Re: RIP over circuit switched media

[Gerry Meyer <[e-mail]>]

Art Berggreen <[e-mail]> wrote:

>I see this approach as having a very different philosophy and different
>mechanisms from most current DV routing protocols.

The philosophy is somewhat different, but I would like to think it is not
too great a leap that I won't pull you along.  If routing on the WAN is to
improve then the philosophy (or something) must change.

>But it is certainly not directly interoperable with exisiting protocols,
>making it a new protocol (albeit based on existing technology).

Sure, it requires consenting routers on the WAN.  You can't get something
for nothing.  Calling it a new protocol is going a bit far.  It is
essentially just RIP (evolution not revolution).

>IMHO, we also should be exploring ways of making link state protocols
>friendlier to WAN enviroments (because I consider them superior to DV
>protocols).

I think some of the general techniques could to a greater or lesser extent
be applied to other routing protocols.    I'm not biased either way.  There
has been some discussion on the ospf mailing list over the last few weeks.
But its still early days.

>It is clear that most of our existing routing protocols have evolved to
>favor LANS and relatively high performace, dedicated WAN links.  Routing
>protocols do need to evolve to deal with new technologies (e.g. Frame Relay),
>and new constraints (i.e. bandwidth, economics and policy).

Agreed.

    Gerry