About the Process Engine and Clustered Operation
In its most basic implementation, the Process Engine is installed as the WmPRT package on a single Integration Server, connected to its own Process Engine database. A webMethods Broker (deprecated) is also included for routing and delivering messages.
For process model development and testing purposes, or for very small process solutions, a single Process Engine, Integration Server, and webMethods Broker (deprecated) is adequate. However, most process solutions require more resources for efficient operation, and it is common practice to run two or more Integration Servers in one of the following configurations:
Distributed (non-clustered) mode. All
Integration Server nodes have their own
webMethods Broker (deprecated) client connection but the process model is designed so that only specific parts of the process run on any one Integration Server node. For details, see
Operating
Process Engine in an
Integration Server Cluster.
External cluster. This is functionally equivalent to the
Integration Server cluster from a BPM perspective, except that the
Integration Server nodes themselves are not part of a
Integration Server cluster. Customers can chose this configuration when they do not want to use the actual
Integration Server clustering feature. For details, see
Operating
Process Engine in an
Integration Server Cluster.
For more information about working with Integration Server and webMethods Broker (deprecated) in a clustered environment, see webMethods Integration Server Clustering Guide.