Re: [rrg] draft-narten-radir-problem-statement-05.txt

[[e-mail]]

In einer eMail vom 02.03.2010 01:19:31 Westeuropäische Normalzeit schreibt  
[e-mail]:

On  2010-03-02 12:36, [e-mail] wrote:
...
> Statement:   Neither LISP nor any of all the other submitted  solutions 
do 
>  reduce the number of routes
> - not even by the number 1.

IMHO  the issue is not to reduce the number of routes but to limit
their growth.  At the moment they seem to be limited loosely like
the square root of the  size of the Internet, and our goal is presumably
to limit them more  strongly than that -- log(N) would be lovely.

Well, TARA would reduce the number of stored routes to zero. If there are 1 
 million routes, it would reduce them by 1 million, if there were 100 
million  routes, it would reduce them by 100 million.
And in addition: it would provide a multitude of (detouring) routes of  
whichever of the taken figure which DV will never be able to provide.


There is a lot of speculation in this, but since the pressure  towards
route de-aggregation seems to come mainly from PI addressing and  the
demand for multihoming, a solution that enables PA aggregation  seems
certain to limit growth, compared to doing nothing. There are a  number
of solutions in the list that appear to do this.
See the other email from/to [e-mail]_ 
(mailto:[e-mail]) : This extra-factor due  to de-aggregation is taken care of by LISP, or 
isn't it ?
 
Heiner


Brian