Balancing resource demand

You can distribute the workload by distributing IMS applications and their associated transactions and terminals between the available processors. Depending on the complexity of the application and the capability of the processor, you can avoid overloading any single processor.

If the current design of the databases is such that the databases and their associated applications cannot be distributed across the available processors, you can:

• Duplicate inquiry-only databases; this allows more than one system to reference the whole of the database. (This is called vertical partitioning.)
• Split the databases into several component databases. (This is called horizontal partitioning.) The component databases must be completely independent for distribution among the available processors. For example, it might be possible to divide a database by key range intervals. The new databases and their associated applications can then be distributed among the existing IMS systems, and you can use the Terminal Routing exit routine to route incoming transactions to the correct IMS system. Another possibility is to divide the database by geographic area. Each IMS system could process the transactions that refer to the databases for its own geographic area and route transactions that refer to a remote geographic area.

In addition to balancing the workload across processors, you might also need to balance the workload on physical links. This occurs when a physical link between two systems is of the SDLC type and multiple physical links have been installed. You can balance the workload on physical links by:

• Specifying, during IMS system definition, proper logical link paths and logical links for each remote application.
• Using a user-written TM and MSC Message Routing and Control user exit routine (DFSMSCE0) to distribute the transaction load on each of the alternative physical links.