Distributed units of work

When a transaction interacts with a transaction in another CICS region, or with an external resource manager, changes in all the regions and resource managers are committed or rolled back together. In this case, the unit of work is known as a distributed unit of work.

In this situation, one of the CICS regions, or one of the resource managers, is the coordinator of the unit of work and is in overall charge of committing or backing out the changes.

During sync point processing, the CICS regions and the resource managers exchange information that enables them to agree to commit or back out changes. Until the information exchange is complete, the coordinator does not know whether the distributed unit of work should be committed or backed out; the outcome of the distributed unit of work is said to be indoubt.

If communication fails during sync point processing, any indoubt distributed units of work remain indoubt until communication is restored. The coordinator then exchanges information with other regions and resource managers to enable the indoubt units of work to commit or back out. This process is known as resynchronization.