Study Of Distributed Data Mining Architecture Based On Grid

While information technology is applied in every field of the human society quickly, people regard collecting data as an important affair, and build a lot of databases used on the institution of business, government, education, and scientific research. Moreover, for the sake of picking out useful information from these data resource, the researchers raise data-mining technology and distributed data-mining technology. The first one may automatically find out the knowledge from databases, the last one implement data-mining on distributed technology. Currently, the distributed data-mining is the primary form.On the other hand, as a new distributed technology, the grid technology has been mature. It provides an efficient managing way of distributed resource, strong computational power, excellent system expansibility, breaks the limit of computational power, storage, resource distributing, the way of sharing resource. It should be a fire-new experiment combining grid and distributed data-mining technology. As a result, this paper do research on the architecture of the distributed data mining, proposes a solution of distributed data-mining based on grid, namely, above the grid layer, a new layer of distributed data-mining is build as a universal data-mining platform.Associated technologies are researched before chapter 5 in the paper. Firstly, distributed data mining are discussed and analyzed, and the problem occurred on current distributed data mining system is proposed. Secondly, the grid technology is analyzed summarily, including its concept, goal and main application. The, Web Service Resource Frame is discussed in detail, as well as the grid implement project—Globus Toolkit 4.It is the design part of the distributed data-mining layer in the chapter 5. The architecture of distributed data-mining on grid is analyzed. Distributed data-mining on grid is brought out. In addition, the components of the whole architecture, global web service resource, local web service resource, algorithm web service resource and datamark Web service resource are designed respectively, including interface defining, service workflow.Finally, all web service resources are implemented. The resource properties documents of Web Services are Described, the operation interfaces of Web Services are implemented, and the static topologies of Web Service resources are protracted by UML. Furthermore, a local grid is built through installing the grid middle ware on some computers linked together. And on which a mining instance running successfully prove the feasibility of the proposed architecture in the paper.