基于云平台的弹性Web集群扩缩容机制的研究

Research on Elastic Web Cluster Capacity Expansion Mechanism Based on Cloud Platform

弹性Web集群即随着工作负载的变化,集群规模可以弹性变化的Web集群。本文研究弹性Web集群的扩缩容机制,首先分析现有扩缩策略的常用算法的优缺点,然后当前策略的不足,提出新的扩缩策略,使用更丰富的指标进行负载预测,提高预测的准确性;使对负载评估的结果带有预测性,相对于简单的非预测性算法可以更早地做出扩容决策,最终支撑Web应用系统对外提供高可用和高性能的服务。

Resilient Web Cluster is a Web cluster with flexible scales as the workload changes.In this paper,westudy the expansion and contraction mechanism of flexible Web clusters.Firstly,we analyze the advantages anddisadvantages of the commonly used algorithms for existing expansion strategies.Then,we propose a new expansionand contraction strategy based on the current strategy.We use more abundant indexes to forecast the load toimprove the accuracy of the result of load evaluation predictive.Compared with a simple non-predictive algorithm,it can make the expansion decision earlier,and ultimately support the Web application system to provide high availabilityand high performance services.