Re: [RRG] Tunnel fragmentation/reassembly for RRG map-and-encaps architectures

[Yakov Rekhter <yakov@juniper.net>]

Dino,

> > Just for the record, CEF did not do any caching.  In fact, the whole  
> > point of the design was to stop caching, since it was ineffective.   
> > The working set was essentially the full space, so there was no  
> > benefit and significant overhead.
> 
> Well, you know I know that. But I was staying with Yakov's lingo. What  
> Yakov is referring to is cisco's original fast-switching cache where  
> host addresses were populated by software switching.
> 
> It's not at all bad with LISP because we are aggregating all site- 
> based addresses into an EID-prefix so any new host at a site with a  
> cached EID-prefix for a destination site that talks to a new host at  
> the destination site, won't have a "cache fault" or have to drop or  
> send data probes.

Whether "it's not at all bad" depends on the locality of the traffic.
To preserve "truth in advertising" the LISP spec needs (a) to clearly
spelled out the assumptions you make about the locality of the
traffic, as well as (b) to document the system behavior if these assumptions 
turned out to be incorrect.

Yakov.

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg