摘要
基于网格资源的异构性和不同所有权现象,为满足中国教育科研网格监控子项目CGSV(chinagrid supervision)中用户的模糊监控需求,提出一种基于语义关联的网格监控元数据管理方案。在局部范围内对资源及其监控元数据分别进行逻辑和语义建模;在全局范围内关联异构资源模型的监控元数据,建立语义关联网络,并依据关联深度来进行监控元数据的查找。实验结果表明:实现语义关联的CGSV组件(注册服务)每秒钟能处理120次以上的关联查找,验证了该方案的有效性。
A metadata management scheme based on semantic links was developed to meet the vague grid monitoring requirements for users in the chinagrid supervision （CGSV） project. The heterogeneity and various ownerships of grid resources require logically modeling of the physical resources with monitoring metadata characterized by shaping specific monitoring metrics into several semantic dimensions. Metrics and semantic dimensions are further linked together across various resource models. This results in a semantic link network of grid monitoring metadata at the global level with a query mechanism based on semantic distance to locate monitoring metadata. Tests show that the semantic link registry for the chinagrid supervision project can handle more than 120 requests per second; thus, the scheme is effective and practicable.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2008年第1期109-112,共4页
Journal of Tsinghua University(Science and Technology)
基金
国家自然科学基金资助项目(90612016
60673152)
国家“九七三”基础研究基金项目(2004CB318000
2003CB316907)
关键词
计算机网络
网格监控
元数据
computer network
grid monitoring
metadata
