Working With Process Flow

A process flow is a set of activities arranged in a sequence to perform a specific task. Combining various activities, that is, Source, Target, Schema, Mapping, or Transformer creates a process flow.  A Process Engine controls the execution of a process by starting the execution of process flow instances and all the activities present in a process flow. When the process flow is executed, data from the source is converted to the intermediate form and dispatched to the target. The transformer does the conversion of data. There are two types of transformers in Adeptia:

• Stream2XMLStream/XMLStream2stream: This transformer converts the source data to XML (Stream2XMLStream) and then XML to target data (XMLStream2Stream).
• SchemaStream2Record/SchemaRecord2Stream: This transformer converts source data to an intermediate format (SchemaStream2Record) and then intermediate format to target data (SchemaRecord2Stream).

Repository
When the process flow is executed, data from the source is converted to the intermediate form and then dispatched to the target. The intermediate data is stored in a repository folder (.../AdeptiaInstallFolder/AdeptiaServer-x.x/ServerKernel/web/repository). Another folder is created representing the group which executed the process flow. All the files that are created during the activity, design, and running of the process flow, are stored in the group folder and this group folder is stored inside the repository folder. While creating a process flow, a user can specify whether to keep or delete the repository folder.

Activities of a process flow are executed in the following two ways:

• Synchronous Activity: If an activity is synchronous, process flow initiates the execution of that activity, and waits for its completion. The execution of the subsequent activity is initiated only after the completion of the first activity.

• Asynchronous Activity: If an activity is asynchronous, process flow initiates the execution of that activity and subsequent activities, as defined in the process flow without waiting for the completion of the first activity.

  Repository file is not generated for the activities used in the Asynchronous mode.

  Controlling Synchronization

  In a process flow, if an activity is synchronous, no signal is required to acknowledge the Process Engine about its completion as it runs in the main thread. If an activity is asynchronous, running parallel to the main thread, some mechanism is required to notify the Process Engine about its completion. This is achieved by signalling.
  
  

  Signaling
  Signaling facilitates coordination of asynchronous activities in a process flow. A signal is raised as an asynchronous activity completes. The Synch Node in the process flow waits for signal and ensures that all the activities raising that signal are completed before control moves further in the process flow. This is done using the Synch Node option while defining a process flow.
  
  

  Time Out
  When an activity is used as asynchronous, a new property TimeOut is added to it. By default, its value is 300 seconds. If an asynchronous activity cannot connect to the next activity within this time limit, it gets aborted. For information on how to use an activity in asynchronous mode and changes its property, refer to Creating Process Flow.

Following are the major subsections: