On the Optimal Petri Net Representation for Service Composition

Abstract

Service composition has received significant attention in the research community, and the focus has been on service semantics and composition algorithms. Surprisingly, the problem of representation of the composition outcome has been largely ignored. Ad-hoc workflows are often employed, which typically sacrifice alternative paths and parallelism for the sake of simple representation. In this paper, we show how theory of regions, which was originally developed to derive Petri nets from finite state automata, can be applied to find the optimal representation of composition. To apply the theory, we first propose an automaton-based composition framework that incorporates most existing composition techniques without changing the service semantics or its description language. Then based on the special requirements of the composition representation, we develop our own Petri net synthesis algorithm that combines the benefits of two well known algorithms from the theory of regions. We demonstrate that AND/OR workflow nets can limit the concurrency even for simple input/output based service composition, while our Petri net representation is optimal in terms of flexibility and parallelism. Our experimental evaluations include a case study on composing Google Checkout Service, and the study on Oracle BPEL samples, for which our algorithm obtains better concurrent representations for almost all non-trivial cases.

From Communicating Machines to Graphical Choreographies

“…This is transformed into a BPMN diagram via a Petri net transformation based on [19]. The work [32] gives an algorithm to compose several services. Each service is presented as an automaton and a set of automata are composed by a parallel product.…”

Workflow Refactoring for Maximizing Concurrency and Block-Structuredness

“…Generally, workflows are expected to be free of both control flow errors (e.g., deadlocks) and data flow errors (e.g., activity input is undefined) [5], [6], [7], [8]. Maximizing concurrency in the control flow, with respect to the data flow, is also critical, and has recently received more attention in the literature [9], [10], [11], [12]. Concurrency maximization refers to the property that activities without dependences (including control dependences, data dependences [13]) should be executed in parallel.…”

“…Recently, a number of approaches to workflow refactoring have been proposed [17], [9], [10], [11], [16], [12]. For example, Wang et al first transform BPEL workflows into automata then employ Petri nets synthesis [18] to obtain a workflow with maximized concurrency [11]. Jin et al propose to leverage workflow mining (i.e., the α-algorithm [19]) for workflow refactoring such that activities without data dependences can be executed in parallel [12].…”

“…Furthermore, it is difficult for programmers to transform unstructured workflow models into popular block-structured workflow languages such as BPEL [14], [21]. For an approach which obtains the block-structured control flow (i.e., the approach in [11]), maximized concurrency is not ensured. One may argue that we can maximize concurrency and then block-structuring or reversely by combining these two existing approaches.…”

“…One may argue that we can maximize concurrency and then block-structuring or reversely by combining these two existing approaches. Unfortunately, this solution is infeasible because the approach in [11] sacrifices concurrency for block-structuredness. In this paper, the control flow of a workflow is regarded as optimal if it ensures maximized concurrency and block-structuredness.…”