Performance estimations of clustered workflow architectures

In this paper, we are particularly concerned about issues of performance to see how well workflow enactment architecture is dealing with the large amount of workcases. Almost all of the traditional workflow management systems are based upon the client-server workflow enactment architecture that reveals a certain degree of limitation on the issues, such as performance and scalability. This paper presents a performance analysis model and its mathematical estimations for analyzing the performance aspect of the clustered workflow enactment architecture. We are expecting that the estimations should be helpful in the understanding of the spectrum of possibilities for large-scale workflow architectures and systems. In especially order to comparatively study the clustered workflow enactment architecture, we also analyze the performance of the client-server workflow architectures by using the same approach. Finally, we give a comparison result of performance estimations, and strongly believe that the result is going to be very useful understanding about designing and implementing workflow management systems aiming at the very large-scale workflow applications.