Study And Design On The Architecture Supporting Negotiation Mechanism Under Service-Oriented Architecture

Service-Oriented Architecture (SOA) is a new generation of architecture ideology applying for distributed software development. SOA has solved the problems of applications for flexibility and interconnectivity wonderfully because of its features of e reusability, and complication etc. SOA has been the hot topic at present. It will be the dominator of software architecture.In ideal SOA environment, there is an assumption that provider exactly knows what services to publish. The assumption enables SOA to be a process that is from solution to problem, not from problem to solution. Because of the assumption, in the real SOA environment, a conflict exits between provider and customer.. And some problems that we need to solve have appeared. Firstly, registry needs to earn high credibility and reject some low quality services. But also registry needs to attract more high quality services, so it can not immediately and simply reject the service on the publishing time, while decide it after negotiation. Secondly, with the increasing of services, there is many function similar services in the registry, then it is important for providers to augment their own service competitive capacity(mainly the QoS). This problem can also be solved by the negotiation between registry and provider.To solve the problems above, in the thesis I study and design the architecture of supporting negotiation mechanism under SOA which negotiate QoS between registry and provider and we also study the aspects related to negotiation.Firstly, in the thesis I propose QoS description language(QSDL), accoding to which provider can publish their QoS to registry. And we study the content of contract that is created after successful negotiation, design the storage structure of QoS and contract in the registry.Secondly, in the thesis I design an efficient and practical decision-making algorithm applying to registry. I describe the process of the algorithm and prove the efficiency and practicability through simulation experiment. Thirdly, in the thesis I redefine the concept of negotiation and improve on the messages and protocols of negotiation mechanism to be more suitable for SOA architecture. Finally, I design the overall architecture supporting negotiation mechanism under SOA, and illustrate the main modules and workflows.In a word, the problems above have been solved successfully by the architecture supporting negotiation mechanism under SOA that I studied and designed. In the real application of SOA, the designer can negotiate not only QoS, but also other issues, so the theory in the thesis that I have studied is very important in SOA field in the future.