基于网格的数据挖掘体系结构  

Data Mining Architecture Based on Grid

作　　者：郑世明[1] 邵荣明 苗壮[1] 

机构地区：[1]解放军理工大学指挥自动化学院,南京210007 [2]解放军沈阳炮兵学院教保科,沈阳110867

出　　处：《指挥信息系统与技术》2011年第5期41-46,共6页Command Information System and Technology

基　　金：国家高技术研究发展计划("八六三"计划)资助项目(2007AA01Z126)

摘　　要：由于现有的网格体系结构异常复杂,技术细节过于全面,因此,难以权衡其整体性能,大规模的资源共享和协同并不理想,难以满足高可用性、安全性和互操作性等要求。结合开放网格服务体系(OGSA)的思想,采用分层描述的方法,围绕功能、结构、调度成本和网格服务目标等,设计了一种五层的网格数据挖掘体系结构(GDMA)。该体系结构以服务为核心,通过统一的网格服务接口屏蔽资源的异构性,并针对用户的特殊需求,设计了基于工作流和W eb服务的数据挖掘客户端。As the existing grid architecture is highly complex with too many technical details, its overall performance is hard to balance, and the sharing and the collaboration of resources in large scale are unsatisfactory, thus it cannot meet the requirements for availability, security and interoperability. Combining the conception of the open grid service architecture （OGSA） using the hierarchical descriptive method, with the functionality, the structure, the scheduling cost, and the objectives of the grid service as the focus, a grid data mining architecture （GDMA） with five levels is designed. Centering on the service, the architecture overcomes the heterogeneity of resources by unified grid service interfaces. Data mining clients based on the work flow and the Web service are designed based on users＇ special needs.

关 键 词：网格数据挖掘体系结构 开放网格服务体系 数据挖掘 网格计算 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]