A generic methodology to derive domain-specific performance feedback for developers

Abstract

The performance of a system directly influences business critical metrics like total cost of ownership (TCO) and user satisfaction. However, building responsive, resource efficient and scalable applications is a challenging task. Thus, software engineering approaches are required to support software architects and developers in meeting these challenges. In this PhD research abstract, we propose a novel performance evaluation process applied during the software development phase. The goal is to increase the performance awareness of developers by providing feedback with respect to performance properties that is integrated in the every day development process. The feedback is based on domain-specific prediction functions derived by a generic methodology that executes a series of systematic measurements. We apply and validate the approach in different development scenarios at SAP.