Florence A Paulin, Shanthi V, Simon C B Sunil
Sathyabama University, Chennai 600 119, India; St. Joseph's Institute of Technology, Chennai 600 119, India.
St. Joseph's College of Engineering, Chennai 600 119, India.
ScientificWorldJournal. 2016;2016:9328070. doi: 10.1155/2016/9328070. Epub 2016 Apr 28.
Cloud computing is a new technology which supports resource sharing on a "Pay as you go" basis around the world. It provides various services such as SaaS, IaaS, and PaaS. Computation is a part of IaaS and the entire computational requests are to be served efficiently with optimal power utilization in the cloud. Recently, various algorithms are developed to reduce power consumption and even Dynamic Voltage and Frequency Scaling (DVFS) scheme is also used in this perspective. In this paper we have devised methodology which analyzes the behavior of the given cloud request and identifies the associated type of algorithm. Once the type of algorithm is identified, using their asymptotic notations, its time complexity is calculated. Using best fit strategy the appropriate host is identified and the incoming job is allocated to the victimized host. Using the measured time complexity the required clock frequency of the host is measured. According to that CPU frequency is scaled up or down using DVFS scheme, enabling energy to be saved up to 55% of total Watts consumption.
云计算是一项新技术，支持全球范围内基于“按需付费”的资源共享。它提供诸如软件即服务（SaaS）、基础设施即服务（IaaS）和平台即服务（PaaS）等各种服务。计算是IaaS的一部分，并且整个计算请求要在云中以最佳的功率利用率高效地得到服务。最近，人们开发了各种算法来降低功耗，甚至从这个角度也使用了动态电压和频率缩放（DVFS）方案。在本文中，我们设计了一种方法，该方法分析给定云请求的行为并识别相关的算法类型。一旦识别出算法类型，就使用它们的渐近符号计算其时间复杂度。使用最佳匹配策略识别合适的主机，并将传入的作业分配给受害主机。使用测量的时间复杂度测量主机所需的时钟频率。据此，使用DVFS方案向上或向下缩放CPU频率，可使能源节省高达总功耗的55%。
