, long term and step by step.
The other point about this approach to strategy B is that it doesn't
break strategy A, if a little care is taken. That's because B is
really not a change in the IP model and is pretty much orthogonal
to the routing system. (Actually, that's exactly why it emerged as
the deployment model for IPv6.)
So, my conclusion is to forget about strategy B as far as IPv4
is concerned, but treat it as orthogonal to strategy A as far as
IPv6 is concerned. That doesn't mean the RRG has to do anything
about it, except bear in mind that IPv6 already supports a form
of strategy B, and move on.
Brian
_______________________________________________
rrg mailing list
[email protected]
https://www.irtf.org/mailman/listinfo/rrg
