Scenario 2 – External Nodes Communicate with Individual IBM Connect:Direct Servers
In this configuration, each external IBM® Connect:Direct® node communicates with a specific IBM Connect:Direct Server. This configuration makes better use of the IBM Connect:Direct/Plex environment by spreading the work from the external nodes between the IBM Connect:Direct Servers.
To accomplish this configuration, change the UNIX.CD2 network map to point to the APPLID for SERVER2 (callout 1).
Then, make the following changes to the IBM Connect:Direct/Plex network map (callout 2):
- Route all Processes from UNIX.CD to SERVER1 by defining SERVER1 as the default PLEXCLASS in the adjacent node definition for UNIX.CD.
- Route all Processes from UNIX.CD2 to SERVER2 by defining SERVER2 as the default PLEXCLASS in the adjacent node definition for UNIX.CD2.
The advantage of this approach is that work from each UNIX node runs on a different IBM Connect:Direct Server, so work from one node does not interfere with work from the other. You do not need to change Processes submitted from either node to run on the specified servers (unless the Process itself specifies a TCP/IP address).
The disadvantage of this approach is that you still cannot use the IBM Connect:Direct/Plex workload balancing to its full potential. You cannot perform workload balancing on Processes received from or sent to the external nodes. Work originating and done entirely on the Connect:Direct for z/OS® system can use IBM Connect:Direct/Plex workload balancing.