Home > Research > Publications & Outputs > EPIC

Links

Text available via DOI:

View graph of relations

EPIC: Profiling the propagation and effect of data errors in software

Research output: Contribution to journalJournal articlepeer-review

Published
Close
<mark>Journal publication date</mark>1/05/2004
<mark>Journal</mark>IEEE Transactions on Computers
Issue number5
Volume53
Number of pages19
Pages (from-to)512-530
Publication StatusPublished
<mark>Original language</mark>English

Abstract

We present an approach for analyzing the propagation and effect of data errors in modular software enabling the profiling of the vulnerabilities of software to find 1) the modules and signals most likely exposed to propagating errors and 2) the modules and signals which, when subjected to error, tend to cause more damage than others from a systems operation point-of-view. We discuss how to use the obtained profiles to identify where dependability structures and mechanims will likely be the most effective, i.e., how to perform a cost-benefit analysis for dependability. A fault-injection-based method for estimation of the various measures is described and the software of a real embedded control system is profiled to show the type of results obtainable by the analysis framework.