Optimizing the Performance of Live Migration of Virtual Machines using Compression Technique
Virtual Machines provide the environment for the cloud users same as they would be working on dedicated hardware in Cloud Computing. Virtualization enables use of computing resources in efficient manner, which reduces the physical hardware systems required to provide services. Virtual Machines need to be moved across different hosts for balancing load across the servers and for maintenance of servers. While migrating the Virtual Machines, there is a high-risk factor in performance degradation due to high workloads on virtual machines and migrating virtual machines over low bandwidth channels with existing algorithms. In this situation the memory pages of virtual machines are dirtied faster and increases the migration downtime. In this work, compression based Adaptive Virtual Machine Migration (AVMM) method is proposed, which compresses the memory pages during transmission in order to reduce the downtime and increase the migration throughput. Simulation results illustrate that Adaptive Virtual Machine Migration method outperforms the standard pre-copy algorithm by reducing downtime by 60 percent. In another scenario a virtual machine with streaming video server is migrated with acceptable downtime and the picture has frozen for six seconds with standard approaches. The proposed Adaptive Virtual Machine Migration (AVMM) method successfully migrates the virtual machines with in the acceptable downtime compares to standard approaches in cloud environment.