8.1.6 Notes on Multiple-Protocol Simultaneous Operation

<Structure of this section>

(1) Concurrent operation of OSPF or RIP-2 and RIP-1

The routing protocols OSPF and RIP-2 use variable-length subnet masking that disregards the class (A, B, or C) of the IP address. In contrast, RIP-1 cannot use variable-length subnet masking because it recognizes the class of the destination address. For this reason, caution is required when deploying these routing protocols concurrently in the same network, as illustrated below. This section uses an example of the relationship between OSPF and RIP-1, but applies equally to the relationship between RIP-2 and RIP-1.

(a) When subnet routes learned by OSPF are not advertised by RIP-1

Note that RIP-1 does not advertise routes to a subnetted network when either of the following conditions is met:

The route is to a subnet whose subnet mask length differs from the network address of the interface using RIP.
The route is to a subnet whose network address differs from that of the interface using RIP.

<>Connections between subnets with different subnet mask lengths

In the following figure, Switch A registers the route to network B in its own routing table. At this time, because network B matches the first of the conditions given above, Switch A does not advertise its routes to network A by the RIP protocol.

Figure 8-2: Connections between subnets with different subnet mask lengths

For Switch A in Figure 8-5: Example of connection between subnets, Network A and Network B advertise routes for subnets of the same subnet length in the same network.

<>Connections between subnets with different network addresses

In the following figure, Switch A registers the route to network B in its own routing table, but because network B meets the second of the conditions given above, Switch A does not advertise the routes of network B to network A by the RIP protocol.

Figure 8-3: Connections between subnets with different network addresses

For Switch A in Figure 8-5: Example of connection between subnets, Network A and Network B advertise routes for subnets of the same subnet length in the same network.

(b) RIP Internetwork Connectivity with OSPF

You can use one of the following configurations to provide an OSPF connection between networks running RIP.

<>Disable a subnet

In the example in the following figure, routing information to Network A and Network B is advertised to Network B and network A, respectively.

Figure 8-4: Example of not using a subnet

<>Used for connections between subnets of the same subnet length in the same network.

In the example in the following figure, routing information to Network A and Network B is advertised to Network B and network A, respectively.

Figure 8-5: Example of connection between subnets

<>Advertise the default route.

Configure static routes with the default address as their destination in Switch A and Switch B, and advertise the routes to the networks running RIP.

In the example in the following figure, packets whose destination address does not match the local network reach Switch A and Switch B via the default routes, and are distributed to the other network via OSPF routes.

Figure 8-6: Example of advertising a default route

<>Advertises summary routes.

Switch A can summarize the routes to network B that it learns by OSPF/OSPF-ASE (OSPF route from external AS) into one natural mask route for advertisement to networks running RIP.

In the example in the following figure, by advertisement of the summarized route, packets destined for Network B reach Switch A and are distributed to Network B via OSPF/OSPF-ASE routes.

Figure 8-7: Example of advertising summary routes

(2) Notes on Learning Routes to the Same Destination by Multiple Protocols

In certain network topologies, routing loops can occur when more than one protocol learns a route to the same destination. To ensure that routing loops do not occur, use route filtering if required by your network topology.

In the network topology shown in the following figure, network 10.0.0.0 uses OSPF and network 10.1.0.0 uses RIP.

Figure 8-8: Example of network configuration

In this topology, the routes generated for network 10.2.0.0 can be categorized into three as follows:

External AS route advertised by router C ((a) in the figure)
Routes advertised from OSPF to RIP ((b) and (c) in the figure)
Routes advertised from RIP to OSPF ((d) and (e) in the figure)

In this example, a routing loop develops if Switch B selects (d) and Switch A selects (c), or if Switch A selects (e) and Switch B selects (b). That is, each router specifies the other as its next hop. In such a scenario, you need to configure route filtering such that routes to 10.2.0.0 advertised from OSPF to RIP by Switch A and Switch B are not learned as external AS routes advertised from RIP to OSPF.