On data logging in real-time process control systems

Real-time process control systems require hard real-time data logging tasks to access all of the data sampled from plants and monitoring tasks to analyze the status of the plants. This paper presents an integrated scheduling scheme for hard real-time data logging tasks and soft real-time monitoring tasks, both types of which require disk I/Os. Conventional disk scheduling policies like Shortest-Seek-Time-First (SSTF), SCAN, and C-SCAN, reorder disk I/O requests to minimize the disk head´s seek time, thus making it very difficult to guarantee the timely completion of the tasks requiring disk I/Os. Our scheduling scheme provides timeliness guarantees by serializing all disk I/O requests of these tasks based on their priorities. Because it doesn´t require any special-purpose real-time disk scheduling algorithm, the proposed scheme can be implemented on top of general-purpose operating systems like UNIX