An Adaptive Pre-Copy Live Virtual Machine Migration Framework Using Fire Hawks Optimization Algorithm for Efficient Cloud Resource Management

Abstract

The Live Migration of Virtual Machine (VM) is essential in cloud computing to ensure the service is not interrupted. This becomes especially necessary when performing system maintenance, load balancing, or resource management-related operations. One of the most popular migration strategies is the pre-copy, in which the virtual machine (VM) continues to operate while its memory contents are transferred to the target host. Nonetheless, this process can be quite inefficient depending on the Dirty Rate (DR), this is the number of pages of memory that are changed while the migrating is occurring. An inappropriate DR may lead to longer migration time, longer down time and ineffective resource usage. To tackle these challenges, this research paper proposes an optimized Pre-Copy Live VM Migration (PLVM) technique that adopts the Fire Hawks Optimization (FHO) algorithm to adaptively select the best Dirty Rate during migration. Total migration time, downtime, and resource are used to create a fitness function utilization to optimize performance for migration. The process of optimization seeks to reduce migration overhead and enhance effective utilization of computational resources. The FHO-based PLVM technique is implemented using CloudSim simulation environment and evaluated using NASA workload traces to represent real user request pattern. The proposed method's effectiveness is contrasted with that of current migration methods, which are Machine Learning-based PLVM (ML-PLVM) and OMA-based PLVM (OMA-PLVM). Test results show that the method reduced downtime and migration time while significantly improving CPU and memory utilization. The results indicate that the efficiency of live VM migration in cloud data centres may be enhanced by the suggested optimization strategy.