Smart real time operating system

Abstract

The work reported in this thesis proposes a practical approach to enhance real-time property of Standard Linux and build a smart real-time operating system. In such a system, a standard time-sharing operating system and a real-time execution module cooperate to satisfy the timing requirements of the applications. The design, implementation and system performance of this system are explored. Besides this, this system enhances the timing response accuracy using multiple approaches. The proposed Queue Manager (QM) mechanism helps to reduce the scheduling overhead. A flexible scheduling framework helps to develop the customized real-time scheduling disciplines. Moreover, a hybrid process model is presented to model the realistic real-time transactions and the schedulability analysis is conducted in this model. A flexible sporadic server algorithm is presented to schedule aperiodic tasks in hybrid process model. The response accuracy of FSS algorithm is compared against sporadic server algorithm. The experimental results show FSS algorithm helps to achieve the better response accuracy.