Research On Service Component Dependency Relationship Resolving And Applications

With the popularization of Internet and the rapid development of the software application technology and mode, software has shown a development trend of the network, platform and service. Wherein, Software as a Service (SaaS) is a new type of Software Service mode, and has changed the user use the Software and Software vendors provide Software mode, with advantages of on-demand rent, do not need to maintain and easy to extend. In addition, service-oriented computing (SOC) and service oriented architecture (SOA) have been widely accepted, and become the computing paradigm and architecture of constructing the next generation distributed application, supporting for low cost and rapid development. The application software system is composited of the distributed, domain-independent and diverse-grained services. Service composition links the available distributed services together to implement business function to solve the tenant demand problem, promote the rapid development, reuse services, and realize dynamic and changeable business of multi-tenant application.In multi-tenancy applications, tenants can customize and adjust the corresponding services shared by tenants based on their business requirements with strong elasticity and expansibility. Service components are equivalent and shared by tenants. On the premise of load balancing, service is a single instance used by multiple tenants at the same time. There are several problems to be solved in the research of service registration, service management and service composition. Firstly, a large number of services emerge in the Internet, which will inevitably affect the efficiency of service discovery and the efficiency of the tenant application building. Secondly, as services do not exist independently, service management and mining information among services are of great importance. Thirdly, product update cycle becomes shorter, the market response requirements increase and personalized demand of the application system is higher and higher. In service resources sharing mode, we also need to satisfy the different, dynamic and changeable business requirements of mass tenants.This thesis bases on the theory of service computing, and comprehensive study of service dependency and service change through service components and dependency relationship in the hierarchical service model. The main contributions of the thesis are as follows:1. The system definition and description of service dependenciesServices dependencies of the service component in the hierarchical multi-tenant model do not exist alone, which are the incidence relation hidden among service components. KIF with description semantics can support any logical sentences, describe all kinds of logical relationship and implement logic understanding. This thesis mine service dependencies effectively, and descript the relative definitions and concepts of service dependencies, service dependency effect and service dependency type by KIF.2. Hypergraph-based service dependency resolving algorithmsBased on the theory knowledge and algorithms of graph and bipartite graph, several service dependency resolving algorithms are given, supporting for service composition, service changes in the process of the subsequent service model instantiation.3. A service composition algorithm based on service dependenciesOn the basis of service dependencies resolving, this thesis presents the steps and procedure of service composition. Under the guidance of service dependencies, existing available small particle size, distributed and loosely coupled services components mix into the large particle composite service to meet the business requirements of tenants.4. Service dependency-based service changesAccording to the service dependency graph and characteristics of service changes in hierarchical service model, we propose the service change propagation graph to research on the propagation direction and effect range. Service changes can be divided into functional changes and non-functional changes, and manage the change of service components evolution by using the services version control method. Service versions are transparent. We implement dynamic references of different versions with the aid of service component service binding tools. Service changes can make the adjustment of the service is fast and efficient, to adapt to the change of the environment and business development, meet customer demand timely, and improve the combination service dynamic adaptability and fault tolerance.In the thesis, we manage the abundant service components reasonably, mine the dependency relationship between services effectively, and research on the service dependency resolving, service composition and service changes. We apply the research achievements to the cloud service platform construction, and guide building, operation and maintenance of multi-tenancy application service. Other research and techniques, such as performance and fault tolerance of service components and application services and so on are the future work.