IPv6 routing issues

[Brian Carpenter CERN-CN]

Pedro,

I don't see anything in RFC 1992 that covers the semantics of
Nimrod EIDs when a transaction is rehomed (as I've just been
discussing with Ohta over on another list). Otherwise, I agree
that it's a perfectly good abstract definition. We don't have
a concrete definition though. And we don't have consensus
whether IDs should be squeezed into the regular address, or
separated. (Nimrod makes an abstract separation between locator
and identifier, but we need bit layouts.)

  Brian

>--------- Text sent by Pedro Roque follows:
> 
> >>>>> "Ran" == Ran Atkinson <rja@cisco.com> writes:
> 
>     Ran>   Similarly, I'm hearing major commercial customers express
>     Ran> dismay that IPv6 does not explicitly include support for
>     Ran> EIDs.
> 
> Maybe we should move the EID argument to a "Strong ES models for IPv6" mailing
> list. Also, there is already a full blown EID based soluction for IPv6, whose
> base architecture document was recently published as an RFC:
> 
> RFC 1992  I   I. Castineyra, J. Chiappa, M. Steenstrup, "The Nimrod Routing  
>               Architecture", 08/30/1996. (Pages=27) (Format=.txt) 
> 
> All in all, we can say the market as nothing to fear... At least not lack
> of proposals, for sure.
> 
> regards,
> ./Pedro.
>