The Research And Implementation Of Availability Detection Mechanism For Networked Embedded Systems

As CPS (cyber-physical systems) applications which combine embedded systemstechnology with networking technology are becoming more and more common. There are anumber of factors, such as the instability of the hardware components, the instability of theartificial software modules and the interference of external application environment, resulting inthe unstable embedded system. Therefore, the study of the availability detection mechanism toenhance the stability of networked embedded systems is necessary.In order to enhance the stability of networked embedded systems, embedded availabilitydetection mechanisms are researched and implemented. The mechanisms can be divided intoself-testing and remote detecting of embedded systems. They include the mechanisms usingICMP-based detection technology to detect network connection status; using watchdog-baseddetection technology to detect thread running state; and using anomaly performance detectiontechnology to detect their own states. In the implementation of mechanisms, the methods ofacquisition of detected data from kernel and multi-point threshold judgment are used toimprove the efficiency and accuracy. In addition, a remote detection module which is based onthe CIM/WBEM (Common Information Model/Web Based Enterprise Management)architecture is implemented, which relies on acquisition of kernel-based detected data, hasbetter scalability, and is independent of the database.The ICMP-based detection and watchdog-based detection approaches are tested andvalidated by imitated fault events. The approach of detecting anomaly performance that relieson acquisition of kernel-based data detection is validated by experiments, the determinationmethod of multi-point threshold values is quantitatively analyzed, and its accuracy is shown.Finally, the feasibility of remote detection scheme is validated through viewing memoryinformation stored in embedded systems from the CIM client.