arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1906.05008eess.SYcs.SY
收藏

边缘计算中虚拟化计算平台的自适应资源管理

Adaptive Resource Management for a Virtualized Computing Platform within Edge Computing

Thembelihle Dlamini, Ángel Fernandez Gambın

摘要:

在虚拟化计算平台中,能量消耗与计算加通信过程有关。然而,大多数文献中提出的能耗模型和节能解决方案只考虑活动的虚拟机(VM),因此,整体运行能耗通常只与计算过程有关。为了解决这个缺点,在本文中,我们考虑了计算加通信的能量模型,在多访问边缘计算(MEC)范式中,然后提出了一种基于流量工程和MEC位置服务的在线服务器管理算法与能量收集(EH)功能的组合,称为能量感知服务器的自动资源控制器(ARCES),用于自动缩放和重新配置计算加通信资源。主要目标是在硬的每任务延迟约束(即,服务质量(QoS))。ARCES联合执行(i)短期服务器需求和收获的太阳能预测,(ii)VM软缩放、工作负载和处理速率分配,以及最后,(iii)打开/关闭传输驱动器(即,快速可调谐激光器)与位置感知流量调度相结合。我们的数值计算结果表明,ARCES实现了平均69%的节能,和从31%-45%和从21%-25%的能源消耗范围内的每虚拟机的重新配置成本的不同值,相对于没有能源管理的情况下,应用。

英文摘要:

In virtualized computing platforms, energy consumption is related to the computing-plus-communication processes. However, most of the proposed energy consumption models and energy saving solutions found in literature consider only the active Virtual Machines (VMs), thus the overall operational energy expenditure is usually related to solely the computation process. To address this shortcoming, in this paper we consider a computing-plus-communication energy model, within the Multi-access Edge Computing (MEC) paradigm, and then put forward a combination of a traffic engineering- and MEC Location Service-based online server management algorithm with Energy Harvesting (EH) capabilities, called Automated Resource Controller for Energy-aware Server (ARCES), for autoscaling and reconfiguring the computing-plus-communication resources. The main goal is to minimize the overall energy consumption, under hard per-task delay constraints (i.e., Quality of Service (QoS)). ARCES jointly performs (i) a short-term server demand and harvested solar energy forecasting, (ii) VM soft-scaling, workload and processing rate allocation and lastly, (iii) switching on/off of transmission drivers (i.e., fast tunable lasers) coupled with the location-aware traffic scheduling. Our numerical results reveal that ARCES achieves on average energy savings of 69%, and an energy consumption ranging from 31%-45%and from 21%-25% at different values of per-VM reconfiguration cost, with respect to the case where no energy management is applied.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑