Policy-based dynamic server selection
Policy-based dynamic server selection (DSS) provides a flexible mechanism for controlling the flow of work to CICS® servers.
In a high availability topology that uses policy-based DSS, workload can be dynamically managed across a number of CICS servers by using round-robin workload distribution; alternatively one or more backup servers can be made available to a CICS server, if that server becomes unavailable (failover).
Policy-based DSS enables failed requests to be automatically retried on alternative CICS servers during abnormal conditions, for example during planned maintenance or overnight batch processing, and for restarted CICS servers to become available for work again, without operator intervention.
At runtime, policy-based DSS maps logical CICS server names to CICS server groups, where a group definition includes a preassigned workload distribution algorithm. A request specifying a logical CICS server name that arrives at CICS Transaction Gateway is routed to one of the servers in the CICS server group that maps to the logical CICS server name, in accordance with the workload distribution algorithm.
Blank server names can be mapped explicitly to a DSS server group or can be configured to use a single default CICS server.
Server names for which no explicit mapping has been defined can be mapped to a default DSS server group; this mapping includes blank server names unless there is an explicit mapping for them. Such a mapping overrides any default server definition.
For information about how to configure a DSS policy see Configuring a dynamic server selection policy.