Adaptive Resource Prediction in Containerized Cloud Environments with CNN-LSTM Optimized by SCA

Many different businesses and professions employ cloud-based microservices because of their exceptional performance capabilities. Containers have evolved into the most trustworthy and lightest virtualization technology for delivering cloud services owing to its flexible sorting, mobility, and scalability. But a quick change causes issues with container management and automation. Numerous research projects have thus far concentrated on resolving the remaining problems with container automation and management. Since resource allocation in containers immediately impacts system and resource consumption, it is actually the cloud providers' top security concern. In order to address resource allocation in containers, this research optimizes the make span and energy using a hybrid optimization technique. Based on metaheuristic optimization and deep learning, specifically Convolutional Neural Network with Long Short-Term Memory networks, this paper presents a hybrid technique that is applied to improve the accuracy of load forecasting models on cloud settings. SCA metaheuristics is used to optimize the hyperparameters of the improved CNN-LSTM model thus capitalizing upon the predictive capability of CNN-LSTM. The general objective of this approach is to improve the accuracy of load prediction, believed to be necessary for proactive load balancing and migration in cloud infrastructures.