RE: Why is MTU an issue? (RE: Please read - proposed W G termination)

snip...

>> 
>> The MTU problem is  a n^2 problem but it is solved distributively... Each 
>> host decides which MTU to use for each one of the connections to the peers 
>> it has. In an N hosts fabric, each host may have (N-1) connections to 
>> peers. The MTU selection is trivial, it's the minimum between my local MTU 
>> and my peer MTU, negotiated at the time the connection is established.
>
> It should also be noted that one would not be limited to only one connection 
> pair since from a technology perspective, there is no reason why multiples 
> cannot be used for QoS purposes / bandwidth management purposes.  The same 
> applies to the UD method though few people have implemented it as such.

Yes, that is true. One may choose to create multiple connections between two
peers. IPoIB-CM allows this. The implementation can then choose to negotiate 
the same MTU across all connections between the two peers.

Vivek
>
> Mike