The IP Over NBMA (ION) Archive

[Albert Manfredi <manfredi@engr05.comsys.rockwell.com>]

mohta@necom830.hpcl.titech.ac.jp wrote:
> 
> Yakov;
> 
> I can't see your points.
> 
> > > > Minimizing topological mismatch could reduce the need not just for
> > > > unicast cut-through, but for multicast cut-through as well.
> 
> Your point, here, shoud be not to fully align the topology
> but to minimize mismatch to make multicast cut-through
> unnecessary.
> 
> But, it makes unicast cut-through unnecessary.

Minimizing (abolishing) topological mismatch makes cut-through easy. It 
makes the _kludginess_ unnecessary.

There's still a difference between routing packets between points A and B 
via the IP layer and routing cells via the ATM layer, ignoring the IP 
layer entirely. If the topologies match and points A and B are both 
reachable over the ATM net, using the ATM net's native capabilities can 
provide advantages.

The real problem here, in my view, is the term "cut through." It implies 
something physical. Instead, it can just be "cutting through" the lines 
someone drew on a piece of paper.

Bert
manfredi@engr05.comsys.rockwell.com