Asynchronous Remote Event Notification Using a Distributed Object Model for Heterogeneous Remote Monitoring Systems and Control Systems at the 1.8-GeV Tohoku Synchrotron Radiation Source

Abstract

An asynchronous remote event notification (AREN) for heterogeneous remote monitoring systems and distributed control systems has been developed using a distributed object model, Java remote method invocation (RMI), for the 1.8-GeV synchrotron radiation source at the Tohoku Synchrotron Radiation Facility (TSRF), which will have fifty beamlines for soft X-ray researches. The beamlines are controlled by distributed control systems composed of fifty outlying node computers and eleven remote monitoring systems connected to a network. There are more than three thousand event status signal data available from the beamlines. The event status signal data are transition status signals generated at any instance by valves/shutters, driving units, and safety interlocks when closing/opening. To acquire such event signal data, the monitoring system had to make continuous synchronous inquiries (CSI) or performed polling for an outlying node over the network. However, the monitoring system must synchronously wait each time until the outlying node transmits event status data back to the monitoring system, causing an unacceptable low event rate of approximately 40 events/second and heavy CPU load. On detecting a transitional change in an event status signal of a component, AREN automatically sends event status signal data to the monitoring systems over the network. Performance measurements proved that AREN increased the maximum event rate by one order of magnitude compared with that of CSI. By using AREN, the monitoring systems are capable of predicting signs that suggest malfunctioning of a beamline component at TSRF. The design and performance of AREN for monitoring systems and control systems for the synchrotron radiation source are described in this paper