摘要
如今,基于云计算的软件服务对自适应资源分配提出要求,这种分配可以根据需要动态调整资源,以保证良好的服务质量(QoS,Quality of Service)和低成本资源.然而,在复杂波动的负载环境下,以具有成本效益的资源量来分配资源并满足QoS是一个挑战.本文引入一种同时考虑当前工作负载和未来工作负载变化的自适应资源分配策略,该策略在预测资源分配操作中,以QoS预测模型为基础,使用面向负载时间窗口的方法,将当前负载以及窗口内未来的负载加入资源分配方案计算过程中,最后使用基于PSO-GA的运行时决策算法来搜索合理的资源分配方案.我们在RUBiS历史数据的基准上评估了我们的方法.实验结果表明,基于本方法的云应用资源分配的有效性有所提高.
Nowadays,the emerging software services based on cloud computing require adaptive resource allocation,which can dynamically adjust resources according to needs to ensure good quality of servie(QoS)and low-cost resources.However,in the complex fluctuating load environment,it is a challenge to allocate resources and meet QoS with cost-effective resources.This paper introduces an adaptive resource allocation strategy which considers both the current workload and the future workload changes,this strategy based on the QoS prediction model,uses the load time window oriented method to add the current load and the future load in the window to the computing process of resource allocation scheme,and then uses the runtime decision algorithm based on PSO-GA to search for a reasonable resource allocation scheme.We evaluated our method on the benchmark of RUBiS historical data.The experimental results show that the efficiency of resource allocation based on this method is improved.
作者
杨立坚
陈星
黄引豪
YANG Li-jian;CHEN Xing;HUANG Yin-hao(College of Mathematics and Computer Science,Fuzhou University,Fuzhou 350108,China;Fujian Provincial Key Laboratory of Network Computing and Intelligent Information Processing,Fuzhou 350108,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2021年第5期953-960,共8页
Journal of Chinese Computer Systems
基金
国家重点研发计划项目(2018YFB1004800)资助
国家自然科学基金项目(61972165)资助
福建省自然科学基金项目(2019J01286)资助
福建省高校杰出青年科研人才计划项目.
关键词
云计算
自适应资源分配
QoS预测模型
粒子群算法
cloud computing
self-adaptive resource allocation
QoS prediction model
particle swarm optimization
