Routing over Demand Circuits

[Gerry Meyer <[e-mail]>]

Tony,

>   Which is strightforward to handle (but unlikely to occur) in a UDP solution.
>   But is a major headache in a TCP solution - which has to be added to your
>   cost to implement a TCP solution.
>
>I found that it's trivial to implement, and looks almost identical to
>the UDP solution.  [Oops, I blew it.  Set a flag to remember to try
>this again.]
>
>   >Yes, BGP has to do this.  I still don't see that it's cost effective
>   >to do otherwise for RIP.
>
>   Because with prolonged delays and alternative routes it can lead to subtle
>   side-effects.
>
>   A bad solution is NOT a cost effective solution.
>
>This is not bad.  The delays are not "prolonged".  They are on the
>order of ms.  Certainly nothing to be concerned about in an X.25 net.

Msecs?  Wish I could use your X.25 net.

If there is a triggering event you are talking about sending an update to
each of the participating peer routers.  I think it would be nice to not
additionally clog up the link with stale TCP retransmissions contending
for VCs, thrashing the circuit manager, and otherwise delaying routing
datagrams.

>Let's not fix RIP as it's clearly not a better solution.

It is the *only* applicable solution today to routing many protocols
over demand circuits.

	Gerry