You can do more with backup than just cloning your data

Make the most of the opportunities

Open access

With or without data reduction, a content store can still be used to restore a system – its metadata tracks which files need to be restored – but can also be accessed directly.

In endpoint and mobile backup, the collation of data that was formerly accessible only on that specific device is now a big opportunity to wring value from what was once a cost sink.

“The value to customers is less and less the backup, it is that the data gets centralised and made available to the administrators,” says White.

“Once you have all your data in a single repository with metadata, you can add other workloads to that. For example, we can add discovery enablement, governance, audit compliance, and also secure file system shares.

“It is becoming converged data protection. We are crossing the boundary between endpoint backup and the likes of tiering and governance.”

Lock agrees. “Data protection inevitably collects metadata and that can be turned into a content store, with appropriate governance layered on to control who can see what,” he says.

He notes that several other multi-role products exist, such as CommVault's Simpana, which uses a single repository to support a range of services, including backup/restore, archiving and enterprise search. But he adds that even with the trend to converge backup and archiving, it is important to remember that those two main activities still have independent roles to play.

“Data protection should be evolving toward generalised information management, but for most organisations backup is still an insurance policy,” he says.

“I would argue there is still a role for traditional backup – for example, once you are making decisions on big data, you may be legally required to keep a copy of what went into that data pool.

“Archiving will grow and grow in importance, though. There's so much data on a live system that is not accessed frequently, but you don't know if you can afford to get rid of it.

"A collateral benefit of archiving that data off can be simpler and quicker backups, because you have shrunk your primary data.”

Box of tricks

The complexity involved in building such data protection systems is also leading more users towards an appliance approach. This is a prime area for startups and established storage vendors alike, as they compete to build the most effective and attractive combination of features into a single, easy-to-manage platform.

For example Infrascale combines local backup for hot data with WAN acceleration and a cloud gateway for disaster recovery and archiving; and Rubrik claims its converged data management box can offload all those backup, versioning and data reduction tasks, again with cloud at the back end.

Others talk of unified data protection. Fujitsu, for example, uses Simpana software in its Eternus CS turnkey backup and archiving appliances.

But Lock warns that information management and archiving is still not simple. “It needs some form of data classification upfront. It needs a fresh look, which is difficult for IT to do. You need to ask, could the business get more use out of its data?” he says.

“It turns what is normally a cost into business value, plus there's the possibility of actually saving money, for example if you can reduce the amount of primary storage you need to buy.”

Death greatly exaggerated

And of course that content store itself will need protection, perhaps via replication. It could even be stored on tape, which shows no sign of dying off.

It remains the most cost-effective deep storage technology for the likes of media archives or static archives, especially in the form of robotic libraries such as Spectra Logic's giant Tfinity. This can scale to more than 400,000 tapes and several exabytes in capacity.

Lastly, while the need to back up and deduplicate virtual servers gets a lot of attention, much of the most valuable and at-risk data in any organisation is out at the edge. So when re-thinking enterprise backup, we also need to approach it from the opposite direction.

Backup has traditionally tended to take a system-oriented approach, albeit with the ability to retrieve individual files, perhaps via a self-service mechanism for end-users. This will still be required in the data centre, but in the wider enterprise the key point is increasingly the profile of the user and the files associated with that user.

This Dropbox-like approach treats users as the key element, using the metadata in the content store to track what they require to do their job.

Whatever platform you use, whether it be a laptop, a smartphone or some web-based service running in the cloud, the files and resources you need are the same. You do not need a dead PC completely restored as it was, you simply need access to your files plus the applications to open them with.

Not for the first time, it seems that businesses can learn from the consumer. ®