Crown: A service grid middleware for e-science

In the past few years, the Grid computing paradigm has emerged as an instance of cyber infrastructure, promising to enable resource sharing and collaborating across multiple domains. In the research community there has been an intense interest in designing and studying of such system. CROWN (China R and D Environment Over Wide-area Network) project is an e-Science project funded by China Natural Science Foundation Committee, and China 863 High-tech Program. The main goal of CROWN project is to empower in-depth integration of resources and cooperation of researchers nationwide and worldwide. CROWN was started in late 2003. The main goal of CROWN project is to build the middleware infrastructure and wide area testbed to support computation intensive, data centric e-Science applications. Recently, with the evolution of Web services, the service-oriented architecture has become a significant trend for grid computing, with OGSA/WSRF as the de facto standards. CROWN has adopted the service-oriented architecture, connecting large amount of services deployed in universities and institutes. Up till mid 2007, lots of applications in different domains have been deployed into CROWN grid, such as gene comparison in bioinformatics, climates pattern prediction in environment monitoring, etc. The long-range goal for CROWN is to integrate home user resources in a fully decentralized way with a robust, scalable grid middleware infrastructure. In this chapter, based on a proposed Web service-based grid architecture, a service grid middleware system called CROWN is introduced. As the two kernel points of the middleware, the overlay-based distributed grid resource management mechanism is proposed, and the policy-based distributed access control mechanism with the capability of automatic negotiation of the access control policy and trust management and negotiation is also discussed in this chapter. Experience of CROWN testbed deployment and application development shows that the service-oriented middleware can support the typical scenarios such as computing-intensive applications, data-intensive applications and mass information processing applications.