Design, Implementation, and Environmental Effects of a Load-Balanced Smart Waste Management System Powered by Fog Computing in Iraq

Abstract

This paper designs and evaluates a load-balanced fog-computing architecture for smart waste management in Iraq. The study addresses the gap in context-specific evidence on how fog-layer task allocation affects latency-sensitive waste-monitoring services under constrained communication conditions. A three-tier IoT-fog-cloud model was implemented in iFogSim2 with 25 fog nodes distributed across five governorate clusters and 1,250-2,500 simulated IoT devices. Four strategies-Round Robin, Weighted Round Robin, Throttled, and Honey Bee-inspired allocation-were compared over 30 independent replications using latency, energy consumption, throughput, recovery time, and data loss. Throttled allocation achieved the strongest overall profile, reducing average latency by 34% compared with Round Robin under normal conditions (p < 0.001; Cohen's d = 2.14) and lowering energy use by approximately 18.5%. The findings indicate that dynamic, load-sensitive allocation is more suitable than static allocation for smart waste monitoring in Iraq, provided that minimum connectivity, device security, operator training, and institutional data governance are addressed.