Dual Memory Search - Harmony Search Algorithm for Load Balancing in Cloud Environment

In Cloud Computing (CC) environment, load balancing among Virtual Machine (VMs) with various task lengths, start time, and execution time presents a significant challenge. Effective load balancing ensures that workloads are distributed dynamically across VMs, preventing any from being overloaded or underutilized, which otherwise can cause excessive energy consumption. In this research, the proposed Dual Memory Search - Harmony Search Algorithm (DMSHSA) efficiently searches multiple solutions in parallel, allowing for optimal distribution of workloads across VMs. The DMSHSA enhances load balancing by optimizing task allocation based on the execution requirements, helping to minimize the response time and energy consumption. The proposed load balancing algorithm is implemented using Python for evaluating its effectiveness in terms of energy consumption, latency, and throughput. The developed DMS-HSA achieves a makespan of $\mathbf{2. 5 5}, \mathbf{2. 7 8}, \mathbf{2. 9 9}, 3.23$ and 3.75, respectively, with $20,30,40,60$ and 80 VMs. The simulation outcomes show that the DMS-HSA based load balancing scheme outperforms the existing techniques namely, Particle Swarm Optimization (PSO) Enhanced Q-Learning algorithm (QMPSO) and Parallel PSO (PPSO) algorithm.