Date of Graduation

2003

Document Type

Thesis

Degree Type

MS

Committee Chair

Hany Ammar

Abstract

The ability to control and quantify the software quality attributes is the ultimate goal for many researchers; non-quantitative measures of quality attributes are not good enough for sound judgment and are often ambiguous. In this work we choose to focus on software models specified in Unified Modeling Language (UML). The main objective of this work is aimed at studying timeliness nonfunctional requirements through timing and performance analyses. We first study a simulation-based analysis methodology and its application on a case study. We then introduce a hybrid methodology that incorporates both analytical and simulation models. We recognize the threats that an unacceptable performance may impose on the software and we introduce a methodology to evaluate performance based risk. The objective of performance risk assessment is to help identify risky scenarios that should undergo further in-depth performance analysis. The proposed methodologies are further illustrated on case studies in the GSM domain.

Share

COinS