|
Revealing Relationships Between SW Faults & Failures & Their Effect on SW Reliability Assessment
|
|
Point of Contact
|
Katerina Goseva-Popstajanova
Katerina.Goseva@mail.wvu.edu
|
|
Dates
|
January 2007 - December 2009
|
|
Problem
|
Our previous research work has shown that some simplifying assumptions made while analyzing software faults and failures and assessing software reliability are not fully justified and may lead to significant errors. It has also shown that using large, real systems as case studies is extremely important for advancing the software assurance and IV&V areas since some phenomena can only be observed on large scale empirical studies. This project is focused on two main topics - exploring the relationships between faults and failures based on large empirical studies which will include open source software and NASA applications and development of architecture-based software reliability models based on more realistic assumptions that will account for the observed phenomena.
|
|
Objective
|
The main focus of this research is to explore the unknown relationships between software faults and failures based on large scale empirical case studies, including open source software and NASA projects and incorporate the newly discovered phenomena into a set of more realistic assumptions used to develop architecture-based software reliability models. The objectives and the dates by which they are planned to be accomplished are as follows.
- Explore the problem/change tracking systems of different projects (open source projects, International Space Station, and projects from Metrics Data Program). Automate to the highest possible extent the process of data extraction. Based on the preliminary analysis of the raw data, propose an appropriate classification system as a vehicle that will allow more efficient study of the fault/failure relationships. (December 31, 2007)
- Classify the fault/failure data for multiple projects and refine the classification system if needed. Conduct analysis aimed at revealing the relationships between software faults and failures and identifying common patterns and unusual dependencies. Compile a list of best practices and lessons learned. (December 31, 2008)
- Based on the discovered phenomena build a set of realistic assumptions and develop a new architecture-based software reliability model (analytical or simulation based) capable of accounting for these phenomena. Explore the model accuracy on multiple case studies. (December 31, 2009)
|
|
Results
|
SAS_07_Exec_Brief_Reveal_Relationships_Fault_Failure_Goseva-Popstojanova.ppt
SAS_07_Tech_Pres_Fault_Failure_Goseva-Popstojanova.ppt
|
|
Keywords
|
Fault, failure, empirical studies, inter-component failure dependence, architecture-based software reliability.
|
|
Categories
|
Software Reliability
|
|
