Automatic derivation of conceptual database models from differently serialized business process models

Automatic derivation of the initial conceptual database model from a set of business process models

“…Apart from BPMN and UML AD, which are dominantly used, the existing POM-based approaches also take source models represented by: Petri Net, RAD (Role Activity Diagram), EPC (Event-driven Process Chain), TCD (Task Communication Diagram), and A-graph. [17,14])…”

“…Based on the experimentally confirmed [7,20] semantic capacity of data-centric BPMs, a two-phase BPM-driven approach to the CDM synthesis was proposed [21], and the first online service-oriented BPM-driven CDM generator (M-lab Generator) was implemented [19]. Finally, the AMADEOS system [36,14] was launched, as the first online web-based system for BPM-driven CDM synthesis. AMADEOS implements the CDM synthesis process through an orchestration of the M-lab Generator services, and enables the CDM synthesis based on BPMs that may be represented by BPMN or UML AD.…”

“…Although the idea of MDSDM, taking BPMs as a starting base, dates back in the early 1990s, in the existing literature there is only a small number of papers presenting the implemented automatic model-driven generator of the target data model with the corresponding evaluation results, while the great majority of papers only present modest achievements in (semi)automated, or even manual, data model synthesis. The surveys [17,69,14] show that the existing approaches are characterized by the direct synthesis of the target model based on BPMs represented by a single concrete notation such as BPMN [58] or UML [59] activity diagram (UML AD). Furthermore, the existing tools are mainly implemented as plug-ins or transformation programs within development platforms (typically Eclipse-based), and also able to process only a single BPM (represented by a single diagram), although a real source model contains a finite set of models (diagrams).…”

“…The surveys [17,69,14] show that a fully automated MDSDM approach is still the subject of extensive research, and the aforementioned problems of limited functionality, portability, effectiveness and efficiency of the existing MDSDM tools remain a significant research goal. In order to contribute the solving these problems, we have started a longterm research project focused on the automatic CDM synthesis based on a collection of models representing business processes of an enterprise, with the main objective to develop an online platform-independent tool for the automatic CDM derivation from a collection of BPMs that may be represented by different notations, and differently serialized, as well.…”

“…The first very important milestone in our progress was the indirect two-phase CDM synthesis [21], which enables and facilitates derivation of the target data model from differently represented BPMs. By applying the two-phase CDM synthesis approach, we have launched the AMADEOS 3 system [19,36,14], which was the first online web-based system enabling automatic CDM derivation from BPMs that may be represented by two different notations and differently serialized. However, AMADEOS was able to automatically generate a CDM based on a single BPM, as the large majority of the existing MDSDM tools.…”