Software reliability from a system perspective

Abstract

The authors address two aspects of software reliability. First, some of the key tools are identified to improve and assure software reliability through process control, failure mode and effect analysis, and so on. These activities actively augment the reliability of the design and code. Second, the software and hardware metrics are combined to yield an estimate of system reliability. The very nature of measuring the code and feeding back the estimate to the designers will work to increase program focus and to improve the code. Checking out and validating interfaces are concluded to be critical to highly reliable designs. This is especially true across the hardware/software boundary.<>