Research On Adaptive Self-tuning Data Buffer Based On The Simulation Technology

The Adaptive Self-tuning algorithm based on Simulation technology, while working to ensure that the database to collect data buffer operating characteristic data, to develop and implement optimization. To avoid a series of problems of the traditional manual way. Meanwhile, the data buffer as an important part of the database memory architecture, the algorithm can be used for self-management of the database in memory.In order to achieve the load in a complex environment, the data buffer size dynamically adjusted so as to achieve the purpose of enhancing performance database. In this paper, have the following work:1.Dfferent adaptive optimization algorithm is compared and studied by experiment,while under different load types of data buffer working conditions were observed and analyzed, summarized the data buffer adaptive optimization is the key to distinguish the data buffer missed in the type,and search the relationship of data buffer size and the buffer miss ratio.2. Based on research results, the working principle of the data buffer, the use of simulation technology to create simulation data in the memory buffer (Simulated buffer pool, SBP), memory the data blocks which droped from the data buffer in the unique identifier to simulate the working status of the expanded buffer pool to determine the different types of buffer miss, compared to the current load in the actual performance data buffer, for a number of key calculations. And according to the simulation monitoring results combined with buffer miss equation (Buffer miss equation, BME) determining adjustment plan .In order to achieve adaptive of the data buffer, and according to the above algorithm to establish the corresponding adaptive optimization system.3.Working with AS3AP benchmark to simulate a complex load environment for database . Respectively, before and after optimization of database performance testing, we compare the databases system with Automatic Shared Memory Management System (ASMM), and analysis the results. To improve the insufficient aspects of this study.The results of this experiment show that the simulation method is indeed able to achieve optimizing adaptation of the data buffer. Before the same database is not optimized, the data buffer miss rate remained at a lower level. Through study, alogrithm are capable of making the data base work in a more efficient way and improve the database performance..