A Dimensionality Model Approach to Testing and Improving Software Robustness

Jiantao Pan, Philip Koopman & Daniel Siewiorek

ECE Department & Institute for Complex Engineered Systems
Carnegie Mellon University, Pittsburgh, Pennsylvania, USA

Published in Proceedings of Autotestcon 99, San Antonio, Texas, August 30 - September 2, 1999.


Software robustness problems may hinder the use of Commercial Off-The-Shelf (COTS) software modules and legacy software modules in mission-critical and safety-critical applications. This research focuses on hardening COTS and legacy software modules against robustness failures triggered by exceptional inputs. An automated approach is presented that is capable of identifying the triggers of the robustness failures. A fault model - the Dimensionality Model - is used to guide analysis. An experiment is described which demonstrates the feasibility of automating the process of analyzing failure causes and hardening against certain data types in POSIX function calls, for example, NULL pointer values and scalar data types such as INT and FLOAT. The final goal of this research is to provide users a tool to harden COTS and legacy software modules automatically.


Slides from conference presentation:

BALLIST HOME PAGE Ballista Home Page