>We have recently had to 'back peddle' on some of our pm's from ospf TO
>ripv1 due to a compatibility issue with USR's lack of OSPF capability
>(doh!). What suprises me, is, that with only RIPv1 on the pm's, how do
>they manage to broadcast the /28 mask (16 ip addresses) back and forth
>between the other Pm's, but our Cisco, listed below sees only the /32 (the
>single ip address) of hte connection.
This is explained in the Livingston Products Configuration Manual. Bottom
line is that they "assume" proper netmask when they see one Portmaster
advertise a whole class C that they have pool addresses for. They just
send out a host route for the part of the pool that they own. This is
using private knowledge and definitely is NOT an RIP feature. It is only
guarranteed to work with Livingston products, and certainly does not
work with Cisco. You are right in saying that Livingston is talking
an enhanced RIP V1. Since USR bought ComOS from Livingston, get them
to buy ComOS 3.5 from Livingston <g>. Livingston (as with most vendors)
bypassed RIP V2 and went straight to OSPF, avoiding the high broadcast
rates and slow convergence of RIP V2. Go back to OSPF, you'll be glad
ya did. And if ya have to talk RIP with the USR's, you can control that
by setting up a Cisco to redistribute the OSPF into RIP and vice versa
to talk to those guys. Make sure only one router is doing the
redistribution, and no other machine is listening/broadcasting two routing
protocols on the wire at the same time.