SLA-Aware and Energy-Efficient Dynamic Overbooking in SDN-Based Cloud Data Centers
SLA-Aware and Energy-Efficient Dynamic Overbooking in SDN-Based Cloud Data Centers
复制标题
DOI：
10.1109/tsusc.2017.2702164
复制
发表时间：
2017-04
影响因子：
3.9
通讯作者：
Jungmin Son;A. V. Dastjerdi;R. Calheiros;R. Buyya
中科院分区：
文献类型：
--
作者：
Jungmin Son;A. V. Dastjerdi;R. Calheiros;R. Buyya
由于云数据中心能耗高，运营成本高，因此其电源管理受到了业界和学术界的高度关注。虽然主机在消耗电力方面占主导地位，但网络占数据中心总能源成本的10%到20%。资源超预订是通过向相同数量的资源放置更多请求来减少活动主机和网络使用的一种方法。软件定义网络（SDN）可以整合流量并动态控制服务质量（QoS），从而促进网络资源的超额预订。然而，现有的方法采用固定的超额预订比例来决定要分配的资源量，这在现实中可能会导致过度违反服务水平协议（SLA），并且工作负载不可预测。在本文中，我们提出了动态超额预订策略，该策略联合利用虚拟化能力和SDN进行虚拟机和流量整合。随着工作负载的动态变化，该策略可以更精确地为虚拟机和流量分配资源。此策略可以增加主机和网络中的超额预订，同时仍然提供足够的资源以最小化SLA违规。我们的方法基于来自主机和网络利用率在线分析的历史监控数据来计算资源分配比例，而无需预先了解工作负载。我们在大规模的模拟环境中实现了它，以证明它在维基百科工作负载上下文中的有效性。我们的方法节省了数据中心的能源消耗，同时减少了SLA违规。
Power management of cloud data centers has received great attention from industry and academia as they are expensive to operate due to their high energy consumption. While hosts are dominant to consume electric power, networks account for 10 to 20 percent of the total energy costs in a data center. Resource overbooking is one way to reduce the usage of active hosts and networks by placing more requests to the same amount of resources. Network resource overbooking can be facilitated by Software Defined Networking (SDN) that can consolidate traffics and control Quality of Service (QoS) dynamically. However, the existing approaches employ fixed overbooking ratio to decide the amount of resources to be allocated, which in reality may cause excessive Service Level Agreements (SLA) violation with workloads being unpredictable. In this paper, we propose dynamic overbooking strategy which jointly leverages virtualization capabilities and SDN for VM and traffic consolidation. With the dynamically changing workload, the proposed strategy allocates more precise amount of resources to VMs and traffics. This strategy can increase overbooking in a host and network while still providing enough resources to minimize SLA violations. Our approach calculates resource allocation ratio based on the historical monitoring data from the online analysis of the host and network utilization without any pre-knowledge of workloads. We implemented it in simulation environment in large scale to demonstrate the effectiveness in the context of Wikipedia workloads. Our approach saves energy consumption in the data center while reducing SLA violations.