cs 16(8): e1

Research Article

A Fair Cloud Resource Allocation using Data Envelopment Analysis

Download246 downloads
  • @ARTICLE{10.4108/icst.qshine.2014.256193,
        author={Hsin-Hung Cho and Chi-Yuan Chen and Hao-Wen Li and Timothy Shih and Han-Chieh Chao},
        title={A Fair Cloud Resource Allocation using Data Envelopment Analysis},
        journal={EAI Endorsed Transactions on Cloud Systems},
        volume={2},
        number={8},
        publisher={IEEE},
        journal_a={CS},
        year={2014},
        month={9},
        keywords={cloud computing; resource allocation; data envelopment analysis},
        doi={10.4108/icst.qshine.2014.256193}
    }
    
  • Hsin-Hung Cho
    Chi-Yuan Chen
    Hao-Wen Li
    Timothy Shih
    Han-Chieh Chao
    Year: 2014
    A Fair Cloud Resource Allocation using Data Envelopment Analysis
    CS
    EAI
    DOI: 10.4108/icst.qshine.2014.256193
Hsin-Hung Cho1, Chi-Yuan Chen2, Hao-Wen Li2, Timothy Shih1, Han-Chieh Chao2,*
  • 1: National Central University
  • 2: National Ilan University
*Contact email: hcc@niu.edu.tw

Abstract

Internet technology is advancing with each passing day, the user's demand is also increasing. Of course the users will more concern to quality of service. The vendors must find out a win win method of resource allocation to meet users and itself. Therefore, the resource allocation of cloud computing has become one of hottest topics. In literatures, some researchers have proposed resource allocation methods which include allocation of virtual machines and service classification, etc. However, these methods are based on subjective observations that lead to the overall cloud architecture becomes imbalance. In order to prevent such situation happened, we use the Data Envelopment Analysis (DEA) to solve the imbalance problem. In this paper, our analysis is start from the user's requests, and use the DEA to evaluate the whole cloud parameters. Then we can find out the resource allocation policy which is the most suitable between the users and vendors.