An experiment-driven energy consumption model for virtual machine management systems

Abstract

As energy consumption is becoming critical in Cloud data centers, Cloud providers are adopting energy-efficient virtual machines management systems. These systems essentially rely on "what-if" analysis to determine what the consequence of their actions would be and to choose the best one according to a number of metrics. However, modeling energy consumption of simple operations such as starting a new VM or live-migrating is complicated by the fact that multiple phenomena occur. It is therefore important to identify which factors influence energy consumption before proposing any new model. We claim in this paper that one critical parameter is the host configuration, characterized by the number of VMs it is currently executing. Based on this observation, we present an energy model that provides energy estimation associated with VM management operations, such as VMs placement, VM start up and VM migration. The average relative estimation error is lower than 10% using the transactional web benchmark TPC-W, making it a good candidate for driving the actions of future energy-aware cloud management systems.

An Energy and Performance Aware Consolidation Technique for Containerized Datacenters

“…The energy cost will depend on the host's power profile P while the performance cost will be subject to several factors such as available bandwidth. We assume P a linear function of host's utilisation level -the more it is utilised, the more energy it will consume [27], [28]. The relationship between power consumption and CPU utilisation can be described as given in Eq.…”

“…The part (P max − P idle ) × u is known as dynamic power consumption, and is treated as a linear function of u. This simplified model estimates the power consumption of a non-virtualised host with less than 5% error, however, it needs modification to account for virtualisation and containerization [28]. Moreover, we assume ∼10% performance degradation due to migration [11].…”

Comparative experimental analysis of the quality-of-service and energy-efficiency of VMs and containers' consolidation for cloud applications

“…5.5.54; Apache 2.0; and PHP Version 5.4.45. Each service runs the web service benchmark RUBiS 4 . RUBiS benchmark is accepted in literature as a good example of a typical Cloud application.…”

“…However, placing multiple virtualized services on the same server has been demonstrated to lead to a QoS degradation. While this limit has been studied for other virtualization techniques [4], with the appearance of containerization technologies it becomes necessary to evaluate this behavior in consolidated environments using containers, and how this consolidation affects QoS and EE.…”

“…Also, no work in literature offers an empirical limit of containers running on the same physical machine, as it has been already addressed for VMs [4]. We believe that an empirical investigation on the performance of containers is useful to better understand their potential.…”