Research On Technology Of Web-based API For Grid Application

With the rapid innovation of grid technology, grid is increasingly being used by ordinary users. Grid applications are developed in the direction of popularity, which has a great significance for the popularization of grid technology. However, ordinary users are unlikely to develop grid applications constrained by the complexity of grid platform and technology. Currently, there are many existed grid application program interfaces. However, ordinary users can benefit little from them since those program interfaces are mainly designed to support grid researchers.In order to solve the above problems, WAGA (Web-based API for Grid Application) provides a WYSIWYG(what you see is what you get) way for grid application programming by adopting participation, interaction and sharing features of Web2.0 technology. In this way, ordinary users are able to exploit their own Web grid applications without understanding the underlying complexity of grid. Using a demand-oriented regulating pattern to analyze the demand of grid platform, we can conclude the primary target of WAGA as follows: API is divided into non-functional and functional API. Non-functional API describes the non-functional requirements of WAGA, which dealt with some relevant issues to achieve grid application functional requirements, including: error handling, session management, context management, attribute model, monitoring model and task model. Functional API is responsible to deal with functions relating to grid application: job management, namespace management, document management, copy management, workflow management and remote procedure calls.We evaluate and analyze a WAGA-based design system. Functional testing is to validate the major functions provided by the system. The results proved that the system has a better ease of use than traditional grid applications. We proceed to the performance test from two aspects: One is a system loading performance test according to the loading time of testing system. Comparing to the traditional grid application client GPE, the WAGA-based system represents a better loading performance. The Second one is the processing time of grid task. The result shows that the efficiency can be guaranteed in the current system while enhancing the ease of use in grid system.