A low level analysis of the realtime Mach distributed operating system

Abstract

The use of measuring techniques to assist in the process of evaluation has always been a speculative endeavor. There is the presence of doubt as to what should be measured, and once this measurable entity has been determined, then what constitutes proper measurement of the entity itself. Current work in the real-time operating system software development arena presents us with this same enigma (how to, and what should be measured). The domain we have been pursuing to this end is a segmented appraisal consisting of analyzing specific elements which are properties of real-time operating systems. Six candidate elements of primary importance in the evaluation of real-time operating systems have been suggested and interpreted by the Rhealstone benchmark (Kar, 1989, 1990). These elements include: task switch time, preemption time, interrupt latency time, semaphore shuffling time, deadlock breaking time and interprocess communication latency time. We have researched these elements in relation to the Real-Time Mach Operating System and present the results of our findings. To minimize the variation on the end results, this appraisal was conducted under a specifically conditioned environment and we acknowledge the limitations of the results as such.