Tenasys Corporation, Real-time Virtualization Experts

 

DSM Role in Reliability

Complex real-time tasks, especially those distributed over multiple hardware elements, require solutions beyond a hard real-time execution environment. Insuring integrity requires a system manager that can notify processes and threads of the availability or failure of key resources.

INtime remote nodes are the mechanism by which real-time Windows applications can be scaled from a single shared hardware module (Windows+INtime) to one distributed over multiple systems (Windows+INtime+INtime+…). The INtime distributed architecture can be implemented without the need to rewrite, or even recompile, your software! The INtime Distributed System Manager (DSM) is the key to providing this flexibility. More info more info

What is the DSM?

The Distributed System Manager (DSM) is a cooperative, multiple-process application for managing INtime systems. It tracks the state of the system, monitors the health of its components, and cleans up in the event of a tracked component’s termination or failure, including the Windows operating system and applications. More info more info

The INtime DSM performs these key tasks:

  • tracks dependencies between Windows processes and threads and real-time processes, threads, and objects
  • provides notification when tracked processes, threads, and objects terminate or fail
  • cleans up shared resources when a tracked Windows process or thread, or the entire Windows environment, terminates or fails

The DSM consists of two parts, a Windows DSM process and a real-time DSM process. These two processes (“the DSM”) cooperate to insure integrity of the entire system. The DSM continuously monitors processes, threads, and shared resources (or simply, “objects”) that have been explicitly declared with dependencies. If an object is no longer available the DSM notifies all interested parties of that event.

For example, when the DSM recognizes that a communication link is no longer present or a process has undergone a catastrophic error (e.g., a Windows blue screen is imminent) it notifies all interested threads of the event. Notification triggers a thread to perform any application-specific procedures necessary to handle the event.

The DSM can be used to build reliable distributed real-time systems that can be scaled from a single Windows/INtime hardware platform to those incorporating multiple INtime real-time modules optionally connected to a supervisory Windows host.