Menu
Tech Watch: Virtually protected

Tech Watch: Virtually protected

To look at the backup of a virtual environment, a very simple statement rings true: it’s different to backing up a physical environment.

The reason behind that is just as simple: a physical server in a virtual environment might have 10 workloads or virtual machines in the server. With those 10 agents, one backup drive is not going to sufficiently scale because it simply will not back up the data in time.

“If you think about the typical time it takes to back up an exchange server overnight, it could be three hours,” Citrix server virtualisation regional manager, Asia-Pacific, Dino Soepono, said.

“If you had another 10 of those workloads on a virtual server – you times that 10 by three, and you have 30 hours.”

Obviously enough, 30 hours is more than a day’s possible backup window, so by sticking to old processes, organisations are introducing the risk of not being able to back up all the data in the server, and on top of that is a performance degradation that is introduced to the server for all users to compound the pain.

The tools are readily available to resolve these issues – if anything there is too many choices available. There are a plethora of agentless solutions available that do backups of virtual machines while they are running throughout the day.

With an agentless solution, a user will be looking at offloading the requirement for an agent using a proxy technology or some form of interface to allow that backup to be taken away from the guest and managed in a centralised fashion.

VMware, and the other major virtualisation vendors, Citrix and Microsoft, offers an agentless solution as part of its vSphere 4 product. vSphere 4 integrates with VMware snapshot technology, which takes a “shot” of an entire virtual infrastructure environment, and is then able to take that off to a proxy server and from there move the data to tape or disc.

VMware senior product marketing manager for vSphere and vCentre, Asia-Pacific, Michael Warrilow, said the vendor was all about meeting recovery time objectives and recovery point objectives.

“The attraction to taking a more virtualisation-aware approach is that you’re shrinking that potential frustration window a business user might have in regards to recovery,” Warrilow said.

“Backup and recovery go hand-in-hand.”

But how do you test that a recovery will go smoothly? According to rising star Veeam, it’s a genuine concern for customers. To that end, it is building into the next release (slated for a Q3 launch) of its solution - a feature called Virtual Lab. Through it a customer will be able to take a backup and then verify that backup without additional hardware requirements.

“Customers are looking to achieve higher server consolidations, and what I’m seeing is what’s driving that is application workloads, whether it’s Exchange or SQL or Oracle. They’re looking for solutions where they can recover quickly and have this in mind basically,” Veeam Software strategic alliance and channel manager, Asia-Pacific, Simon Higgins, said.

It’s only a problem when it’s a problem

Of course, customers are not necessarily going to care about these issues - until something goes wrong. A virtual backup solution should epitomise what the typical customer wants from its IT infrastructure: for it to work, quietly and in the background.

“I used to joke that backup is boring, but backup should be boring because once it becomes interesting that’s when there is a problem,” Microsoft Windows server and infrastructure product manager, Rosemary Stark, said. “I think sometimes people do a set-and-forget, or they don’t think about the tiers of data they have in their environment.

“It’s those kinds of things that we’re focused on is ease of use and democratising the solutions so it’s available to more people, and more people can have no just the tools but the best practices.”

Microsoft backup technology, too, takes snapshots of the operating environments. The vendor parks the snapshot up to the application so that the snapshot of the application is in a known state, and not in a crash recovery state.

In a bid to make the virtualisation backup solutions run more smoothly, Microsoft partners with various storage vendors to create a more integrated package that maximises overall efficiency.

The operating should not do all the heavy lifting, Stark claims, so extending a backup solution into the storage subsystems should make life easier for the customer.

“If you’ve got that intelligence in the storage layer, then backups get passed down to the storage layer to go the heavy lifting because they’re the ones that know storage best,” Stark said.

HP, too, approaches server backup from an infrastructure level. HP Storageworks business manager, Mark Nielsen, said the vendor considers storage and the virtualisation platform intrinsically linked.

“One of the things we do for our resellers is train them on how to talk to their customers about virtualisation in a holistic manner. If you look at our converged infrastructure message, it talks about looking at infrastructure from a holistic point of view, so servers, storage, networking, power, cooling and management,” Nielsen said.

“It’s important to look at how all of those things integrate together to provide a single virtual environment from both a server and storage perspective.”


Follow Us

Join the newsletter!

Error: Please check your email address.

Tags VMwareCitrixvirtual serverDino Soepono

Show Comments