Manage Your Workflows : A Classification Framework and Technology Review of Workflow Management Systems

Abstract

Managing workflows and complex asynchronous operation flows is a common problem that needs to be solved in a variety of software products. Workflow management systems are used to provide solutions that implement those features and orchestrate their execution. The infrastructure and data models in those products vary significantly and the amount of them can make the choice of a single workflow management system a tedious task.

In this thesis, we try to make this task easier by providing a common classification framework that can be used to compare different workflow management systems with each other. By using the classification framework we can distinguish the project and technical viewpoints from each other and provide a more objective baseline for the comparison of different workflow management systems. A systematic mapping study is used as a method to derive an initial classification framework for the workflow management systems.

In the scope of this study we focus on cloud-native and open source products to get a clear view on the freely available modern solutions on the field. A set of the most popular products with those characteristics is chosen by using a variety of different popularity metrics. As a result we have ten different workflow management systems that meet our popularity and study scope requirements and that can be reviewed against the classification framework.

The initial classification framework is refined with the results of a documentation analysis done on the selected workflow management systems. After that a full technology review is conducted on them using the classification framework. The steps and results of this technology review are documented in the thesis.

Finally, the learnings of that process are gathered into a set of guidelines for selecting a workflow management system. Those guidelines can be used to recreate the study with a new set of systems and iterate through them until a final choice can be made. By offering a classification framework, guidelines for its usage and an example of the review we believe that the work can be extended on any set of workflow management systems and used to perform a review on them against each other.