Container Management Strategy with Dynamic Scalable Task Scheduling Load Balancing and Resource Allocation in Cloud Environment Using Hybrid Optimal and Deep Learning Techniques

Abstract

Cloud-based micro services have gained remarkable traction across industries due to their exceptional performance. Cloud containers have emerged as a lightweight and reliable virtualization solution for delivering cloud services, offering advantages such as scalability, portability, and flexible deployment. Essential to cloud container services are planner components, which optimize resource management, workload performance, and cost efficiency. Unlike conventional methods that focus on virtual machine allocation and migration, the crux of the challenge lies in efficiently allocating resources for containers, directly influencing system performance and resource utilization. This study introduces a dual-pronged methodology. Firstly, it presents a hybrid approach amalgamating optimal and deep learning techniques for dynamic scalable task scheduling (DSTS). The approach encompasses a novel modified multi-swarm coyote optimization (MMCO) method to expand container virtual resources, a tailored Modified pigeon-inspired optimization (MPIO) technique for task clustering based on priority, a rapid adaptive feedback recurrent neural network (FARNN) for pre-virtual CPU allocation, and a deep convolutional neural network (DCNN) for real-time task load monitoring. Secondly, the study addresses the concurrent optimization of load balancing and resource allocation within the cloud through an optimal container management strategy. This strategy employs an improved backtracking search optimization (IBSO) algorithm to efficiently allocate resources between end users/IoT devices and the cloud, all while accounting for service-level agreements. Additionally, a mechanistic quantum recurrent neural network (MQ-RNN) is innovatively designed for container allocation, consolidation, and migration within cloud environments. The efficacy of the proposed methodologies is validated through extensive results and comprehensive comparative analysis.