Date of Publication :29th March 2024

Abstract:Resource management in cloud-based big data systems presents significant challenges due to the dynamic nature of workloads, varying resource demands, and the need for efficient utilization of computational resources. This paper proposes a novel approach for resource management in cloud-based big data systems using a hybrid mechanism called Firefly Optimization Convolutional Neural Network (FFOCNN). The FFOCNN mechanism combines the strengths of FFO and CNN to optimize resource allocation, workload scheduling, and system performance. Firefly Optimization is employed to dynamically adjust resource allocations based on changing workload patterns and optimize the utilization of computational resources, storage, and network bandwidth. CNNs are utilized for workload prediction, anomaly detection, and resource demand forecasting, enabling the system to make informed decisions regarding resource allocation and optimization. The proposed FFOCNN mechanism is designed to address the challenges of scalability, adaptability, and efficiency in cloud-based big data systems by leveraging advanced optimization techniques and deep learning models. Experimental results demonstrate that the FFOCNN mechanism which is implemented in Python software outperforms traditional resource management approaches like CNN-LSTM (Long Short Term Memory), CNN-GRU (Gated Recurrent Unit), and Deep Neural Network with an accuracy of 99.21%. The integration of FFO and CNNs provides a robust framework for resource management in cloud-based big data systems, offering improved performance and scalability for handling large volumes of data and dynamic workloads. This research contributes to the advancement of resource management techniques in cloud computing environments, offering insights and methodologies for optimizing resource utilization and enhancing system efficiency in the era of big data.0

Reference :