| Software systems based on SOA are referred to as service-based software systems (SBS). The working of SBS system relies on effective collaboration among heterogeneous and autonomous Web services in the Internet, where the QoS depends on not only the Web service with autonomy itself, but also the change of the Internet environment and the Web service space in the open environment. Therefore, SBS system needs a dynamic and "soft" quality guarantee mechanism, being able to adapt to environment changing. The SBS that is able to autonomously adjust the runtime behavior with the environment in order to improve the system performance is called adaptive service-based software systems (ASBS) by the researchers. Through monitoring the service and dynamically adjusting the implementation plan, ASBS realizes the SBS’s environment adaptive ability. Changes in the environment will usually affect the quality of service, thus affecting the adaptive behavior. Facing the requirement of adaptive, timely and dynamically adjusting the organizing method of alternative services according to the quality of service can better support ASBS adaptive. Therefore, how to effectively organize and manage alternative services to support ASBS adaptive becomes a problem must be addressed.Aiming at these problems, combined with the platform to support composite service environment adaptive that our research group had built up, virtual service community dynamically adjusting method is proposed. After the analysis of virtual service community adjusting requirement for ASBS adaptive, the data structure of virtual service community, operational mode and the corresponding management modules are designed. On the basis of virtual service community, the ASBS adaptive processes based on virtual service community for ASBS passively adaptive and actively adaptive methods are designed. Adjustment of the virtual service community will directly affect the composite service execution plan, so as to realize supporting the ASBS adaptive. This thesis divides the adjustment process into three stages, analysis, decision and adjustment. To support the information transfer between the virtual service community adjusting stages, environment and state information of composite service in formalized representation model is provided. In the research of analysis and decision method supporting passive adaptive and active adaptive, critical task analysis, urgent degree analysis, cost-benefit analysis and service advanced execution method are proposed. In virtual service community adjusting, after virtual service community adjust action been designed, local constraint modification algorithm, service state transition method and service sorting algorithm considering the factors such as user preferences and local urgent degree are proposed. Finally, this thesis verifies virtual service community can effectively support the process of ASBS adaptive by the example analysis, and analyses the time complexity of the process of virtual community service adjustment, and through the experiment analyses the relationship between the time of virtual service community to adjust and each influence factor. |
