Thursday, May 9, 2013

Why Virtual Machine Backup Is Better Than Your Backup Manager?

1:44 PM
Every company understands how important it is to back up data or even archive emails for future reference. Backups are not only crucial for disaster recovery purposes, but also for the long-term business continuity requirements of your company. With the recent growth in virtualization as a way to maximize resources and implement consolidation measures, companies will need to change the way they think about backups. Now, more than ever, it is necessary to consider your virtualized assets when deciding what to back up and how.

virtual machine backup

Virtualized data just as stored in a physical environment

"The big thing that’s changed relatively recently is the criticality of what we’re virtualizing," says Dave Russell, research vice president at Gartner (

It’s important to come up with a VM (virtual machine) backup policy that is equal to your policy for traditional data backups. You need to prioritize virtualized data and applications just as you would data stored in a more physical environment or else you may not be able to fully restore your critical systems in the event of a temporary outage or widespread failure.

The important thing to remember about virtual machine backup is that it isn’t exactly the same as traditional backup environments and that doing business as usual from a backup perspective won’t typically work effectively in a virtualized environment.

One of the biggest problems facing companies today is the sheer amount of solutions on the market that are specifically built for smaller implementations. For instance, you may start with a small pilot project and your virtualization administrator can download a tool and feel very confident it’s going to work, but once you start to deploy 300 or 500 virtual machines at the same time and reach critical mass, "the cracks will start to show."

physical server

VM backup isn't same as traditional backup>

Traditional backup is a resource-intensive process that can fill up your input and output streams as can other resources such as CPU and memory. Because virtualization is all about maximizing efficiency and putting numerous virtual machines on one physical server, it can lead to overloading your equipment. If you were to implement 20 VMs on one physical machine, stack or increase the density of the VM images, and then try to back it up with 10% of your resources, you would be oversubscribed by 200%. You need to make sure you have systems in place that can handle that amount of traffic and prevent overtaxing your infrastructure.

Another key to successful VM backup is to consider what you’ll need to protect in the future. Rather than focusing on the here and now, you’ll need to project where you see your company a year or more from now. One solution may work perfectly fine with five virtual machines, but as you grow to 50, 500, or 1,000 machines. From there, you have to decide what types of plug-ins, applications, and reporting tools, you’ll need. That’s where doing your due diligence upfront will really pay off.

When choosing a virtual machine backup solution, most any program will do in the short run, however it is important to choose a solution that will meet your company’s needs over a long period of time.

Some backup solutions are wizard-like and walk you through the selection process. Russell says, "will require encoding from the ground up." Your company's backup needs and staff expertise will likely dictate whether you’ll need a fully formed solution or something more customizable.

You'll also want to consider disaster recovery and other factors when shopping for backup solutions. You may be able to store VM images on a server, but will you be able to access them immediately if your company experiences an outage.

standby server

Choose solution that meet company’s needs over a long period

Russell recommends finding a solution that lets you make the target location (where you’re writing the virtual machine) a cold standby server for your company to use in case of an outage. That means that if there’s a failure of any kind in your primary system, you essentially have an up-to-date replacement waiting in the wings.

As with any technology investment, cost will always be a factor. It’s important to speak to vendors to figure out how they charge for their services and whether they are compatible with your internal systems.

Some vendors also charge by the socket or by the terabyte, You will need to feel what type of model is going to work for you at scale when you deploy it throughout the environment. The more accurate you are with your projected number of virtual machines, the better chance you have of not overpaying for a solution.

Once you have a backup solution in place, you can’t simply set it and forget it. You need a game plan that details the systems you need to back up and how often. The frequency really depends on several things but most fundamentally the criticality of the data and the change rate of the data.

With some systems, the new and modified data is pretty infrequent, so you could argue that you only need to protect them once a week. Other systems are either so important or are being added to and extended on such a regular basis that maybe snapshotting them once an hour may be appropriate. You can save money as well as crucial resources by backing up only what is necessary and following a strict policy.
This is the most recent post.
Older Post


Post a Comment

Toggle Footer