Research On Web Service Management Technology In Cloud Environment

Web service is one technology for building services based Service Oriented Architecture(SOA). Currently, Web service has been widely used because of its advantages. However, with the increment of internet services, some problems gradually are exposed. For example, management and maintenance of services become more and more difficult. By making services running on Cloud and scaling resource for services, we can solve these problems.In this thesis, we introduce a management model for services running on Cloud to solve the problem that distributed services do not have a universal management standards and their management is difficult. At the same time, in-depth research is proceed in some area, including of Configuration management research and Performance management research.(1) Management model for Web services running on CloudBecause traditional internet services are distributed, the management of services is not universal. In this thesis, we firstly construct a model for management based on the third part for Web service running on Cloud. This model is composed of Configuration Management, Performance Management and Fault Management. In detail, we introduce the attributes and operations for management and validate the effectiveness of the model by constructing a prototype system. Thinking of Crossed-Clouds, we also design a system management model for Web services running on different Clouds.(2) Configuration management research for Web servicesAlthough Web services based on UDDI as Register Center have been widely accepted, however, with the increment of applications, the Register Center based on UDDI has exposed many defects. In this thesis, we extend Register Center. Firstly, we introduce services' performance states with time changing to increase credibility of the Register Center.For finding services that are not registered in system better to manage system, we use DPI technology to find services and write the services' name into services' information database.At the same time, we use DPI to identify services relationship for scaling performance faster and making services' selection easier.For satisfying the need of users, we also extend UDDI by providing a service selection algorithm based on society relationship and users' utility. We take society relationship of services as parameters for selecting services and calculate value to sort order so that we can select the right service. For this method, we introduce services' society relationship and obscurity of users' selection to develop selection algorithm.(3) Performance management research for Web services running on CloudThinking of every service's provider and Cloud operator hoping to get more benefits, we should maintain services performance and decrease resource consummation. In this thesis, we design a regulative framework. For composed services, we solve service performance bottleneck location and resource deficiency identification. Lastly, for regulating faster, we design a workload prediction based performance state shifting.(4)Construction of prototype systemAt last, associated with the research of key technology, we design a management system for Web services running on Cloud based on Open Nebula. For this system, we implement Configuration management and Performance management for Web service and validate key technology for managing Web services by experiments.