A Distributed Architecture for Large Scale Multimedia Visualization and Surveillance

Distributed multimedia-based surveillance has been widely used for monitoring and control systems as well as many other industry applications. These applications enable various multimedia sources to be transported, processed, and visualized. Of many requirements of these systems, real-time and high reliability are often required and proved to be challenging. In this paper, we presented an IP-based architecture for implementing large-scale distributed video surveillance systems with real-time, synchronized visualization, and high reliability requirements. The main contribution of our work is the dynamical flow control mechanism integrated with fault-tolerant system architecture to improving real-time performance and system reliability on top of an inherently unreliable IP network. Experiments with large systems are also discussed.