In einer eMail vom 26.01.2009 19:35:37 Westeuropäische Normalzeit schreibt  
[email protected]:

> From: David Conrad <[email protected]>

>> The question to ask is whether LISP is an appropriate solution  to
>> the problems discussed at the IAB's October, 2006  Routing and
>> Addressing Workshop.
>> ...
>> Yet the LISP proponents still did not  adequately address these concerns.

> Perhaps the  reasons LISP proponents haven't adequately addressed those
> concerns is because they are not addressable with the underlying  LISP
> assumptions?

When thinking about the  routing, it's also important to remember that LISP,
in the sense of 'a  map-encaps scheme plus a mapping subsystem plus ancillary
glue' is likely  (almost certainly?) just a component, almost more of a piece
of necessary  substrate (to give us a way out of the current paint corner),
than a  full-blown new routing approach. (I've referred before to how LISP
reminded  me of simply the deployment plan for a new routing architecture I
worked  on.)

Yes, LISP may do some routing-like things (e.g. support  multi-homing, and
provide a certain amount of provider independence), but  there's a limit to
what you can do with simply separating location and  identity. 
Right.
 

At the  end of
the day, we're still working with BGP, and if you want to do more  with
_routing_ than BGP allows (e.g. control the paths for traffic  aggregates),
then you need something _else_ new.
Formerly I also thought, overcoming the DV algorithm required to replace  
BGP. But meanwhile I see it differently: The fact that BGP is all present is a  
big help and not a big obstacle. Routing without disseminating any user  r
eachability prefixes is certainly something_else_new. But it can be  enabled by 
enhancing BGP !

And  deploying that new thing would almost certainly start with....  deploying
something that looks a lot like LISP. Add that to the fact that  LISP gives
you a decent bang/buck _in and of itself_ (multi-homing, etc),  and you soon
decide to focus on getting LISP done (right, with hooks for  the future), and
worry about the rest later...
LISP will never get IP forwarding on the Moore's law applicability  track. 
That's for sure :-(
 
Heiner
 



Noel
_______________________________________________
rrg mailing  list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg



_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to