针对当前网格资源管理系统扩展性差的问题,提出了具有分布式本体知识库的网格资源管理模型。新模型由分布的多个自治管理域组成;自治域之间通过DHT覆盖网连接。网格资源信息以本体知识库的形式管理:本地知识库管理域内网格知识;全局知识库则提供分布本体的语义映射以及知识的索引和定位。资源发现转换为本体知识库的实例检索推理,并且支持跨管理域的资源集发现。新模型极大地提高了网格资源管理系统的可扩展性。
To improve the scalability of grid resource management system, a new resource management model, Knowledgebased Resource Management Model( KRMM), which is based on distributed ontology knowledge base, was presented. KRMM is composed of distributed autonomy domains, which connected by DHT overlay network. Resource information is maintained by knowledge base. Local knowledge base manages the knowledge of local domain, and the global knowledge base provides distributed ontology integration and knowledge indexing. Resource discovery in the Gird is converted into inference operations on the two-layer knowledge base. Furthermore, KRMM provides resource set discovery across grid domains.