Analysis and differentiation of software system environments

Abstract

To be certain that prerelease reliability and performance levels can be expected in operation, one must establish that the testing environment emulates the end-use environment. If these environments differ, then one must know the nature of this difference to know its impact on customer satisfaction. While an ability to differentiate software environments is important in software engineering, there are few publications regarding this concept. Of the publications that we know of, none provide an empirical setting for comparing environments. In this paper we analyse two software systems as they operate in selected environments. We present two graphical characterizations of these systems, and we apply a statistical method to determine whether two environments are distinct. The results suggest applications of statistical methods in the detection of environmental differences.