Coupling Facility Structures

Instead of accessing data in a coupling facility by address, you can allocate three types of objects, called structures, and access data in the structures as logical entities (by name, for instance). The ability to access data in this manner frees coupling facility users from having to be concerned with the physical location or address of the data.

A coupling facility can hold one or more structures of any type, however, each structure must reside entirely in a single coupling facility. Applications are not limited to using a single coupling facility structure. They can use multiple structures of the same type or different types.

Products and subsystems that exploit the coupling facility indicate their coupling facility structure requirements as part of their installation information. For instance, a product might require a lock structure of a certain size, with particular attributes, and a certain name. Or, a product might require that a structure be allocated in a certain level (CFLEVEL) of a coupling facility because of the functionality it provides.

When system administrators or system programmers install software that requires a coupling facility structure, they create a coupling facility resource management (CFRM) policy that specifies the name, size, and attributes of each structure to be allocated. The CFRM policy also allows the installation to limit the amount of storage each structure can occupy and control where each structure is allocated, through a “preference list”, for multiple coupling facilities.

Once the policy is defined, the operator needs to issue the SETXCF command to activate the policy. The activated policy does not cause the structures to be allocated. A structure is allocated only when the first user connects to the structure.