The Routing Over Large Clouds Mailing List Archive by date

["Andrew G. Malis" <malis@nexen.com>]

Masataka,

> If you have a box
> that participates in PNNI routing, but TTL doesn't decrement when
> cells traverse that box, then a bug in the routing code can result
> in infinitely looping cells. Thus, a true ATM switch
> (with the emphasis on swich) decrement TTL, UNLESS it doesn't
> participate in the PNNI routing algorithm.
> 
> Hmmm, so, ATM lacking TTL, you think ATM switchs shouldn't
> participate in PNNI routing.

The above statement is incorrect.  PNNI routing is used only during VC
setup, and it is source routed - the entire path followed by the VC
setup request is determined by the source node.  To quote the PNNI
spec,

"In source routing, the source is responsible for selecting the path
to the destination. It does this based on its local knowledge of the
network topology. Since only one database is involved, loops are not
possible, nor will the paths be inefficient due to database
inconsistency."

Once the connection has been set up, of course, cell routing simply
follows the pre-established labels from hop to hop.  Again, no loops
are possible.

It's IP's hop-by-hop routing that makes TTL so important.  From
reading your email from over the weekend, I believe you and I are in
agreement on the importance of properly decrementing TTL.

Cheers,
Andy