Java Grid Computing, Application Programming Model

With the rapid development and widely use of Internet, enterprise applications are becoming more and more complex. Accordingly, more requirements are enforced on the design of the applications. However, the applications (data or computing) that use the power of grid computing must be programmed in accordance with the semantics of the grid infrastructure. Therefore, applications are encumbered with an additional programming model and set of APIs that must be adhered to. Thus the code becomes bulkier and interspersed with calls that interface with the grid infrastructure. The inadequate response to this, we use Java identity of cross-platform and proposed a framework -GPMJ. GPMJ is a framework that abstracts the details of interfacing with a grid infrastructure, resulting in a simple programming model that can be of substantial aid in the quicker development of maintainable grid client applications.First of all, the author analyzes the concept of grid computing, the progress of the study and application areas. It's focused on the concept of grid. It will be impossible to start the study if the concept of grid is not clear.Secondly, the author introduces the content of Java programming language and grid programming technology and a brief summary of the model designs, which required grid application development tool of Java.On this basis, from the core technology of Java programming, Java grid programming model is put forwarded by the author. GPMJ supports two models, distributing tasks over the grid and distributing threads over the grid. The paper gives a detailed account of the design and realization of the model.The model of distributing threads over the grid using Java distributed parallel programming environment to create a master programming models, so grid application developers can also chooses a number of distributed computing model based on grid applications.The model of distributing tasks over the grid can identify suitable grid nodes, transfer the executables, input resources and dependent libraries to remote locations, start the remote job and monitor its execution and transfer the results from the remote...