On 2009-01-02 13:39, Tony Li wrote:
>
>
> |Certainly. So if you were product manager for a highly reliable
> |distributed application, I'm sure you would insist that it was
> |coded to detect permanent transport failures and try alternative
> |addresses. That may not be elegant computer science but it's one
> |way that the Internet routes around damage.
>
>
> And that again results in trying to fix the routing architecture at layers 7
> and 8. You'll pardon me if I don't find that a particularly satisfying
> 'solution'.
It isn't, which is why we're here, I think.
To be clear what I'm thinking, it seems that if nothing new comes out
of the routing world, we'll eventually hit the wall where PI-style
multihoming cannot grow. Then economics will rule it out except for
very large customers. At that point, medium or small customers who want
multihoming will either have to use standard multi-prefix PA IPv6, or
NAT66 (assuming IPv4 addresses are exhausted by then). In both cases,
non-SCTP transport sessions will break when multihoming fails over.
The applications community will resolve that either by supporting
address failover at layer 7 or by converting to SCTP (or some
other agile transport).
Brian
_______________________________________________
rrg mailing list
[email protected]
https://www.irtf.org/mailman/listinfo/rrg
