One means of obtaining time predicable systems is to build them from predictable components which are themselves built from predictable subcomponents. But this is not the only approach. In this paper we illustrate the benefits that are possible if components are designed to exhibit independent random behaviour. The worst-case execution time of application code can be estimated (with low probability of failure) to be not much greater than what one would expect the average to be. Unfortunately this approach of obtaining predictability as an emergent property of the code's execution cannot be delivered by today's hardware. Means by which hardware can be made more accommodating to this objective are explored.
|Title of host publication
|4th Workshop on Compositional Theory and Technology for Real‐Time Embedded Systems
|Robert I. Davis, Linh T.X. Phan
|Number of pages
|Published - 2011