A new demand-oriented and fine-grained resource elastic scaling method which is based on the benefit analysis model was proposed from the angle of improving load-adaptive capacity and resource utilization of the cloud system by researching on critical issues of building elastic cloud platform. In order to maximize the scalability of the system and eliminate the limiting of single point of centralized cluster,the resource hierarchical management strategy was adopted to achieve the resource management architecture that combines cluster vertical scaling and inter-cluster horizontal scaling;in order to improve system resource utilization and overcome the waste of resource of the traditional virtual machine,the lightweight virtualized container was used to realize the fine-grained resource allocation;in order to ensure the real-time performance of the system and avoid the lag effect,the method of benefit estimation was applied to realize resource pre allocation and load balancing. Theoretical analysis and simulation results showed that this method can effectively realize the resource elastic scaling of cloud platform,reduce job waiting time,and enhance the resource utilization.%针对构建弹性云平台的关键问题展开研究，从提升系统负载自适应能力与资源利用率角度，提出一种基于效益分析的面向作业需求的细粒度云平台资源弹性伸缩方法。为最大化系统可伸缩性，消除集中式集群单点限制，采用资源分层管理策略，实现集群内垂直伸缩与集群间水平伸缩相融合的弹性资源管理架构；为提高系统资源利用率，克服虚拟机方案的资源浪费问题，采用轻量级虚拟化容器，实现面向作业的细粒度资源分配；为保证系统动态资源调整的实时性，避免滞后效应，采用效益预估方法，实现资源预分配并兼顾负载均衡。理论分析与仿真实验表明，本方法能有效实现云平台资源弹性伸缩，减少作业等待时间，并显著提升系统资源利用率。
展开▼
