基于多策略的私有云资源弹性调度方法

发布时间：2018-04-08 23:19

本文选题：云计算　切入点：基础设施即服务　出处：《计算机应用》2017年S1期

【摘要】：针对云计算环境下资源使用呈现的动态、随机、异构等特性,以及资源需求的不断提升给基础资源的管理、调度带来的巨大挑战及问题,提出了基于多策略的私有云资源弹性调度方法。首先在传统协作模型研究的基础上,通过提炼和概括云计算调度的主要组成构件,提出私有云资源调度总体框架;其次,在调度总体框架基础上,以资源的协同管理为核心,给出面向多触发条件的协同调度流程;最后,给出了触发条件中应用集群负载的计算方法。实验结果表明,提出的弹性调度方法能够自动实现虚拟机资源调度,从而优化应用集群CPU、内存、网络等性能。所提方法具有较强的可行性和实用性。
[Abstract]:In view of the dynamic, random and heterogeneous characteristics of resource usage in cloud computing environment, as well as the enormous challenges and problems brought by the continuous improvement of resource requirements to the management and scheduling of basic resources,A multi-policy based elastic scheduling method for private cloud resources is proposed.Firstly, based on the research of traditional collaboration model, by refining and summarizing the main components of cloud computing scheduling, the paper proposes a general framework of private cloud resource scheduling. Secondly, on the basis of the overall framework of scheduling, the cooperative management of resources is the core.Finally, a method of computing cluster load in trigger condition is given.Experimental results show that the proposed flexible scheduling method can automatically implement virtual machine resource scheduling, thus optimizing the performance of application cluster CPU, memory, network and so on.The proposed method is feasible and practical.
【作者单位】：上海计算机软件技术开发中心;上海市计算机软件评测重点实验室;上海海事大学信息工程学院;
【基金】：上海市科技人才计划项目(16XD1421500) 上海市科研计划项目(15511101503) 上海张江高新区专项(201505-ZB-C104-013) 上海市研发平台专项(14DZ2291200)
【分类号】：TP393.09

【参考文献】