Author(s): Lucas Reinhardt, Elena Rossi and Johan Svensson

Abstract: Lightweight operating systems have become foundational components of contemporary embedded, cyber-physical, and edge computing platforms where constrained resources and deterministic behavior are primary design requirements. Process scheduling in such systems directly influences latency, throughput, energy efficiency, and real-time responsiveness, yet empirical evaluations remain fragmented across platforms and workloads. This research presents an empirical analysis of process scheduling behavior in representative lightweight operating systems, focusing on how scheduler design choices affect execution fairness, response time, and context-switch overhead under realistic operating conditions. Controlled experiments were conducted using synthetic and application-oriented workloads to capture scheduler performance across varying task arrival rates, priority distributions, and computational intensities. Quantitative metrics including average waiting time, turnaround time, response time variance, CPU utilization, and pre-emption frequency were systematically measured and compared. The analysis reveals that while priority-based pre-emptive schedulers offer superior responsiveness for time-critical tasks, they may induce starvation risks under sustained mixed workloads. Conversely, round-robin and time-slice-based approaches demonstrate improved fairness and predictability but incur higher context-switch overhead in high-frequency task environments. The results further indicate that scheduler tuning parameters, such as quantum length and priority aging mechanisms, significantly moderate performance trade-offs. By correlating observed scheduling behavior with workload characteristics, this research highlights practical design implications for selecting and configuring schedulers in lightweight operating systems. The findings contribute empirical evidence that supports informed scheduler selection for embedded and real-time applications, emphasizing that no single scheduling strategy is universally optimal. Instead, performance efficiency emerges from aligning scheduler policies with workload demands, timing constraints, and resource limitations inherent to lightweight operating environments. These insights provide guidance for system designers, researchers, and practitioners seeking to balance responsiveness, fairness, and efficiency when deploying lightweight operating systems across diverse embedded scenarios and evolving edge workloads under practical constraints, real-time demands, and long-term maintainability considerations within constrained hardware ecosystems globally applicable.

DOI: 10.33545/27076636.2026.v7.i1a.151

Pages: 31-35 | Views: 101 | Downloads: 56

Download Full Article: Click Here