期刊名称：International Journal of Computer Science & Information Technology (IJCSIT)
印刷版ISSN：0975-4660
电子版ISSN：0975-3826
出版年度：2014
卷号：6
期号：2
页码：1
出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：Reliability of a performance model is quintessential to the robust resource management of applications onthe cloud platform. Existing studies show that the contention for shared I/O induces temporal performancevariations in a guest VM and heterogeneity in the underlying hardware leads to relative performancedifference between guest VMs of the same abstract type. In this work, we demonstrate that a guest VMexhibits significant performance variations across repeated runs in spite of contention free hosting of asingle guest VM on a physical machine. Also, notable performance difference between guest VMs createdequal on physical machines of a homogeneous cluster is noticed. Systematic examination of the componentsinvolved in the request processing identifies disk I/O as the source of variations. Further investigationestablishes that the root cause of the variations is linked with how MySQL manages the storage of tablesand indexes on the guest VM's disk file system. The observed variations in performance raise the challengeof creating a consistent and repeatable profile. To this end, we present and evaluate a black box approachbased on database population from a snapshot to reduce the perceived performance variations. Theexperimental results show that the profile created for a database populated using a snapshot can be usedfor performance modeling up to 80% CPU utilization. We validate our findings on the Amazon EC2 cloudplatform
