Research On Some Key Technologies Of Grid Service Application Model

With the expansion of the scope of shared resources and the demand for the development of business enterprises, grid service computing, which is characterized by its capability for safely sharing cross-organizational resources and providing problem solving environment for virtual organization, has become an important trend. The demand of industry and commerce request for service application system based on user demand for rapid and flexible integration of various existing and new business, make them collaborating to afford on demand service. How to open, dynamic and difficult to control the grid environment to achieve a wide range of services to share and integrate resources and the development of complex business applications, has become a challenging research topic of computing grid services facing. Distributed grid workflow sharing and reuse technology has become a core technology of intelligent problem-solving. Based on customer demand as the center of service-oriented architecture (SOA) designed to study grid workflow technology, many scholars also have made some encouraging results, but did not make full use of the advantages of SOA, the results is very difficult to adapt to the current real-time applications demand ever-changing objectives. Effect for the realization of the rapid restructuring and upgrading of the business processes was not obvious. Therefore, in order to achieve resource sharing and collaboration shield the complexity of the integration of on-demand services and to improve the modern enterprises ability of on-demand, the in-depth study of dynamic construction technology for grid service application are very necessary.From the perspective of grid service-oriented architecture (SOA), this paper based on artificial intelligence planning, Particle Swarm Optimization, Contract Net Negotiation Protocol, workflow management technology as well as the combined Identity certification, focus on the following critical theory and technology of grid service applications:(1) Grid Workflow Merging Model Based on user demand We put forward the grid service workflow model based on user demand, which is composed of the grid services organizations model, the business concept model, business logic model, custom command model, business scheduling model and business enactment model. In order to improve the retrieval grid service and efficiency, our proposed grid service organizations model made up of grid service registration system(GSRS) and grid service spanning tree (GSST), and we are given a building algorithm for GSST and a business logic spanning graph algorithm. Throughout the in-depth study of planning methods, we use artificial intelligence planning techniques to give a modeling method for grid service workflow.(2) Grid Service System Architecture Self-adaptive grid contract computing management (SACCM) system includes application layer, grid contract computing management layer and the application proxy layer. Its core is grid computing contract management layer, which have four main components: contract computing services workflow manager, contract computing services manager for negotiating, C_SLA life cycle manager and contract computing services restructuring. We give grid services failure recovery algorithm for workflow from composing workflow instance level and workflow relations level.(3) Based on Particle Swarm Optimization Algorithm for Scheduling Workflow By studying the mechanism of Particle swarm Algorithm and linking grid service workflow scheduling features, according to quality of grid service, IBPSOA is given, which is an alternative path to the integer-binary coding operations, and dynamic service scheduling algorithm. IBPSOA algorithm uses services type and instances of services of the integer - binary encoding, redesigns of the location of particle swarm optimization, speed, add/subtraction and multiplication and increases the absolute value operator. From two aspects of simulation, we educe that the IBPSOA algorithm is more effective.(4) Negotiating Model of Grid Services Introducing the contract computing into the Multi-Agent negotiating mechanism, the grid contract computing model based Agent is proposed, which adds features that the use of grid services to provide a common interface and the entire process of contract management the regulatory mechanism of grid services to enhance QoS. The FIPA contract net negotiating approach with mental state of Agents is given to the achievement of contract status migration in contract-oriented computing model. Meanwhile, from the workflow point of view, it realizes a business concept workflow (BCWf) to the business logic workflow (BLWf) to the business executive workflow (BEWf) conversion. In Multi-Agent system is running continuously adjust the process entrusted with the task-related mental state; the system has good flexibility and adaptability.(5) Grid Security Architecture for Authentication /Certification based on Identity The Grid Security Infrastructure (GSI) has been accepted as the primary authentication mechanism for the Grid. However the efficiency of the grid system is still required improvement because exchanging certificates and verifying validity are compulsory in the authentication procedure. We propose a Combined Public Key Cryptosystem- CPK-CCC and a novel Authentication Architecture for Grid Security (AAGS), which adopt seeded secret key and combine key technique and employs the features of conic curves over the residue class ring Z_n into the authentication security, and we give precise definitions of the grid infrastructural protocols for AAGS. Compared with the SAP (SSL Authentication Protocol) in the GSI, AAGS is more efficient and convenient for its implementation than PKI.(6) Implementation of Prototype System and Application In order to verify correctness and effectiveness of the key theory and technology for our proposed application model of grid services, we developed a prototype system. We give an online travel management system for grid services as an example, which described Agent and the planning design and implementation of workflow.