数据缓冲区自适应优化算法的模拟研究

摘要

To decide the data buffer pool size and direction when DBMS running in complex workload, adaptive optimization algorithm based on simulation is proposed, which decide direction by building simulated buffer pool (SBP). By buffer miss equation (BME), determine tuning size is decided by some critical data, which is obtained by monitoring database combined adaptive optimization algorithm based on simulation. The experiment shows this algorithm can improve the overall performance, and keep performance stable in long time.%为决定数据库在复杂负载下运行时,数据缓冲区自适应优化所要求的大小与方向,提出了基于模拟仿真的自适应优化算法,通过建立模拟数据缓冲区(simulated buffer pool,SBP)来决定调整的方向.应用缓冲错失公式(buffer miss equation,BME)结合基于模拟仿真自适应优化算法监测数据库运行状态时所获得的一系列关键数据,来确定优化调整的具体方案.实验结果表明,该算法提升了数据库整体性能,并在长时间内保持性能高效稳定.