Virtual Machine Migration Services in Cloud Computing
Virtual Machine Migration:
Virtual Machine Migration is one of the vital strategies in the field of physical machine virtualization that permits applications to be moved with their implementation settings across physical machines. Virtual machine migration is required for server consolidation, resource scheduling and load balancing. For a successful migration, interruption and migration time should be minimized.
Live virtual machine migration is a method which moves the whole operating system and its related application from one physical machine to another. The virtual machines are moved live with no interruption in the application working on it. The advantages of virtual machine migration comprise the maintenance of physical server energy, load balancing amid the physical servers, and malfunction tolerance in situations of an unexpected breakdown.
Virtual Machine Migration Services:
The various virtual machine migration services methods are given below:
1. Fault Tolerant Migration:
Fault tolerance permits virtual machines to continue work if any system component crashes. This method migrates the virtual machine from one physical server to another physical server based on the prediction of the breakdown. This method is used to enhance the accessibility of the physical server and avert the degradation of applications.
2. Load Balancing Migration:
The load balancing migration method intends to deliver load across the physical servers to enhance the scalability of physical servers in a cloud setting. Load balancing helps reduce resource utilization, augment scalability, accomplish fail-over, over-provisioning of resources, avoid blockages, etc.
3. Energy Efficient Migration:
The power utilization of a data centre is mostly based on the consumption of servers and their cooling mechanism. The servers characteristically require up to 70% of power utilization even at a low consumption level. So, there is a want for migration methods which preserve the energy of servers by best resource consumption.
4. Stop and Copy-Based Migration:
It is a non-live migration method. The virtual machine stops working on the source machine and all the memory pages are copied to the target machine. After copying all the memory pages, VM happens on target. Downtime and migration time is similar for stop and copy-based migration since VM doesn’t begin on the goal host until the entire pages are transmitted to the goal.
5. Live Migration:
Live migration of virtual machines permits the VM to migrate without any disruption while implementing its application. Virtual Machine Migration is a significant way for running applications and resources in big virtualized systems. It facilitates the handling of resources to be dynamically reasonable in the complete virtualized system across physical host limitations and it permits applications to be dynamically shuffled to develop consistency and performance.
6. Post-Copy-Based Migration:
In Post-copy-based migration, the VM stops working on the source and only its implementation state (CPU, register, memory pages vital to begin VM on goal) is passed on to the aim machine and the VM starts working the goal. Even then, all the memory pages aren’t transferred and continue to subsist on the source. When the VM requires a memory page, it creates a page fault and that consequent page is transmitted from the source to the target machine. When each of the memory pages is shifted to the target machine, the VM begins on the target host.
In this technique, the page fault is one of the overheads as shown below figure. The steps included in the post-copy technique are as follows:
1. The virtual machine is on the source machine.
2. The execution state of the virtual machine is transferred before memory is transferred and if the page generates page fault then the requested page is transferred.
3. The virtual machine is activated on the target machine, later.
7. Pre-Copy-Based Migration:
Pre-copy technique is usually employed for live migration. In the beginning, it shifts each of the memory pages to the target machine. Then, iteratively, copies of pages are altered in the last round as shown below figure:
The procedure is continued until the writable working set (WWS) happens to be small. When WWS turns small, it executes stop and copy functions and shifts dirty pages and the CPU state. WWS includes the pages altered in every round. The steps included in pre-copy migration are as follows:
1. Send all the memory pages to the target machine.
2. Transfer the modified pages.
3. Virtual machine active the target machine.
Function of Live Virtual Machine Migration:
The functioning of any live VM migration approach can be estimated by the following metrics:
1. Preparation Time: This is the time between starting the migration procedure and shifting the VMs processor status to the aim node, throughout which the VM continues to perform.
2. Down Time: This is the time when the VMs implementation is blocked. It comprises the transport of processor status.
3. Pages Transferred: This is the complete sum of memory pages shifted, comprising copies across all the time stages.
4. Resume Time: This is the time between resuming the implementation of VMs and the end of migration, all necessities on the source are eliminated.
5. Total Migration Time: This is the total time occupied by the migration procedure, from the beginning till the end. Total time is extremely significant as it influences the discharge of resources on both source and target nodes.
6. Application Degradation: During the migration of virtual machines from one host to another, the application functioning, which is working on that VM is violated.