Modeling and Analyzing Time Constrains for Service Composition

Abstract

As a way to compose independent services together to fulfill a function, service composition is widely applied in different applications. One key issue is how to model and analyze time-related properties in service composition so that it can satisfy user's deadline requirements. A model called service composition time-constrained Petri net (SCTNet) is introduced in the paper, which is capable of capturing time behavior of services accurately. The basic relationships, priority and preemption constraints between service are also characterized based on SCT-Net. In order to reduce the state space of model, we propose four simplification schemas which can maintain structural feature and time delay of SCT-Net. We advance the concept of greatest concurrent set, and a heuristic algorithm for schedule computation is given. Finally, two specific experiments demonstrate the feasibility and practicability of our method.