From: Mason Harris (MHarris@xxxxxxxxxx)
Date: Fri Jul 16 1999 - 12:38:56 GMT-3
Hey all--
First, thanks to Ben for that heads-up account on his experience at the RTP
lab. I live here in RTP and am scheduled for late October. Just as I was
feeling a wee bit confident......*bam* to hear about that. Well, I was
planning on heading to the beach this weekend, but I guess I need to get
back in the lab. :-)
Anyhow my question.....
I have been playing with RIP 1 and RIP2 configs in the last couple of days
and wanted to get some feedback about some things. First since RIP 1 sends
as RIP 1, but listens for RIP 1 and 2, I didn't see the need to configure an
interface running RIP 2 to send as RIP 1 because the RIP 1 interface (at the
other end of a point to point link) would pick up the route tables anyway
and redistribute automatically to its RIP 1 neighbors. So my first thougt
was "that's dumb why configure a RIP 2 interface to send as RIP 1 when RIP 1
will pick it up anyway?"
Ah, my naivety. However, last night when configuring RIP 1 and 2 interfaces
on shared media (Ethernet) I noticed something in the debug output that the
night before I had missed. Looks like it comes down to multicast vs.
broadcast updates. In my practice scenario on the Ethernet segment, some
devices needed to see the updates and some didn't. Obviously broadcasting
the updates wouldn't cut it.
It seems to me a major factor in choosing whether or not to send RIP 2 as
RIP 1 over an interface has to do with what kind of updates are required by
the design (multicast vs. broadcast). I know both can be set up to unicast
with the neighbor command, but I was curious if I was missing any nuances
between RIP 1 and RIP 2 (apart from the classful and classless routing
updates).
Any insight would be appreciated.
Thanks,
Mason
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:21:42 GMT-3