> For the complete documentation index, see [llms.txt](https://hanfak.gitbook.io/workspace/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://hanfak.gitbook.io/workspace/general-paradigms/enterprise/workflow.md).

# Workflow Design Patterns

* Workflows are used in software applications to automate business processes.
* A workflow defines the sequence of steps or tasks that a business process should execute.
* The control flow workflow patterns describe the standard ways of task execution in a business process.
* Patterns used in use case to describe the behaviour

## links

* <http://www.workflowpatterns.com/patterns/control/>
* <https://www.sciencedirect.com/topics/computer-science/workflow-pattern>
* <https://docs.aws.amazon.com/amazonswf/latest/awsrbflowguide/programming-workflow-patterns.html>
* <https://en.wikipedia.org/wiki/Workflow_pattern>

## Libraries

* <https://github.com/j-easy/easy-flows>
* <https://github.com/copper-engine/copper-engine>

## Types

![workflows-examples.jpg](/files/-M1oFarLVkG9z9jUt-Zj)

* (a) Sequential
  * Several tasks follow after one another
  * Task A needs to be done first, then Task B then Task C
  * Blocking tasks.
    * Task C needs to wait for Task B and Task A to be finished before starting
* (b) And split
  * After Task A is done, several tasks can begin at the same time
  * These tasks may not finish at the same time
  * These can be done in parallel and on seperate threads or processes
* (c) Synchronization
  * several concurrent activities to terminate before an activity can start
  * Normally associated with (b)
  * Must wait for all tasks to finish
* (d) XOR split
  * requires a decision
  * after the completion of task A, either B or C can be activated
  * Some sort of if/else or case/switch or strategy pattern
* (e) XOR join
  * several alternatives are merged into one
  * in our example task C is enabled when either A or B terminates
  * Race condition determines flow
* (f) OR split
  * is a construct to choose multiple alternatives out of a set.
  * after completion of task A, one could activate either B or C, or both
* (g) multiple merge
  * allows multiple activations of a task and does not require synchronization after the execution of concurrent tasks.
  * Once A terminates, tasks B and C execute concurrently
  * When the first of them, say B, terminates, then task D is activated; then, when C terminates, D is activated again.
* (h) discriminator
  * waits for a number of incoming branches to complete before activating the subsequent activity
  * then it waits for the outgoing branches to finish without taking any action until all of them have terminated. Next, it resets itself.
* (i) N out of M join
  * provides a barrier synchronization
  * Assuming that M > N tasks run concurrently, N of them have to reach the barrier before the next task is enabled
  * any two out of the three tasks A, B, and C have to finish before E is enabled
* (j) deferred choice
  * the choice is not made explicitly and the run-time environment decides what branch to take,