云平台下实验室数据库资源负载优化控制仿真

发布时间：2018-03-23 03:26

本文选题：云平台数据库　切入点：负载控制　出处：《计算机仿真》2017年07期 　论文类型：期刊论文

【摘要】：对云平台下实验室数据库资源负载进行优化控制,可提高实验室数据的管理效率。进行资源控制时,应将实验室数据库资源损耗特征按性能进行分类后,通过提高实验室数据库资源的吞吐量和整体稳定性完成控制,但是传统方法通过控制输入给数据库的负载数量来提高数据库系统的稳定,但是忽略了实验室数据库资源的吞吐量和整体稳定性,降低了数据库资源负载控制的性能。提出一种云平台下实验室数据库资源负载优化控制方法。首先,运用基于多目标的优化算法以数据库负载特征参数为负载控制的依据,根据各个负载具有的不同属性特征,获取每个负载控制时间距离内价值大且平均响应时间少的负载,确定数据库负载控制的形态特征。其次利用改进特征向量增量聚类算法,依据负载控制的形态特征对负载进行分类,采用动态聚类形式提取负载状态特征,将具备同类性能特征和数据库资源耗损特征的负载分为一类,通过负载控制以提高数据库系统的吞吐量和整体稳定性,最终实现云平台下实验室数据库资源负载优化控制。仿真结果表明,通过确定负载形态特征,然后对形态特征进行分类后实现了数据库的负载控制,优化了云平台虚拟实验室数据库资源负载的控制性能。
[Abstract]:The optimal control of laboratory database resource load under cloud platform can improve the management efficiency of laboratory data. In resource control, the resource loss characteristics of laboratory database should be classified according to its performance. By improving the throughput and overall stability of the laboratory database resources, the traditional method improves the stability of the database system by controlling the amount of load input to the database. However, the throughput and overall stability of laboratory database resources are ignored, and the performance of database resource load control is reduced. Based on the multi-objective optimization algorithm, the load characteristic parameters of the database are taken as the basis of load control. According to the different attribute characteristics of each load, the load with large value and less average response time is obtained within the control time distance of each load. Secondly, the improved feature vector incremental clustering algorithm is used to classify the load according to the morphological characteristics of the load control, and the dynamic clustering method is used to extract the load state feature. The load with the same performance characteristics and the characteristics of database resource consumption is divided into one class, which can improve the throughput and overall stability of the database system through load control. The simulation results show that the load control of the database is realized by determining the morphological characteristics of the load and then classifying the morphological features of the database. The control performance of resource load of virtual laboratory database on cloud platform is optimized.
【作者单位】：山西大学商务学院;
【分类号】：TP311.13;TP393.09

【参考文献】