Illustrated TCP/IP by Matthew G. Naugle Wiley Computer Publishing, John Wiley & Sons, Inc. ISBN: 0471196568   Pub Date: 11/01/98

Chapter 111
Subnet Mask Field

This is what is making RIP stick around a little longer. Yes, it still has problems with scaling and still needs split horizon and poisoned reverse to operate properly. But to use the address scheme more efficiently, RIP version 2 now has the ability to support multiple subnet masks per network address. As we learned in other sections of this book, one of the biggest problems with RIP was its inability to support a subnet mask in the routing update. This led to the shortcoming of one subnet mask per network ID. Subnet masking really extends the life of RIP. RIP v1 does not indicate a subnet mask on a route entry. This can create many problems, two of which are learning and updating. How does RIPv1 know how to apply a subnet mask for a learned IP address? How does RIP provide a mask for its updates? Good questions! The answer is not real good though. RIP assumed that the IP address uses the same subnet mask as it does providing the IP network ID portion of the address is the same as its own and there is a subnet mask applied to its interface.

For example, a router has two interfaces: Interface 1 has an IP address of 130.1.1.1 and a subnet mask of 255.255.255.0. Interface 2 has an IP address of 205.1.1.1 and a subnet mask of 255.255.255.0. When interface 1 receives a routing update, any entry that has the same network ID as its own, 130.1.x.x, will apply the subnet mask that is configured to its port to those entries. You will see an entry in the routing table for that learned address. So, if a RIPv1 update was received on interface 1 and the update contained the entry 130.1.4.0, then the interface will record 130.1.4.0 in its routing table. However, if interface 1 received 155.1.1.0 on that interface, it would only place 155.1.0.0 into its table because it does not know the subnet mask for the address of 155.1.0.0. You must be safe when you assume.

Subnet Mask Field

When the router must transmit its table, how does it know to apply a mask to any of the entries in the table? It will depend on the interface from which it is transmitted. On interface 1, the router will transmit two entries: 150.1.1.0 and 200.1.1.0.

So, RIPv1 and subnet masks did not understand each other. RIPv2 fixes that. Notice in the slide that the format of the RIP datagram is preserved. The two fields in RIPv1 that stated “must be 0” are not used for the subnet mask and next-hop entries. Each route entry in the datagram will have an associated subnet mask with it. If the field contains a 0, there is not a subnet mask associated for the route entry. Also, in coordination with RIPv1 routers, a mask that is shorter than the Class’s natural mask should never be advertised.