一种负载预测感知的虚拟机合并与迁移策略

VIRTUAL MACHINE CONSOLIDATION AND MIGRATION STRATEGY BASED ON WORKLOAD PREDICTION-AWARE

虚拟机合并和迁移可以有效降低云数据中心的能耗并提高资源利用率。然而,已有算法多集中于根据当前的资源需求最小化活跃主机数量,忽略了负载变化情况下的未来资源需求,这样会生成过多无用虚拟机迁移,增加SLA违例风险。为了解决这一问题,同步考虑当前和未来的资源利用请求,提出一种新的虚拟机合并算法。该算法利用回归模型对主机和虚拟机的CPU占用进行预测,在虚拟机迁移源主机和目标主机的选择上,同步考虑了当前超载和预测超载问题,较好地避免了无用虚拟机迁移。通过不同类型的负载对算法进行了实验分析。结果表明,与基准的启发式算法和元启发式方法相比,该算法不仅可以降低主机能耗,还可以同步减少虚拟机迁移量和降低SLA违例。

Virtual machine consolidation and migration can effectively reduce energy consumption and enhance resource utilization.However,the existing algorithms mostly focus on the number of active physical hosts minimization according to their current resource requirements and neglect the future resource demands,which can lead to excessive unnecessary virtual machine migration and increase the risk of SLA violations.To solve this problem,a new virtual machine consolidation algorithm is presented that takes into account both the current and future utilization of resources.Our algorithm predicted the CPU workload of the host and virtual machine simultaneously through the regression model,and then synchronously considered the current overload and predicted overload in the selection of the virtual machine migration source host and target host,thus avoiding the useless virtual machine migration.The algorithm was simulated by the different types of workload.The results show that compared with other heuristic and meta-heuristic algorithms,our algorithm can not only reduce the overall energy consumption of the host,but also synchronously reduce virtual machine migration number and lower the SLA violations.