Extending the Watchdog Pattern for multi-threaded windows based traffic management and control applications

Applications in the field of traffic management and control systems have an increasing demand on availability. Typically, hardware circuit watchdogs reset the whole system in case of failure targeting the goal, that after some acceptable time the system is up and the applications are running correctly again. Depending on the hardware and operating systems used, this can take a reasonable time. Software-based monitoring solutions have the ability to restart only faulty software components which shortens recovery duration. Normally, they implement standard-patterns like the watchdog pattern or safety executive pattern. As these patterns are primarily targeting the monitoring of components two approaches for monitoring threads and the activity of every component involved are presented in this paper.