Author(s): Mariana L Soares

Abstract: Packet delay and throughput are critical performance metrics in the design and evaluation of computer networks, particularly in small-scale environments where resource constraints and traffic dynamics strongly influence quality of service. This research presents a simulation-based analysis of packet delay and throughput in a small-scale computer network using a controlled network model that captures realistic traffic patterns, queueing behavior, and protocol interactions. The simulation framework enables systematic variation of key parameters such as packet arrival rate, link bandwidth, buffer size, and routing configuration to observe their impact on end-to-end delay and achievable throughput. Performance results demonstrate that increasing traffic load leads to nonlinear growth in packet delay due to queue saturation, while throughput initially increases with load before reaching a stable limit governed by link capacity. The research further shows that appropriate buffer sizing can reduce packet loss but may increase average delay, highlighting an inherent trade-off between latency and throughput. Comparative observations across multiple simulation scenarios indicate that balanced resource allocation and moderate traffic intensity yield optimal performance in small networks. The findings emphasize the usefulness of simulation tools for understanding complex network behavior without the cost and rigidity of physical test beds. By providing quantitative insights into delay-throughput relationships, this work supports informed network design decisions for educational laboratories, small offices, and experimental test environments. The results also offer a foundation for extending simulation-based evaluation toward more advanced networks incorporating quality-of-service mechanisms, adaptive routing, and heterogeneous traffic types. Such controlled experimentation assists researchers and students in visualizing performance trends, validating analytical expectations, and developing intuition about congestion effects, protocol efficiency, and scalability limits, thereby strengthening foundational understanding of network performance analysis and supporting reproducible, low-cost experimentation. These insights are particularly valuable for preliminary design stages and academic demonstrations involving simplified yet representative network configurations under varied simulated operating conditions.

DOI: 10.33545/27075923.2026.v7.i1a.126

Pages: 47-52 | Views: 95 | Downloads: 51

Download Full Article: Click Here