Research On Partitioning Technologies Of Resource-constrained Web Service Applications

With the coverage and popularization of mobile wireless network, it becomes more and more favored by people to use mobile devices and embedded smart devices in mobile wireless network, which leads to the problem of how to achieve the interconnection of heterogeneous devices in the wireless mobile environment. As a kind of autonomous and open application entity, Web service has the characteristics of loose coupling, platform independence, strong interoperability etc, which is especially suitable for the release and use in the Internet environment. However, in the wireless mobile environment, the CPU computing power, network bandwidth, memory capacity, battery capacity and other resources are very limited. In this resource constrained environment, how to divide and deploy Web service application properly and achieve the goal of saving resources and improving efficiency is a hot problem of current researches.Based on the analysis and research of Web service, resource-constrained Web service and application partitioning, combining with the existing research results of our research group, this paper does a deep research on resource-constrained Web service application partitioning techniques. The framework of resource-constrained Web service application partitioning is designed. The detailed description and execution flow of each module in the framework are presented. This paper focuses on the partitioning algorithm of resource-constrained Web service application based on network flow graph. The resource-constrained Web service application is firstly transformed into weighted directed-graph. Then the concept of network flow graph is introduced and the weighted directed-graph is further transformed into network flow graph. Eventually, the problem of resource-constrained Web service application partitioning is transformed into the minimum cut problem of network flow graph and in this paper Dinic algorithm is adopted to solve the problem. From the aspects of optimal execution time, the lowest execution energy consumption and the lowest network traffic volume of resource-constrained Web service application, the parameters design of single objective optimal partition and multi-objective optimal partition are given. Finally, a Web service application case of traveling is presented to verify the validity of the partitioning algorithm of resource-constrained Web service application. The work in this paper has some theoretical and practical value for realizing resource-constrained Web service application.