Research On Modeling And Discovery Mechanism Of FDR In FDG

Grid is hot topic in distributed computing field, it aims to achieve the computing resources, storage resources, information resources, knowledge resources overall sharing. Using Grid in remote collaborative fault diagnosis can be a very good solution to diagnostic resources island issue and can achieve the diagnosis resource sharing and collaborative working. The characteristics of Fault Diagnosis Grid (FDG) were analyzed in this paper. The grid resource management is the one of the core problems of FDG, FDG resource management makes up of Fault Diagnosis Resource (FDR) description, FDR discovery, FDR location and so on. In the middle of them, that FDR discovery, FDR location are implemented needs a foundation—reasonable FDR description. FDR description and FDR discovery are the key problem of resource sharing and collaborative working. Therefore, this paper mainly focuses on the method about FDR description and then mechanism of resource discovery. This paper makes following contributions:1. Proposed the framework of FDG system. Paper proposed FDG architecture based on the OGSA framework, in combination with RCFD features, and analyzed its business processes. the diagnosis of the characteristics of diagnostic resources, proposed diagnostics grid resource management to meet the functional requirements of the FDG resource management system proposed diagnostics grid resource management based on the characteristics of diagnostic resources.2. This paper has introduced a method of classifying Fault Diagnosis Resource based on service type in grid, then FDR classified model was established by UML tools. Fault Diagnosis Resource modeling and uniformly describing was based on combining UML, XML and RDF. Fault Diagnosis Resource encapsulation is presented based on WSRF, which shields complexity and heterogeneous of FDR and makes a Web service-resource structure of FDR for Grid resource service.3. Proposed diagnosis tree-peer resource discovery layered network architecture. This paper has introduced a method for the aggregation of resources by resource properties similarity, and this method can achieve load balancing. This paper proposed an improvement of DHT algorithm to meet the category search and exact match of FDR. Experiments show that this method can effectively improve the efficiency of resource discovery.