[rrg] [lisp] Economic issues of long-path / stretched routing

[HeinerHummel at aol.com]

In einer eMail vom 26.01.2009 17:24:51 Westeuropäische Normalzeit schreibt  
dino at cisco.com:

>  Prior arguing with long-path, stretch etc. I have another  
>  informative question (maybe I missed something too during the  
>  discussion): Why LISP 1.5 ? Why not LISP 2 ? or directly:
> Why doesn't  the ITR intercept and enhance the DNS lookup as to also  
> request  the eRLOC address in addition to the dest IP address? It can  
>  also intercept the respective response and store (EID,  eRLOC).

OK. My concern is still how to get the geographical coordinates of the  
egress-(DFZ-)router


1)  Because not all communication uses DNS.
OK. There will be other ways, too, as to get the geo.coordinates


2)  Because DNS queries go out one xTR and replies can come in through   
another xTR.
This shouldn't bother neither solution, particularly not if the mentioned  
intercepting router is the ingress router nor in the ideal case when even the  
source host learns to handle  the returned geo.coordinates (e.g. by some  small 
Microsoft update).
 


3)  Because you don't want a architectural circular dependency between   
directory and routing.
OK. Wouldn't affect my solution.


Do  you want more reasons?
Please. I have always appreciated your answering of my  questions.
BTW, I am grateful for the default mapper mechanism, which helps me to  solve 
my incremental deployment issue.
Thank you so much.
 
Heiner
 
 
 
 
 



Dino



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.irtf.org/pipermail/rrg/attachments/20090126/8bf3cff4/attachment.htm>