VerChor: A Framework for the Design and Verification of Choreographies

Abstract

Choreographies are contracts specifying from a global point of view the legal interactions that must take place among a set of services. Such a contract may serve as a reference in the development of concurrent distributed system, whether it is achieved following a top-down or a bottom-up approach. In this article, we present VerChor, a generic, modular, and extensible framework for supporting the development based on choreographies. It relies on a choreography intermediate format (CIF) into which several existing choreography description languages can be transformed. VerChor builds around a set of formal properties whose verification is central to choreography-based development. To support this development process, we propose a connection between CIF and the CADP verification toolbox, which enables the full automation of the aforementioned properties. Finally, we illustrate a practical use of the VerChor framework through its integration with the Eclipse BPMN 2.0 designer.

Checking business process evolution

“…The extension to choreography is possible by adding choreography tasks in PIF and treating them as basic observable atoms in the behavioural equivalences and inclusion relations. However, based on our past experience on choreography verification [9], we believe that specific evolutions relations could be defined for choreographies. Patterns of properties.…”

“…BPMN is an ISO/IEC standard since 2013 but its semantics is only described informally in official documents [3,4]. Therefore, several attempts have been made for providing BPMN with a formal semantics, e.g., [5,6,7,8,9]. In this paper, we abstract the features of BPMN related to data and we focus on the core features of BPMN, i.e., its control flow constructs, which is the subset of interest with respect to the properties we propose to formally analyse in this paper.…”

“…PIF is simpler than BPMO. It originates from a previous intermediate format we had defined, CIF [9]. Its objective is neither to be a comprehensive set of all concepts found in workflows, nor to support automated mapping between notations in their whole.…”

“…[37,38] focus on the semantics proposed in [6,36] and propose an automated transformation from BPMN to timed CSP. In a previous work [9], we have proposed a first transformation from BPMN to LNT, targeted at checking the realizability of a BPMN choreography. We followed a state machine pattern for representing workflows, while we here encode them in a way close to Petri net firing semantics, which favours compositionality and is more natural for a workflow-based language such as BPMN.…”

SBVR2Alloy: An SBVR to Alloy Compiler

“…Work on formal semantics in this area has focused more on the imperative (or procedural) approach [14], [17], [18], [19], [20], [21], [16], [22] and has been geared towards an interleaving semantics which sometimes unnecessarily restricts the possible behaviours, e.g., see PROPANE in [20], or forces the modeller into making premature decisions, e.g., see separated collaboration diagrams [14]. A true concurrency semantics is proposed in [16].…”