Performance analysis of distributed real-time databases

In a distributed process control system, information about the behavior of physical processes is usually collected and stored in a real-time database which can be remotely accessed by human operators. In this paper we propose an analytic approach to compute the response-time distribution of operator consoles in a distributed process control environment. The technique we develop is based on Markov regenerative processes (MRGPs) and described with the assistance of deterministic and stochastic Petri nets (DSPNs). We construct exact models for performance analysis of centralized and decentralized database architectures. However, due to limitations on the exact solution, we also propose an approximate solution which is then used to study response-time distributions of large systems.