Multi-Objective Cuckoo Search Optimization Algorithm for Optimal Resource Allocation in Cloud Environment

Abstract

Cloud technology refer to a pool of virtualized heterogeneous computational resources which are served to different customers as a service through the internet. These virtualized resources get provisioned to cloud users dynamically based on demand. however, dealing with resources in such a dynamic environment is still a challenging task. Deciding the way, the available cloud computational resources get allocated to the workload requires efficient provisioning and allocation techniques. The clients need to operate their various applications on cloud servers with variant Quality of Service (QoS) requirements based on the workloads and type of application whereas the objective of the cloud providers is effectively reducing the energy consumption by optimizing the CPU and bandwidth utilization and maintaining the Service Level Agreement (SLA). This, however, makes resource provisioning and allocation a very complicated task. This article implements the meta-heuristic Multi-Objective Cuckoo Search Algorithm (MOCSA) for efficient resource allocation to dynamic upcoming user workloads and scheduling user tasks on cloud VMs, which aim at minimizing the energy consumption with efficient utilization of CPU and bandwidth. The implementation and simulation experiments were conducted on the Cloudsim simulation toolkit. The evaluation results proved that MOCSA performed well in terms of performance, energy consumption, CPU utilization, bandwidth utilization, and SLA violation.