Author(s): Lukas M Schneider, Anna K Vogel and Tobias R Klein

Abstract: Latency is a critical performance indicator in computer networks, directly influencing application responsiveness, reliability, and user experience in constrained environments. Small-scale networks, such as laboratory setups, campus subnets, and small enterprises, increasingly rely on simulation-based evaluation to understand latency behavior without disrupting production traffic. This research investigates latency patterns in small-scale computer networks using simulated traffic models that represent diverse load conditions, packet sizes, and routing behaviors. A controlled simulation environment was constructed to emulate typical network topologies, including star and mesh configurations, while varying bandwidth, queue management, and propagation delay parameters. Traffic generators produced constant bit rate, bursty, and mixed workloads to capture realistic operational scenarios. Latency metrics were recorded at end hosts and intermediate nodes, enabling the analysis of average delay, jitter, and tail latency under incremental load. Results indicate that even modest increases in offered load can produce nonlinear latency growth due to queue buildup and contention, particularly in links with limited buffer capacity. Bursty traffic was found to exacerbate delay variability, while appropriate queue discipline reduced extreme latency spikes. Comparative observations across topologies show that path diversity mitigates congestion-induced delays when routing is stable. The findings highlight the sensitivity of small networks to configuration choices that are often overlooked in practice. By demonstrating how simulated traffic can reveal latent performance bottlenecks, this research provides a methodological foundation for proactive network planning, testing, and optimization. The insights are intended to support educators, researchers, and network administrators in designing resilient small-scale networks with predictable latency behavior before real-world deployment. Such pre deployment analysis reduces risk, improves service quality, and enables evidence-based tuning of protocols, buffers, and traffic policies across evolving workloads and technologies. It also facilitates repeatable experimentation, comparison of scenarios, and transparent reporting of assumptions for reproducible network performance studies under constrained budgets and timelines typical in practice.

DOI: 10.33545/27076636.2026.v7.i1a.152

Pages: 36-40 | Views: 106 | Downloads: 56

Download Full Article: Click Here