Recent Advances in A Proactive Auto-scaling and Energy-Efficient VM Allocation Framework Using an Online Multi-Resource Capsule Shuffle Attention Network for Cloud Data Centres: A Systematic Review

Abstract

Cloud data centres are essential for supporting modern digital services, including cloud computing, big data analytics, artificial intelligence, and large-scale web applications. These centres host numerous servers and virtual machines (VMs) that deliver on-demand computing resources. However, the rapid expansion of cloud services has significantly increased computational workloads and energy consumption, leading to higher operational costs and environmental concerns such as carbon emissions. As a result, efficient resource management, proactive auto-scaling, and energy-aware VM allocation have become critical research areas. Auto-scaling enables dynamic adjustment of resources based on workload demand, ensuring optimal performance while minimizing wastage. Traditional reactive approaches respond only after performance degradation, often causing delays and SLA violations. In contrast, proactive auto-scaling uses predictive models to anticipate workload changes, allowing timely and efficient resource provisioning. Additionally, energy consumption in virtualized environments remains a major challenge. Inefficient VM placement can increase power usage, whereas energy-efficient allocation strategies aim to consolidate workloads and optimize server utilization. These approaches reduce energy consumption while maintaining system performance and reliability, making cloud infrastructures more sustainable and cost-effective.