RIPv2 is an extension of RIPv1 and is supported only on IPv4 networks. RIPv2 shares the same basic algorithms of RIPv1, along with the support of several new features that provide the additional capability to support subnet mask information, which is required for the implementation of classless addressing needed by VLSM and CIDR. These new features, as described in RFC 2453, include subnet masks, next hop addresses, authentication, and multicasting. All RIPv1 messages are still supported in addition to the new RIPv2 message types.

The MCP TCP/IP RIPv2 implementation uses the subnet masks, next hop addresses, authentication, and multicast features in order to fulfill the requirement of classless addressing. RIPv2 responses update IP routing tables with information based on the route referenced in the IP address/subnet mask pair supplied in the RIPv2 message. Information regarding the next hop address of that route will also be updated if that information is supplied in the message. Finally, the RIP port is configured to listen for messages that are received on the IP multicast address 224.0.0.9. Use of this multicast address reduces the load on those hosts not configured to listen for multicast RIPv2 messages and prevents RIPv1-only hosts from having to process RIPv2 messages. Since periodic RIPv2 broadcasts use this address, an explicit request must be issued to receive datagrams using that address. It is a requirement that the multicast address be assigned via the MulticastAddressList connection attribute in the connection definition of the TCP initialization file.

Previously only one route to a destination was maintained in the routing table. Note that with the support of multiple parallel routes, all of the routes discovered by RIP will be added to the routing tables with a preference value based on the route metric.