Onegine: an Engine for the Description, Analysis and Execution of Block-structured Business Processes

Abstract

The core of modern enterprise resource planning (ERP) systems are business process engines. The vast majority of them use the BPMN description language (for example, jBPM, Activiti, Camunda). Despite the adoption of the standard, editor’s BPMN support is platform-oriented and has limited ability to transfer definitions. In addition to editors, a business analyst uses a whole set of auxiliary tools to create and maintain BPMN descriptions up to date, including verification tools (for example, WoPeD) and productivity analysis (Bizagi, BIMP). Involvement of its own features for identifying and updating business processes is the Process Mining methodology, which uses its own set of tools (Disco, ProM), where BPMN support is limited. This paper describes the Onegine business process engine implemented by the authors, which integrates description and analysis. The developed engine uses the previously proposed process trees as a model for description and execution, which are higher-level than the BPMN representation. The main property of process trees is the impossibility of defining a logically incorrect (from the point of view of control flow) scheme. Process trees are also the target model of a number of effective Process Mining algorithms and allow you to obtain analytical expressions for some performance indicators (time, cost). All of this allows you to integrate description and analysis tools into the execution engine, reducing the need to use external tools, thereby reducing the burden on business analysts and creates the basis for designing more efficient enterprise management systems.