一种基于ICFV的云计算数据库负载访问优化方法

A LOAD ACCESS OPTIMISATION METHOD OF CLOUD COMPUTING DATABASE BASED ON ICFV

为了优化云计算服务器数据库负载访问性能,采用基于特征向量的增量聚类方法(ICFV)对数据库负载访问性能进行优化。ICFV算法提取数据库负载的特征向量然后进行聚类划分,它与传统的基于负载特征向量聚类不同,采用的是增量聚类方法,不需要对所有负载集合进行重新中心距离计算,从而优化特征向量维数。通过实验证明,采用ICFV算法实现数据库负载自适应优化,提高负载聚类效率。在负载个数相同时,负载类别的增加对云计算数据库的访问性能影响较小,且ICFV方法运算效率优于CFV方法。

In order to optimise the access performance of cloud computing server database load, the incremental clustering method based on feature vector （ICFV） is used to optimise it. ICFV algorithm extracts the feature vectors of database load and then clusters and partitions them. Differing from traditional load feature vector-based clustering, it uses incremental clustering method, without the need of calculating the central distance once again on all load collections, so that optimises the dimensionalities of the feature vector. It is proved through experiment that the use of ICFV algorithm realises the adaptive optimisation of database load, and improves the efficiency of load clustering. When the load numbers are the same, the increase in load types has little impact on the access performance of cloud computing database, and the ICFV method has higher operation efficiency over the CFV method as well.