Some of the things to consider before Migrating to virtualized technology

What is going to be virtualized?

Is a business's entire infrastructure going to be virtualized or just its servers? Is a business going to use application virtualization or storage virtualization technologies? These are important questions to which a clear answer must be given, before any migration takes place.

Preparing for Downtime

No matter how carefully a physical to virtual migration is planned, there is bound to be at least some downtime. During the planning stage a business should work out how long this downtime is and what they can do to minimize it or continue running with the very least amount of downtime.

Sometimes downtime can be as short as switching off a physical machine and replacing it with a new physical machine that hosts virtual machines. Other times it can be as long as waiting for the installation of a bare metal solution to be completed. No matter what the scenario is, a business should take the necessary steps and precautions to prepare accordingly.

Hardware Requirements

This involves working out whether virtual machines can run efficiently on existing hardware or not and if so, how many are required. This is important because if, for example, existing servers could only run two virtual machines efficiently, then the cost of migrating from a physical to virtual environment using existing hardware would be a waste because a ratio of 2:1 is not very good. In this case, the most efficient solution would be to replace existing hardware, so that five virtual machines could be run per physical server, for example. In the long run, this would be much more cost-effective than the first solution.

However, a business also wants to be careful that they do not overcompensate. For example, if a business only required five virtual servers per physical machine and they ordered in servers which were capable of running ten virtual machines, this would be a fifty percent waste of both money and resources.

Storage Requirements

Virtual machines are encapsulated within files on a physical machine's hard drive. This means that hard drive or storage volume requirements are a big factor when dealing with virtualization. If storage is centralized, i.e. physically located within a server, then a specific storage volume or hard drive capacity would be required. For example, if a single physical server was going to be used to host five virtual machines each with a storage volume capacity of 100 GB, then at the very least, this server would require at least one 500 GB hard drive.

If storage was distributed, for example via the use of a SAN (Storage Area Network), then an organization would need to ensure that their SAN was large enough to accommodate all of its virtualized machines. If it was not large enough, then an organization could increase the capacity of their SAN by adding more storage volumes.

Ensuring that all Software will continue working in a Virtualized Environment

There is nothing worse than finding out that important programs or software applications no longer work after a business's infrastructure has been completely virtualized. Certain software applications won't work in a virtualized environment and may not support virtualization at all. A business should list all the software applications they will be taking forward when migrating from a physical to virtual environment. They should contact the application's vendor or test the applications themselves, to see if they will work.

Naming Conventions

Organizations usually have their own naming conventions in place to help identify servers, end-user desktops, etc. When a P2V migration takes place but physical machines still exist (which do not host virtual machines), it is often practical to name virtual machines in such a way, so that they are easily identifiable from their physical counterparts.

If this is done, then these changes will also have to be made at the network level, meaning an IT professional has much more work to do. If naming conventions are not changed, then the situation is effectively a 'plug and play' one, meaning that a virtual machine can be connected back on to the network without having to reconfigure any network setting (assuming of course that no connectivity changes have been made).

If a fast and easy migration is required, then naming conventions and network parameters should remain the same. If an organization has manageability in mind, then naming conventions and network parameters should be changed so that virtual machines are easily identifiable.

Once the migration process is completed, there are a number of things that IT professionals should do before they actually give the virtual machine a 'green light' and let it go live.

________________________

For more information on Virtualization and to sign up for our Cloud Computing Certification courses, have a look at the Virtualization Certification Kit.