Optimized Resource Management in Cloud-IoT Integrated Environments using Machine Learning

In recent years, the integration of Cloud Computing and the Internet of Things (IoT) has revolutionized data processing, storage, and communication by providing scalable and efficient resource management. However, managing heterogeneous devices and dynamic workloads in Cloud-IoT environments remains a major challenge due to latency, energy inefficiency, and resource imbalance. Conventional methods such as Round-Robin Scheduling (RRS), First-Come-First-Serve (FCFS), and Dynamic Resource Allocation (DRA) have been widely used for workload distribution and resource assignment. Yet, these approaches exhibit limitations including high response time, inefficient load balancing, and poor energy optimization in large-scale IoT deployments. To overcome these challenges, the proposed Machine Learning-Based Resource Optimization Model (ML-ROM) dynamically predicts workload variations and allocates resources intelligently using adaptive learning mechanisms. The model integrates real-time feedback from IoT nodes to enhance decision accuracy and reduce resource wastage. Through simulationbased evaluation, the proposed ML-ROM achieves significant improvements compared to conventional methods—resource utilization increased by 28%, latency reduced by 22%, energy efficiency improved by 31%, and overall throughput enhanced by 25%. The optimized framework ensures sustainable performance, scalability, and reliability, making it highly suitable for smart city, healthcare, and industrial IoT applications.