“Keene Development Process SW Reliability Model:
An Early Prediction Method”
Dr. Samuel Keene, IEEE Reliability Society
Presented in conjunction with the IEEE Reliability Society
How do you estimate the reliability of software before it is finally developed, inspected, tested and released? What assertions can we make about software reliability early on in the process? Suppose you were an auditor and were asked to assess the quality of software being developed. How would you make that assessment? What questions could you ask of the management and the developers? What other metrics would you look at? Historically, some development organizations looked at their past programs and scaled the new one to have a 10 percent lower fault rate. The new model presented here has shape factors that merge code size, historical performance and operational profile as well as the capability of the development organization.
This talk discusses an a priori software reliability prediction model that projects the latent fault rate of the code being released to the field along with its reliability growth profile after release. The latter results from the ongoing fault discovery and subsequent fault removal. The code is refined over time and usage. This model helps set the reliability expectations for the new software and has been successfully used by several major government contractors and development organizations. It is also referenced in the IEEE Standard for Software Reliability 1633.
Sam Keene is a past president of the IEEE Reliability Society and continues to serve on its advisory committee. He also serves on the board of directors of the IEEE Technology Management Council. He received the 1996 Reliability Engineer of the Year award, the Allan Chop Education award from the American Society of Quality in 1999, and the IEEE Education Award in 2000. He was also recognized with the IEEE Millennium Medal and by the FAA for distinguished service on the WAAS flight navigation system’s reliability. He is an IEEE fellow, and he has produced 10 video tutorials on aspects of software development, software reliability and concurrent engineering for NASA, NTU and the IEEE. He has published over 200 technical papers and several book chapters. He was the co-principal developer of the PRISM reliability model and has worked on updates of the MIL HDBK 217 models. He has been a Six Sigma Senior Master Black Belt since 1999. He received his PhD in operations research from the University of Colorado.
[an error occurred while processing this directive]