Research On Scientific Data Grid Service Discovery Architecture And Its Key Problems

Service-Oriented Architecture views the Internet as a large set of services with extensible set of well-defined interfaces. The evolution of grid system shows this trend. In the Open Grid Services Architecture, a grid service instance is a service that conforms to a set of conventions (expressed as WSDL interfaces, extensions, and behaviors) for such purposes as lifetime management, discovery of characteristics, notification, and so forth. And grid is viewed as a set of extensible grid services.Grid services discovery is the key technologies that enable grid application. Based on Scientific Database and Scientific Data Grid project, This dissertation research the problem of information, resource and service discovery in large scale distributed system which spanning multiple administrative domains. Grid service discovery technology supports the integration of heterogenous and dynamic database nodes in large scale distributed system such as grid. Through simple, efficient query language, user can discovery and locate information, resources and services that meet their requirement on interface with uniform view.This dissertation introduces Scientific Data Grid Service Discovery Architecture (SDGSDA). SDGSDA integrates standards and promote interoperability. It covers disparate concepts, interface and protocols under the same umbrella. This thesis also researches in depth the problems which are important to implement service discovery, including dynamic data model of grid service information, which make the distributed grid services information database express logically as a single view that contains all the collections of grid services information; the soft-state mechanism used for consistency maintenance. We designed the grid service information distribution protocol, which implements the strong coherence and weak coherence cache policy; we propose a new query language, structured keyword query language, which is simple and efficient. Combining research works of database and information retrieval community, we devise query engine for grid service information, which can index, query grid service information efficiently, at the same time, which support partial query matching and ranking of query results; finally, we initially explore the problem of how to view the distributed, autonomic database of grid service information as a whole to support uniform service discovery.