Re: [Isis-wg] [rbridge] Why is MTU discovery important?

[Dave Katz <dkatz at juniper.net>]

On Apr 14, 2009, at 7:41 AM, mike shand wrote:

> Yes, the piece which seems to be consistently missed out of this  
> discussion is that it is NOT as simple as using unpadded IS-IS  
> hellos and associated elements of procedure as the TRILL neighbor  
> discovery mechanism. As I understand things from this discussion,  
> TRILL requires a number of changes to the adjacency formation state  
> tables, which at the least require changes from an existing IS-IS  
> implementation, and probably make it a lot harder to have a single  
> implementation (with tweaks).

Well, "single implementation" is probably overstated to start with;   
"significant amounts of shared code" is perhaps a better description.   
But that's an implementation issue...

--Dave