Dynamic Resource Allocation for Latency Optimization and Energy Efficiency in Fog Computing: A Comparative Analysis Across Cloud, Edge, and Hybrid Architectures

For time-sensitive applications including autonomous cars, smart healthcare, and industrial IoT, fog computing's value for controlling latency, resource economy, and energy efficiency is becoming clear. This paper investigates three workload scenarios: Low (10–50 tasks), Medium (50–200 tasks), and High (200–1000 tasks), thereby contrasting Cloud Processing, Edge Prioritizing, and Dynamic Resource Allocation. The CityPulse IoT, Intel Lab Sensor, and Google Cloud Public Datasets are employed to ensure relevance and applicability. A predictive load-balancing algorithm is integrated into the evaluation of node availability, network latency, and energy constraints in real time within Dynamic Resource Allocation. Using measures established by sensitivity, regression, and ANOVA, advanced models evaluate latency, resource use, and energy costs. Dynamic resource allocation is shown to minimize latency, resource use, and energy over all workload conditions. By means of the combination of theoretical frameworks and pragmatic implementations, support is given for designers in constructing scalable, flexible, and efficient fog computing systems. Identified are research areas targeted at improving fog computing systems for next IoT ecosystems and real-time analytics. Global interoperability standards, hybrid fog-edge architecture, artificial intelligence-driven predictive optimization, and real-world validation are discussed here.