Data Network Performance Metrics Optimization in Heterogeneous Environments

Abstract

Although data network performance optimization has received sustained scholarly and industrial attention, integrated assessments of protocol inefficiencies, wide area network (WAN) acceleration, and measurement distortion across heterogeneous private and wireless environments remain limited. This study evaluates throughput, delay, jitter, packet loss, and application response time within a faculty network and a General Packet Radio Service (GPRS)-based wide area wireless network. A quantitative experimental design was employed, combining virtual web hosting, repeated download tests, protocol-level measurements, and simulations of packet-loss sensitivity. Data obtained from representative web workloads, network optimization appliances, and monitoring outputs were analyzed using descriptive performance metrics, Transmission Control Protocol (TCP) throughput models, and comparative latency analysis. The findings demonstrate that web browsing over high-latency connections substantially underutilizes available bandwidth because of sequential object retrieval, connection-establishment delays, and sensitivity to packet loss. Increasing the number of concurrent connections and implementing Hypertext Transfer Protocol (HTTP) pipelining, compression, and proxy-based WAN optimization substantially reduced page latency. Segmented monitoring also restored performance visibility across optimized WAN paths. These findings establish an integrated framework for diagnosing and improving network performance in heterogeneous institutional environments by linking application behavior, transport-protocol limitations, optimization mechanisms, and monitoring accuracy. The study concludes that effective network upgrades should integrate infrastructure design, protocol optimization, monitoring transparency, and context-sensitive reliability planning rather than rely solely on increased bandwidth capacity.