Artifact-Centric Workflow Dominance

Abstract

Abstract. In this paper we initiate a study on comparing artifactcentric workflow schemas, in terms of the ability of one schema to emulate the possible behaviors of another schema. Artifact-centric workflows are centered around "business artifacts", which contain both a data schema, which can hold all of the data about a key business entity as it passes through a workflow, along with a lifecycle schema, which specifies the possible ways that the entity can evolve through the workflow. In this paper, the data schemas for artifact types are finite sets of attribute-value pairs, and the lifecycle schemas are specified as sets of condition-action rules, where the condition is evaluated against the current snapshot of the artifact, and where the actions are external services (or "tasks"), which read a subset of the attributes of an artifact, which write onto a subset of the attributes, and which are performed by an entity outside of the workflow system (often a human). The services are also characterized by pre-and post-conditions, in the spirit of semantic web services. To compare artifact-centric workflows, we introduce the notion of "dominance", which intuitively captures the fact that all executions of a workflow can be emulated by a second workflow. (In the current paper, the emulation is focused only on the starting and ending snapshots of the possible enactments of the two workflows.) In fact, dominance is a parametric notion that depends on the characterization of the policies that govern the execution of the services invoked by the workflows. In this paper, we study in detail the case of "absolute dominance", in which this policy places no constraints on the possible service executions. We provide decidability and complexity results for bounded and unbounded workflow executions in the cases where the values in an artifact range over an infinite structure, such as the integers, the rationals, or the reals, possibly with order, addition, or multiplication.

Verification of relational data-centric dynamic systems with external services

“…For each considered variant, verification is generally undecidable; decidability results were obtained only under rather severe restrictions, e.g., restricting all pre-conditions to be "true" [28], restricting to bounded domains [29,9], or restricting the pre-and post-conditions to be propositional, and thus not referring to data values [29]. [16] adopts an artifact model variation with arithmetic operations but no database. It proposes a criterion for comparing the expressiveness of specifications using the notion of dominance, based on the input/output pairs of business processes.…”

“…Process-centric formalisms focus on control flow while under-specifying the underlying data and its manipulations by the process tasks, often abstracting them away completely. In contrast, data-aware formalisms treat data as first-class citizens [37,30,19,16,40,1]. The holistic view of data and processes together promises to avoid the notorious discrepancy between data modeling and process modeling of more traditional approaches that consider these two aspects separately [9].…”

Decentralized enforcement of document lifecycle constraints

“…In other cases, knowing the workflow allows to statically analyze it and make sure that all declarative lifecycle constraints are respected at all times [13,20,25,[27][28][29]. For example, Gonzalez et al symbolically represent GSM-based business artifacts, in such a way that model checking can be done on the resulting model [30].…”

“…For example, [27] considers an artifact model with arithmetic operations, no database, and runs of bounded length. The approaches in [13,28] impose that domains of data elements be bounded, or that pre-and postconditions refer only to the artifacts, and not their variable values [28].…”

Comparing workflow specification languages

“…Verification for such models is considered in [23,24,11,16,21]. The comparison of such systems is considered in [14] using the notion of dominance, which focuses on the input/output pairs of the workflows. Other models in the same spirit include the Vortex workflow framework [28,18,27], the OWL-S proposal [30,29] as well as some work on semantic Web services [33].…”