一种低成本的云资源提供算法
发布时间：2018-11-22 08:42
【摘要】：针对云资源提供成本较高的问题,结合预留计划和按需计划,提出了一个两阶段的云资源提供算法,以降低服务提供商的运营成本。在预留阶段,根据历史数据估计工作负载需求的分布,将长期的资源预留问题规划为两阶段随机规划问题,并转化为确定性的整数规划问题进行求解;在按需阶段,采用基于卡尔曼滤波的工作负载预测算法预测工作负载需求,并且将按需资源提供问题规划为整数规划问题进行求解。仿真结果表明,所提出的云资源提供算法节省了超过25%的运营成本,同时保证了较低的服务等级协议违约率。
[Abstract]:In order to reduce the operating cost of service providers, a two-stage cloud resource provision algorithm is proposed to solve the problem of high cost of cloud resource supply, combined with reservation plan and on-demand planning. In the reservation stage, the distribution of workload demand is estimated according to the historical data, and the long-term resource reservation problem is programmed into a two-stage stochastic programming problem, which is transformed into a deterministic integer programming problem to solve the problem. In the on-demand phase, the workload prediction algorithm based on Kalman filter is used to predict workload requirements, and the on-demand resource provision problem is solved as an integer programming problem. The simulation results show that the proposed cloud resource provision algorithm can save more than 25% of the operating cost and ensure a lower service level agreement default rate.
【作者单位】： 南通大学电子信息学院;
【基金】：国家自然科学基金资助项目(61403216) 江苏省高校自然科学研究面上资助项目(15KJD520002)
【分类号】：TP301.6
[Abstract]:In order to reduce the operating cost of service providers, a two-stage cloud resource provision algorithm is proposed to solve the problem of high cost of cloud resource supply, combined with reservation plan and on-demand planning. In the reservation stage, the distribution of workload demand is estimated according to the historical data, and the long-term resource reservation problem is programmed into a two-stage stochastic programming problem, which is transformed into a deterministic integer programming problem to solve the problem. In the on-demand phase, the workload prediction algorithm based on Kalman filter is used to predict workload requirements, and the on-demand resource provision problem is solved as an integer programming problem. The simulation results show that the proposed cloud resource provision algorithm can save more than 25% of the operating cost and ensure a lower service level agreement default rate.
【作者单位】： 南通大学电子信息学院;
【基金】：国家自然科学基金资助项目(61403216) 江苏省高校自然科学研究面上资助项目(15KJD520002)
【分类号】：TP301.6
【相似文献】
相关期刊论文 前4条
1 张治斌;刘丽丽;;基于信誉机制的云计算资源预留算法研究[J];计算机仿真;2013年10期
2 胡志刚,吕祯恒;一种基于动态资源预留的任务映射算法[J];计算机应用研究;2005年07期
3 祁超;张t，
本文编号：2348734
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2348734.html
