For SNA network interconnection (SNI), paths are defined the same as for single- and multiple-domain environments within each network (see Network routing for subarea nodes and How to plan routes in your network). Rules for explicit and virtual routes do not cross network boundaries. For example, in Figure 1 assume that you are setting up a session between LU1 and LU2 and that they are in the same network. The same explicit route that is used from HOST1 to GWNCPB (ER0) would have to be used from GWNCPB to HOST2. However, because HOST1 and HOST2 are in different networks, the two explicit route definitions are independent of each other.

Figure 1. Multiple-network paths

Paths do not cross network boundaries. Cross-network routing requires separate paths in each network, each having the gateway NCP as the endpoint. If you are combining networks, you might have to add paths in nongateway NCPs that go to the gateway NCP.

If you are splitting an existing network into two networks, redefine the paths so that each path passing through a gateway NCP is defined separately in each of the networks.