Which models and architectures of distributed real-time computing systems suit which application area?

Abstract

A real time system must produce the intended results at the intended points of real time determined by its environment, otherwise the system has failed (H. Kopetz, 1997). For the purpose of this discussion, we consider two types of environments that contain real time computer systems: (i) static environments, where the frequency of the service requests on the computer system are constrained by a priori known environmental characteristics; (ii) dynamic environments, where it is possible to specify peak-load-service-request rates probabilistically only, e.g., a military command and control system that has to adapt dynamically to the n priori unknown threats of the enemy. We feel that these two fundamentally different environments of distributed real time computer systems demand different system models and different implementation strategies.