RE: draft-kurtis-multihoming-longprefix comments

["Dmitri Krioukov" <dima@krioukov.net>]

There is this obvious thought that there might be
some theoretical results already obtained showing
that no change within model X would solve problem Y
since what we perceive as problem Y is in fact a
theorem in X. In other words, Y is just caused by X.

As the first step, there might be some lower bounds
for local memory requirements of routing schemes with
specific stretch factors...
--
dima.


> -----Original Message-----
> From: owner-multi6@ops.ietf.org [mailto:owner-multi6@ops.ietf.org]On
> Behalf Of Tony Li
> Sent: Tuesday, January 07, 2003 4:41 PM
> To: Iljitsch van Beijnum
> Cc: multi6@ops.ietf.org
> Subject: RE: draft-kurtis-multihoming-longprefix comments
> 
> 
> 
> 
> |   > The current routing technology does NOT scale properly.  
> |   There is not
> |   > going to be any intermeidate approach based on the 
> |   current technology
> |   > that will scale.
> |   
> |   Now I have to go and disagree with you there.
> |   
> |   With a reasonable level of aggregation current technology will scale
> |   just fine.
> |   
> |   Aggregation is simply a means to divide the routing 
> |   information over the
> |   routers so that a single router doesn't have to hold all possible
> |   routes. Today, if routers are owned by the same owner, they hold the
> |   same routing information. Maybe it's time we try 
> |   aggregating based on
> |   something else than router ownership.
> 
> 
> The current architecture has no incentive for end user sites to want
> to be aggregated and without that, the architecture is bound to fail.
> Right now, we have two large incentives not to aggregate: multihoming and
> more-specifics for traffc engineering.
> 
> We need to change something, moving forward.
> 
> Tony