In einer eMail vom 16.02.2009 17:28:23 Westeuropäische Normalzeit schreibt  
[email protected]:

Hence,  if that argument is valid against host based solutions, then we 
are  completely out of luck.

Yours,
Joel



Strategy A: 
 LISP-ALT may be viewed as some outsourcing trick - however  staying inside 
the same inertial frame.
 
Strategy B:
Appears to be out of luck - see above.
 
Strategy C:
Besides overcoming  the scalability problem and making Moore's law  
applicable to routing etc., it would also enable  multihoming where  both  
network 
and/or  end users may impose path changes such that  different egress routers 
from 
different ISPs are passed.
 
Multihoming can be considered as a special kind of ANYCAST whereby it makes  
sense to change the anycast dest. node within the IP packet flow and also a  
single packet's anycast destination node in the middle of the path (because 
each  of these anycast destination nodes would deliver it to the same end user).
Network reasons like traffic jam or temporary drop outs may cause detours  
and a detour might be such that a different anycast dest. node becomes more  
suitable to go for.
Yes, these are not ECMP alternative paths but genuine detours which to  
offer/handle would be part of strategy C as well.
 
Heiner
 
 
 
 
 
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to