Design and implementation of a distributed workflow enactment service

Abstract

Workflows are activities involving the coordinated execution of multiple tasks performed by different processing entities, mostly in distributed heterogeneous environments, which are very common in enterprises of even moderate complexity. In current commercial workflow systems, the workflow scheduler is a single centralized component. A distributed workflow enactment service, on the other hand, should contain several schedulers on different nodes of a network, each executing a part of the process instances. Such an architecture would fit naturally into the distributed heterogeneous environments. Further advantages of a distributed enactment service are failure resiliency and increased performance, since a centralized scheduler is a potential bottleneck. In this paper, we present the design and implementation of a distributed workflow enactment service based on the work of M. Singh (1996). By starting with a block-structured workflow specification language, we avoid a very general set of dependencies and their related problems. In this way, it is possible to present a simple algorithm for the distributed scheduling of process instances. Further benefits of the approach are the ease of testing and debugging the system, and the execution efficiency through having a reduced number of messages.