Computer systems are becoming more complex and more difficult to test but, simultaneously, competitive pressure is mounting to streamline system level test without affecting customer quality. Despite efforts to improve the effectiveness of lower level test processes, system-level testing still cannot be eliminated as a manufacturing process stage and must, therefore, be recognized as a valid discipline. This is happening but it is a relatively immature field with technical and financial drivers hotly debated. Economic modeling can aid objective analysis and has already been used to identify cost effective alternatives for component and board test.

This article briefly describes a system-level test economic model and interprets some key results. As a prerequisite, the features unique to system-level test are identified and an event rate model is used to characterize system behavior during both production testing and early field run-time.