Utilizing multi-population ant colony system and exponential grey prediction model for multi-objective virtual machine consolidation in Cloud Data Centers

Cloud Data Centers (CDCs) play a crucial role in meeting the increasing demand for scalable computing resources, facilitated by virtualization technology. However, their rapid expansion has led to heightened power consumption and carbon dioxide emissions, raising environmental concerns. Furthermore, inefficient resource usage within these CDCs contributes to increased energy consumption. Addressing these inefficiencies, especially in Virtual Machine (VM) consolidation, is imperative. Effectively balancing resource utilization, power consumption, and avoiding Service Level Agreement (SLA) violations poses a challenging paradox. This paper proposes EXGM-MACS, a strategy integrating the Exponential Grey Model (EXGM) for workload prediction and Multi-population Ant Colony System (MACS) for dynamic multi-objective VM consolidation optimization to achieve an acceptable balance between conflicting objectives of decreasing energy usage, eliminating resource waste, and avoiding SLA breaches. The EXGM(1, 1) model excels in requiring minimal historical data for accurate predictions, facilitating effective migration decisions. The MACS algorithm optimally maps VMs to hosts, striking a balance between Quality of Service and power consumption within a reasonable timeframe. Experimental results, evaluated using the CloudSim real dataset, demonstrate EXGM-MACS’s superiority in energy efficiency, migration efficacy, and SLA compliance against benchmarks, presenting promising advancements for optimizing CDC operations.