The Routing Over Large Clouds Mailing List Archive by date

[Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>]

> >> With regard to TTL, CSRs, and NHRP - decrementing TTL guarantees that
> >> if routing screws ups, packets won't loop forever.

> That problem may be better described as a conflict between
> link layer and network layer routing.

Unless you stick to NHRP, there is no conflict.

> >> If you have a box
> >> that participates in IP routing, but TTL doesn't decrement when
> >> packets traverse that box, then a bug in the routing code can result
> >> in infinitely looping packets.

> >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.

> This claim certainly is not true unless an extremely unsophisticated
> architecture for combining ATM switching and network layer
> routing is in use.

It is true that NHRP combines link layer routing and network layer
routing in an extremely unsophisticated manner. But, so what?

I'm afraid you misunderstand it IPNNI. No, it's just PNNI.

Wrongly implemented ATM swiches, which have nothing to do with IP,
generate a loop.

> >Hmmm, so, ATM lacking TTL, you think ATM switchs shouldn't
> >participate in PNNI routing.
> 
> >Use Ipsilon.
> 
> Ipsilon may not solve any problem.

As it, reportedly, supports IP TTL, those ATMers who can't accept
TTL-less cell-relaying (not me) should use it.

> >To me, policed ATM is good enough, because a transient loop of
> >a flow won't affect other flows.
> 
> Now, why should there even be a transient loop of flow?

Now? Read the mail you are responding:

	a bug in the routing code can result in infinitely looping packets.

						Masataka Ohta