> From: Scott Brim <[email protected]>
> NAT won't be everywhere, and .. it will go away eventually
One more time:
No, it won't, not unless the deployed architecture provides _all_ the
capabilities NAT now gives us (e.g. provider independence). As long as we
don't have separate location and identity namespaces, we by definition
can't have provider independence.
>> NAT66 stands as a potential solution, with its pros and cons,
>> independent of such a prediction.
> If NAT66 is everywhere then the RRG doesn't need to do much.
Only if the existing routing architecture (ERA - i.e. distributed
computation by passing destination vector tables around, with limited
built-in naming abstraction tools) does all we need.
I.e. i) only if the ERA can continue to handle the network as it gets
bigger (since it will, even if separation of location and identity
namespace allows us to do some gargage collection on the routing tables),
and ii) we don't need/want any additional capabilities out of the routing
(e.g. multiple metrics, or some user influence over paths, or whatever).
Noel
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
