You can separate the work in a workload based on:
You can create a workload that routes requests from a set of requesting regions to different subsets of target regions based on the name of the terminal, user, or both associated with each occurrence of a transaction. For example, you might want to route all requests initiated by certain individuals from specific terminals to a special subset of target regions.
Figure 3 illustrates what such a workload might look like. In this case, if the user and terminal names associated with a transaction begin with SM and NET, respectively, the transaction is routed to the set of target regions identified as EYUCSG05. If either the user or terminal name begins with any other characters, the transaction is routed to the default set of target regions identified as EYUCSG01 on the workload specification.
During workload processing, CICSPlex® SM evaluates the terminal and user names associated with each occurrence of a request to determine where the request should be routed.
After determining the appropriate set of target regions, CICSPlex SM selects one based on the status of the active target regions in that set.
You can create a CICS® BTS workload that routes requests associated with a certain process type to a specific target region or set of target regions. For example, you may wish to route all the requests associated with the STOCK process type to a special subset of target regions.
Figure 4 illustrates what such a workload might look like, if the process type associated with a CICS BTS transaction is STOCK, the transaction is routed to a set of target regions identified as EYUCSG05. If the process type is anything other than STOCK, the transaction is routed to the default set of target regions identified as EYUCSG01 in the workload specification.
workload group
If you choose to separate a workload by process type, you must set the Luname and Userid fields to *. If you separate a workload by LU name and user ID, you must set the Process Type field to *. If you want to separate an enterprise bean workload, the Luname and Process Type fields must be set to *. You can separate a workload only either by process type or by LU name and user ID.
You can specify either a specific or a generic process type. During workload separation processing, CICSPlex SM evaluates the process type supplied by CICS to determine to where the transaction should be routed.
You can also separate the work in a workload based on the transactions themselves. For example, you might want all occurrences of payroll-related transactions initiated from terminals in an accounting department to be routed to a specific set of target regions for processing.
Figure 5 illustrates how you might separate the work in a workload based on transaction identifiers. In this case, if the user and terminal names associated with any transaction identified in transaction group EYUWMT01 begin with SM and NET, respectively, the transaction is routed to the target regions identified as EYUCSG05. If the transaction identifier, user name, or terminal name does not match the criteria, the transaction is routed to the default target regions identified as EYUCSG01.
During workload processing, CICSPlex SM evaluates the transaction identifier supplied by CICS to determine which transaction group to use.
CICSPlex SM uses the match key value to establish the order in which the terminal and user names associated with the transaction are to be evaluated. The evaluation is used to determine where the transaction should be directed:
After determining the appropriate set of target regions, one is selected based on the status of the active target regions in that set.
The method for workload separation of enterprise beans is the same as that described in Separating by transaction. For example, you might want all enterprise bean-related transactions to be routed to a specific set of target regions for processing.
An incoming IIOP enterprise bean request to CICS includes the bean name, which is matched to a predefined request model definition that is installed in the routing regions (cloned listener regions). The request model identifies, among other things:
Figure 6 illustrates how you might separate enterprise bean-related transactions. The request definition relates enterprise bean beanname with CICS transaction EJB1, which belongs to transaction group EYUTGEJ1. The workload definition identifies the target scope as EYUSGEJ2. The match key is USERID and you can use a specific or generic user id in the transaction group definition. If the user id in the transaction group definition does not match that in the incoming IIOP request, the enterprise bean-related transaction is routed to the default target region EYUSGEJ2.
Link3270 bridge workloads can be separated by user ID, LU name, and transaction group. For Link3270 bridge workloads theLU name can be produced in three different ways:
You can separate Link3270 bridge workloads by LU name only if you are using methods 1 or 3 of those listed, so that the LU name is known in advance. If you are using the method 2, the LU name is not known in advance and cannot be used for workload separation.
To separate by the bridge facility NETNAME and not the name associated with the client program that started the Link3270 bridge you must modify the EYU9WRAM module. You can use the CICS API commands:
to assign the user ID and LU name. You can use the NETNAME returned from the INQUIRE BRFACILITY() command rather than the NETNAME passed via the DFHDYPDS commarea parameter DYRNETNM to separate the workload.
For more information about Link3270 bridge facility definitions see the CICS External Interfaces Guide.