Research Into Mechanism For Improving Large-scale Database Manipulation Efficiency Based On Software Agent

With the development of the information technology in the modern enterprise, the manipulation efficiency of such large-scale database can no longer meet the needs of production. However, there are some defects in the traditional optimization methods, for instance, too scattered to integrate, excessive reliance on manual operation and personal experiences, poor adaptability to the structure of distributed system. Moreover, the traditional software engineering methodology can not provide enough support to design such complex, dynamic and open system. Thus, a new manipulation mechanism for large-scale database is needed to solve the problem of low manipulation efficiency.Based on the theories of software agent and agent-oriented software engineering, an agent-based model of improving manipulation efficiency to distributed database system is presented in this paper. The organizational structure of the model, various agent design, collaboration mechanism among the agents, agent communication language based on KQML, and also wrapper of legacy database are researched and solved. This model has the abilities to interact with the environment, and to integrate with existing optimization methods. Meanwhile, intelligence, initiative and extendibility are embodied in the model. The critical part of the system, which truly improves the efficiency, is the internal algorithms. In this paper, an agent-based distributed semantic caching scenario is designed based on existing semantic caching technique. Then, in order to further the effect of caching, an intelligent pre-fetch algorithm and its application in the model are introduced. Finally, aiming at the application features of oil field, an instance system is designed and implemented, which proves the feasibility and validity of the mechanism introduced above.The agent-based mechanism for improving large-scale database manipulation efficiency brings forth new ideas in traditional succession of heritage system and data retrieval of database, which also compensates the defects of the traditional optimization technologies, such as lack of intelligence, pre-mobility, integration and reuse-difficulties. Furthermore, it enriches the database performance optimization research field, and makes meaningful explorations on this field.