Research On The Virtual View Maintenance Of Information Grid

Information Grid is a system using the Grid technologies to achieve the sharing, managing of information resources and to provide information services. This dissertation is based on the research work of Railway Information Grid, and addresses the virtual view maintenance of layered resource space model of information grid.RIG(Railway Information Grid) employs the advanced technologies of Information Grid to address the specific problems in railway industry environment, and achieve the transparency of information resources accesses by constructing the layered resource space model based on VEGA information resource space model and relational database theories.Layered resource space model leads to layered structure of metadata, in which the requests of resource access are encapsulated in metadata of virtual view layer. The fact that the information access using virtual views are based on validity of metadata mappings from virtual view layer to logical resource layer introduces the problem of virtual view maintenance.To address the invalidation of virtual views caused by evolution of logical resource layer metadata, this dissertation discusses the different change types of logical resource layer metadata as well as their potential influence on virtual view layer in detail. The RIG metadata factory is responsible for the maintenance of metadata. We propose type-based virtual view maintenance algorithms to adjust the affected virtual views, state standards of evaluating the maintenance algorithms according to the railway industry requirements.Although our discussion is placed in the specific environment of railway industry and RIG, the proposed metadata change types and view maintenance algorithms are not limited to this system.In this dissertation, we present some original ideas in following aspects:1. A set of compound evolution types of logical resource layer metadata based on relational database theory is provided, discussed in their preconditions and effects (both on metadata and logic dataset);2. A type-based virtual view adapting algorithm is presented, and evaluated in preserving virtual views'structure and maintaining the logic dataset.