LIFE DATA ANALYSIS OF SERVER VIRTUALIZED SYSTEM

Abstract

The use of reliability metrics and life data analysis has received considerable attention
recently in the software engineering literature. Life data analysis under the actual operational profile can,
however, be expensive, time consuming or even infeasible. In this paper, a systematic approach has been
adopted in order to reduce the experimentation time for estimating time to failure of a server virtualized
system. The study of time to failure (TTF) is very essential in server virtualized system, because it is the crux
of the cloud computing infrastructure. In order to meet service-level agreements (SLAs) like availability,
reliability and response time, prediction of reliability metrics like mean time to failure (MTTF), life
distribution etc are indispensable. The most important contributions of this paper are the reduction of
experimental time, and the life data analysis of the server virtualized systems which were not addressed so far.
Experimental results demonstrate that there is only four percentage deviation from the observed results from
the Normalized Root Mean Square Error and resulting in 96% accuracy of predicting MTTF.