Design-Time Compliance of Service Compositions in Dynamic Service Environments

Abstract

In order to improve the flexibility of information systems, an increasing amount of business processes is being automated by implementing tasks as modular services in service compositions. As organizations are required to adhere to laws and regulations, with this increased flexibility there is a demand for automated compliance checking of business processes. Model checking is a technique which exhaustively and automatically verifies system models against specifications of interest, e.g. a finite state machine against a set of logic formulas. When model checking business processes, existing approaches either cause large amounts of overhead, linearize models to such an extent that activity parallelization is lost, offer only checking of runtime execution traces, or introduce new and unknown logics. In order to fully benefit from existing model checking techniques, we propose a mapping from workflow patterns to a class of labeled transition systems known as Kripke structures. With this mapping, we provide pre-runtime compliance checking using well-known branching time temporal logics. The approach is validated on a complex abstract process which includes a deferred choice, parallel branching, and a loop. The process is modeled using the Business Process Model and Notation (BPMN) standard, converted into a colored Petri net using the workflow patterns, and subsequently translated into a Kripke structure, which is then used for verification.

A Formal Model for Compliance Verification of Service Compositions

“…The resulting approach 1) is process design and implementation independent, 2) allows for correct interpretation of temporal logic specifications, 3) provides full insight into possible parallel service invocations, 4) provides insight into the next local service invocation, 5) supports arbitrary cycles, 6) causes limited state explosion compared to other approaches, and 7) allows further model reduction through equivalence with respect to stuttering. This paper extends the work reported in [8] in the following ways: (i) it provides an extended reduction of the Kripke structure, allowing for a significant performance improvement in case of multiple complex formulas, (ii) it allows the application of CTL instead of CTL-X, (iii) it provides a real-life case study for validation, showing the feasibility in complex scenarios in practice, (iv) it provides an evaluation on expressive power when compared with existing conversion approaches and, as such, shows that our approach provides a more precise representation of process constructs, and (v) it provides a more extensive performance evaluation.…”

“…As a result, well-known model checking techniques and tools can be applied during verification. An initial version of the approach was presented in [8]. An overview is given in Figure 1.…”

Variability in business processes: Automatically obtaining a generic specification

“…CTL defines the following operators over paths: 4 The semantics of CTL are defined on a Kripke structure, which is shown formally in Appendix A. The class of CTL formulas used throughout this paper require a Kripke structure with a business process specific labeling known as a transition graph [15,1]. A transition graph is a Kripke structure obtained from a Petri net which maintains parallel and local next behavior (i.e., the next activity occurrence within a single branch) of business process models.…”

Research on Probabilistic Optimization to Dynamic Composition for Service Replacement

“…In addition, it is also data equivalence when the current service planning is used to replace the failed service fs that ws 1 …”

“…The solution of service oriented architecture (SOA) has become an important method to achieve resources sharing and applications integration through services composition [1][2][3]. In general, service process is defined in term of workflows, a flexible mechanism to implement business logics [4], by which the advantages of high interoperability, crossplatform-ability and loose coupling contribute to dynamically assign services to each task of workflow in run-time environment.…”