Modeling and enacting complex data dependencies in business processes

TLDR: This paper extends BPMN data objects with few annotations to allow data dependency handling as well as data instance differentiation and introduces a pattern-based approach to derive SQL queries from process models utilizing the above mentioned extensions.

Abstract: Enacting business processes in process engines requires the coverage of control flow, resource assignments, and process data. While the first two aspects are well supported in current process engines, data dependencies need to be added and maintained manually by a process engineer. Thus, this task is error-prone and time-consuming. In this paper, we address the problem of modeling processes with complex data dependencies, e.g., m:n relationships, and their automatic enactment from process models. First, we extend BPMN data objects with few annotations to allow data dependency handling as well as data instance differentiation. Second, we introduce a pattern-based approach to derive SQL queries from process models utilizing the above mentioned extensions. Therewith, we allow automatic enactment of data-aware BPMN process models. We implemented our approach for the Activiti process engine to show applicability.

Figures

Fig. 4: Describing create, read, update, and delete of a data object.

Fig. 3: Describing object interactions in (a) 1:1, (b) 1:n, and (c) m:n cardinality.

Fig. 9: Object life cycles of objects (a) ProC and (b) CP derived from the process model.

Fig. 2: Object life cycle of data object D with two representations.

Fig. 8: Build-to-order scenario: purchase order arrangement.

Fig. 10: Setting missing foreign key relation of m:n object Component: Concrete update statement of subprocess 17 to relate all CPs referring to supplier B to the PO with ID 17 indicated by arrows.

Full text

Technische Berichte Nr. 74
des Hasso-Plattner-Instituts für
Softwaresystemtechnik
an der Universität Potsdam
Modeling and
Enacting Complex
Data Dependencies in
Business Processes
Andreas Meyer, Luise Pufahl, Dirk Fahland,
Mathias Weske

Technische Berichte des Hasso-Plattner-Instituts für
Softwaresystemtechnik an der Universität Potsdam

Technische Berichte des Hasso-Plattner-Instituts für
Softwaresystemtechnik an der Universität Potsdam | 74
Andreas Meyer | Luise Pufahl | Dirk Fahland | Mathias Weske
Modeling and Enacting Complex
Data Dependencies in Business Processes
Universitätsverlag Potsdam

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "Modeling and enacting complex data dependencies in business processes technische berichte des hasso-plattner-instituts ; 74)" ?

In this report, the authors address the problem of modeling processes with complex data dependencies, e. g., m: n relationships, and their automatic enactment from process models. Second, the authors introduce a pattern-based approach to derive SQL queries from process models utilizing the above mentioned extensions. The authors implemented their approach for the Activiti process engine to show applicability. 

Q2. What have the authors stated for future works in "Modeling and enacting complex data dependencies in business processes technische berichte des hasso-plattner-instituts ; 74)" ?

In future work, the authors will address these limitations. 

Q3. What is the purpose of enacting processes in process engines?

The purpose of enacting processes in process engines or process-aware information systems is to query, process, transform, and provide data to process stakeholders. 

Q4. What is the number of objects to be created?

The number of object instances to be created is determined by the process variable #items, which is attached to the output data flow edge. 

Q5. What are the two main concepts used to distinguish and reference data object instances?

To distinguish and reference data object instances, the authors utilize proven concepts from relational databases: primary and foreign keys [20]. 

Q6. What is the process variable used to set the state of the data object table?

Each new entry is added to the data object table of D2 with a default primary key value, the values for the specified foreign keys, and the process variable value of $stateVar as state value covered by the corresponding SQL query. 

Q7. What is the syntax for building the join-table between m:n data object D2?

Using the statement JOINALL(D2, D3), the authors can build the join-table between m:n data object D2 and the case object D3 by means of their foreign key relation. 

Q8. Which pattern describes an update operation on multiple data object instances of D2?

pattern 5 describes an update operation on multiple data object instances of D2, which sets the foreign key d4 id that is not set yetModeling and Enacting Complex Data Dependencies in Business Processes 13and moves them to state r. 

Q9. What is the pattern for a read operation on a dependent data object?

The pattern describes a read operation on a dependent single instance data object for which the foreign key value is not yet set, i.e., the data object instance is not yet correlated to a scope instance. 

Q10. What is the downside of this approach?

The downside of this approach is an increased complexity of the process model; however, this complexity can be alleviated through appropriate tool support providing views, abstraction, and scoping. 

Q11. What is the function that ensures that the read activity is enabled?

The guard ensures that the activity is only enabled if all instances related to $ID are in state t.D1:nR2 – Read multiple states.