R2A-DSA: A Robust Resource Allocation Approach for Delay-Stringent Applications in IoT-Fog-Cloud Networks

Now-a-days, the exponential growth of Internet of Things (IoT) applications has intensified the demand for efficient resource allocation strategies to meet critical Quality of Service (QoS) requirements, especially in terms of latency and execution time. While cloud computing offers vast computational resources, centralized processing introduces significant delays, making it unsuitable for time-sensitive tasks. Fog computing mitigates this by enabling edge-level processing, but its limited resources necessitate a hybrid IoT-fog-cloud architecture. This paper proposes Robust Resource Allocation-Delay Stringent Application (R2A-DSA), a novel Robust Resource Allocation approach that integrates task classification, prioritized queuing, and Secretary Bird Optimization Algorithm (SBOA) for dynamically optimizing resource distribution across fog and cloud layers. Our framework categorizes tasks based on their characteristics and priorities, then dynamically allocates resources across fog and cloud layers to decrease latency and increase whole system efficiency. Inspired by the hunting techniques of secretary bird, the SBOA is used to balance exploration and exploitation during resource allocation, achieving optimal configurations even with fluctuating workloads and resource availability. Extensive experiments on the iFogSim simulator demonstrate that R2A-DSA reduces execution time by 32%-45%, cuts task delays by 28%-40% and lowers task failure rates by 50%-65% compared to traditional methods (First Come First Serve (FCFS), Random, Cloud Only) and metaheuristic benchmarks (Harris Hawks Optimization (HHO), Crayfish Optimization Algorithm (COA), Fennec Fox Optimization Algorithm (FFA)). These results highlight R2A-DSA's superiority in enhancing system efficiency, scalability, and reliability for latency-sensitive IoT applications, offering a significant advancement in distributed computing resource management.