There may be situations where you want to distribute your processing over multiple Process Engines, but distribute the steps of a process model among the Integration Server nodes in a cluster.

In this implementation, you design the process model so that generation distributes the steps among the Integration Server nodes. That is, the sum total of all steps across all Integration Server nodes in a distributed cluster equals the complete process. A distributed cluster implementation shares a webMethods Broker (deprecated), but each Integration Server connects to the central webMethods Broker (deprecated) instance with its own client prefix.

As in a standard cluster, a single external RDBMS containing a Process Engine database supports all Integration Server nodes in a distributed implementation.