The Research And Application On The Self-optimization Model In The Database

The promotion of computer networks and the gradual rise in the number of users, the resource utilization and system response time supposed a new challenge to the database optimization. The type of traditional database optimization can’t meet the requirement to the application, and database optimization techniques have recently become a hot research topic.The paper analyzes and improves the research database optimization techniques with the combination of autonomic computing techniques. It achieves the classification of the dynamic load during the database running by the decision tree algorithm. At the same time, we improve a self-optimization model framework choosing the oracle database as the experimental platform and the memory resources as the Turing object. At the last we design and imply each module in the framework of the model. The main work in this paper is flowing.At first, we summarized the autonomic computing and the database optimization techniques and analyzed the influence to the database memory resource utilization from the type of the load. We improved a framework model of self-optimization to the database memory resources.With the help of benchmark factory, we simulated the two database loads operation and obtain the experimental sample data. Using the decision tree algorithm, we derived the decision tree of the database self-optimization model and verified the accuracy of the number of classification decision tree on the experimental platform.We deigned and implied a self-optimization model for the memory resource of the oracle database. First the monitor component monitors the current system performance of the database system. The analyzer component analyzes the operation performance of the current system. The performance analysis and the load classification results as the input to the planner to the planning model and get the optimization strategy. At last, the resource optimization regulator runs the optimization strategies. The whole optimization processes are supported by the knowledge’s component.According to the relations between the load type and the memory resource utilization, we verified the self-optimization model of the database memory resource with the composite load scene in the Benchmark factory tool. The experiment results show that the precise division of the type of load and the real-time optimization regulation to the memory resources has a significant improvement on the resource utilization and system response time to the Oracle database.