Title: Integrated online and offline scheduling of real-time tasks using a co-processor scheduling unit towards dual-mode kernels

Authors: Yacine Laalaoui

Addresses: Department of Information Technology, Taif University, Taif, Kingdom of Saudi Arabia

Abstract: In this paper, we integrate online and offline scheduling of real-time tasks using two processing elements. The scheduling task executes on the first PE (called scheduling PE or the co-processor). User tasks execute on the second PE (called main PE). The main objective is to decrease the scheduling overhead from the main PE. The scheduling PE executes both online and offline algorithms. Thus, it runs on two different modes, scheduling and dispatching. The scheduling mode executes first at selection instants using an online algorithm. The offline algorithm executes in parallel to users tasks. Once the offline algorithm finds a feasible schedule, the scheduling PE switches to dispatching mode. We further describe a new task states that fit the proposed design. Finally, we explain the resolvability of the problem of nonpreemptive scheduling using the IBM ILOG-CP solver and Xu and Parnas's algorithm using the proposed design.

Keywords: real-time tasks; scheduler; dispatcher; optimal scheduling; online; offline; co-processor scheduling.

DOI: 10.1504/IJCSYSE.2022.131032

International Journal of Computational Systems Engineering, 2022 Vol.7 No.1, pp.19 - 29

Received: 04 Mar 2022
Received in revised form: 10 Aug 2022
Accepted: 10 Aug 2022
Published online: 19 May 2023 *