RIP uses autosummarization for network advertisement by default. the RIPv2 network command only allows for a classful network as a parameter, which enables RIPv2 on all router interfaces that are part of that classful network.
Even if a subnet is entered as the network commands parameter the router will automatically compute the corresponding classful network address and store it in the configuration. Enabling RIPv2 on router interfaces allows that port to send RIPv2 updates, listens for RIPv2 updates on UDP 520, and advertises that interfaces connected subnet.
Since there is no way to match one interface at a time, you can instead configure some interface specific RIPv2 commands to disable the features that are turned on once the subnet is included in the RIP network table.
- Sending RIPv2 updates – you can disable this using the passive-interface command
- Listening for RIPv2 updates – you can filter updates either with a distribute list on the RIPv2 process or drop all RIP traffic on an inbound ACL on the interface
- Advertising the connected subnet – you can filter outbound advertisements with distribute lists.
To limit the devices that receive RIP updates on a multiaccess network, you can use the neighbor <ip address> RIP subcommand. This allows RIP to send unicast messages for updates as opposed to broadcast. you could put the interface into passive mode and use the specific neighbor RIP configurations only for the neighbors you want to receive RIPv2 update messages.
RIPv2 allows for discontiguous networks but autosummarization must be disabled for that to work.