A Process Model or Workflow Model is a formal description of a business process containing a set of process definitions and other process artifacts. All information concerning a process model is stored in an XPDL file referred to as (Stardust) model file. The file format itself is described in chapter Stardust File Formats.
A process model can be self-contained or depend on other process models in one of the following ways:
Workflow models created with Stardust are semantically built on top of the WfMC Workflow Reference Model (www.wfmc.org). Nevertheless, the Stardust process model adds a lot of functionality and extension points to the WfMC Workflow Reference Model and can thus be regarded as a semantic superset.
The following figure shows an overview of the basic elements used in the process model:
Figure: Components of a Process Model
This chapter will make you familiar with the elements of a process model as interpreted by the Stardust Process Engine.
Your process model or 'model' in short, serves as container for the business processes in your problem domain with the following components:
A full set of model deployments with all their links is called a Process Model Set.
Each time a deployment or re-link happens, a snapshot of the resulting Process Model Set is stored in the Audit Trail database. At start time of any Root Process Instance the current snapshot of the Process Model Set (as defined by the Model containing the Root Process Instance's Process Definition) is associated with the Process Instance. This ensures that during the lifetime of this Process Instance a consistent definition is used (For example, if a Process Interface is invoked multiple times with static resolution, each time the same Implementation is invoked, namely the Primary Implementation at the time of process start).
The Stardust Process Workbench maintains its own model repository of model files. The model repository manages concurrent access to the model and version control in the modeling environment.
After being ready for execution, process models are deployed to the runtime environment to be executed by the Stardust Process Engine. Deployment of a model means storing it in the audit trail together with some information for the Stardust runtime to evaluate the active model among all deployed models. Only processes from the active model may be started but processes from other models already running will continue to run. Refer to the chapter Working with Multiple Versions for details on multiple model versions in the runtime environment.
It is possible to deploy different models into one runtime. This scenario is described in detail in chapter Deploying Different Models into One Runtime of the Concepts of Using Multi Model Audit Trail Setups.
Process models and process model sets are determined as closed if they either do not contain any references to process models or they contain implementing process models for all abstract process interfaces that they are using.
In contrast, a process model is called open if it is not closed as described above.