Business process discovery methods and process calculus for RESTful systems

Acknowledgements

Work presented on this web page was supported by the Polish National Science Center under Grant No. DEC-2012/05/N/ST6/03051 and the Poznan University of Technology.

Project description

In days of information societies, distributed computer systems need to be: scalable to process lots of requests, elastic to adapt to ever changing user’s requirements, and reliable to provide uninterrupted service. These features, however, result in a number of potential problems. Systems become over sophisticated, hard to maintain, and what follows, expensive to run. Currently, one of the leading approaches to create such systems is Service Oriented Architecture (SOA). Despite different approaches to implemen SOA, developers tend to choose REST for their architectural model to build service based systems (CRS). This approach makes maintainability and managebility of systems, and so calles business processes that are run inside of that systems much easier. Neverless, it is worth mentioning that management of business process in such systems sill leads to a number of problems like: bottlenecks, deadlocks, invokation loops.

Process mining is a field of science and engineering that can be helpful in resolving the latter problems. Process mining, based on the gathered event logs of a system, creates a model of processes executed in it. The model can be then used to search for unwanted features in order to fix them.

Distributed Systems CRS (REST) have different characteristics, which prevents the use of traditional methods of process mining. Similarly to the classical systems they generate a very large amount of data in the form of event logs, which in case of failure or inefficient operation of the system are analyzed by specialists. Unfortunately, a large number of the collected data is not possible to be analyzed by experts which often leads to wrong decisions related to the distributing or improving a wrong element of the system. Consequently, this results in even greater problems in the management and maintenance of the system.

This grant’s aim is to significantly improve the quality of management and maintenance of CRS (REST) distributed systems, by proposing a number of techniques and algorithms aimed at automatic collection and analysis of event logs in a systems using dedicated process mining methods.

During the project, a minimum set of data required to be collect in the event log of CRS systems has been developed, in order to enable discovering of a model of a CRS system process. A method of non-invasive support for creating such a log has been proposed, which does not require reimplementing of an entire system.

In addition, two new notations of business processes in CRS systems have been developed. They take into account the characteristics of these systems through the ability to express: sub-processes, system resources, the hierarchy of resources, communication and semantics of messages.

As a result, two groups of algorithms have been proposed: CRS miners and Algebra miners. They allow to discover business process models, taking into account their special characteristics. On the basis of the discovered model it is possible to analyze processes of CRS systems and their properties. It should be noted that the proposed methods allow comprehensively and efficiently explore the process models of the system.

It should be emphasized that also methods for conformance checking of structural and workflow properties of business process in CRS systems has been proposed.

Project results

Published

  • Dwornikowski, D., Stroiński, A., Brzeziński, J., Towards a process calculus for REST: current state of the art,Foundations of computing and decision sciences, volume 40, number 4, pages 237-265, Poznań, Poland, 2015.
  • Dwornikowski, D., Stroiński, A., Brzeziński, J., Conformance Checking of Communicating Resource Systems with RAs Calculus, IEEE Int. Conf. on Services Computing (SCC 2015), IEEE, pages 759-764, 2015
  • Stroiński, A,; Dwornikowski, D., Brzeziński, J., RESTful Web Service Mining: simple algorithm supporting resource-oriented systems, In Proc. of 21th IEEE Int. Conf. on Web Services, ICWS 2014, IEEE, pages 694-695, 2014,
  • Stroiński, A,; Dwornikowski, D., Brzeziński, J., Resource Mining: applying process mining to resource-oriented systems, In Proc. of 17th Int. Conf. on Business Infor. Systems, Springer Int. Services, pages 217-228, 2014,

Pending

  • Stroiński, A,; Dwornikowski, D., Brzeziński, J., A distributed model discovery of resource communicating systems, IEEE Transactions on Service Computing, (minor revision)
  • Stroiński, A,; Dwornikowski, D., Brzeziński, J., Conformance checking of structural and workflow properties of communicating resource systems with ROC, IEEE Transactions on Service Computing, (in review)

Technical Reports

  • Stroiński, A,; Dwornikowski, D., Brzeziński, J., CRS system process model discovery with Petri Nets - algorithms, Poznan University of technology technical report number RA-1/16
  • Dwornikowski, D., Stroiński, A., Brzeziński, J., CRS system process representation in ROC process algebra, Poznan University of technology technical report number RA-4/16