Configuration and Optimization of Virtual Business in Cloud Computing Environment

Cloud computing enables the realization of virtual business through outsourcing business functions to online services. Due to the large number of available business services and candidate cloud service providers, the configuration and optimization of the virtual business becomes a key issue for a service-oriented business ecosystem. In this paper, the virtual business is regarded as a series of service processes. Constrained service workflow net (CSWF-net) is developed to describe the component and structure of such service processes. Considering the dynamic and stochastic nature of services in a cloud computing environment, a stochastic decision making approach for CSWF-net is proposed in which the state transition of services is characterized as a Continuous Time Markov Chain (CTMC). Based on the four control structures of service processes, the dynamic Quality-of-Service (QoS) aware selection and composition problem is formulated as a Stochastic Linear Programming (SLP) model. Mixed policies generated by the branch and bound algorithm and modified heuristic approach are used to solve the dynamic QoS-aware selection and composition problem at runtime. In the meantime, the most appropriate business services and cloud service providers can be selected to support the execution of the virtual business effectively.