Efficient Cloud Resource Management using Complex-Value Spatio-Temporal Graph Convolutional Neural Network

Efficient resource management in cloud computing (CC) is critical to maintaining high availability, scalability and energy efficiency. Traditional methods often struggle to handle dynamic workloads, resulting in resource waste or overload. To address these challenges, this research proposes an Efficient Cloud Resource Management utilizing Complex-Valued Spatio-Temporal Graph Convolutional Neural Networks (CRM-CVSTGCNN). Initially, the input data is collected from the 2019 Google cluster workload traces dataset. It is then pre-processed using Cauchy Robust Correction-Sage Husa Extended Kalman Filter (CRCHEKF) to combine and normalize data vectors. The pre-processed data is then fed into Complex-Valued Spatio-Temporal Graph Convolutional Neural Networks (CVSTGCNN) to forecast future cloud workloads. Since CVSTGCNN alone does not adaptively optimize its weight parameters, the Portia Spider Optimization Algorithm (PSOA) is employed in this work to enhance prediction accuracy. The proposed approach is implemented in Python and evaluated utilizing metrics like Accuracy, Root Mean Squared Error (RMSE), Mean Squared Error (MSE), Mean Absolute Error (MAE) and Coefficient of Determination ([Formula: see text]). The performance measures of the proposed approach attain 97.5% accuracy, 0.03% RMSE and 0.022% MSE when compared to the existing methods, such as Stable and efficient resource management by deep neural network on cloud computing (SERM-CC-BiLSTM), A proactive auto scaling and energy-efficient VM allocation framework using online multi-resource neural network for cloud data center (EETS-CCE-GGCN), and AI-based energy-efficient task scheduling for CC surroundings (YPR-RWE-CNN) methods, respectively. These outcomes highlight the effectiveness of the proposed method for dynamic and energy-efficient cloud resource management.