This article can also be found in the Premium Editorial Download "Storage magazine: How to improve your virtual server storage setups."
Download it now to read this article plus other related content.
Many storage shops are fighting a losing battle when it comes to data storage and backup protection, with too much data and not enough time. Maybe it's time to rethink the process.
It's 2013 and we're still doing backup. I know things tend to toddle along kind of slowly in IT, but the way we protect our data has barely changed at all in my lifetime, and I'm hardly a kid. Backup keeps getting better, of course, with techs like data deduplication and bigger, faster storage targets to send all that backup data to. But today's data storage and backup process requires essentially the same oversight and administration it did 10, 20 or 30 years ago.
Few things stagnate like that in the tech universe without a reason, and there are plenty of reasons why backup has been lingering in a time warp for so long. Along with everything else in this world, if you follow the money things start to fall into place and a lot of the mystery disappears. Backup is a complex process involving software, hardware, manpower and special expertise. That makes backup a pretty lucrative endeavor for vendors and offers little motivation to disrupt the status quo by doing anything that might simplify or streamline the process.
We can blame ourselves, too. After so many years, we're set in our ways and just accept that we need special people with special skills to ensure all that backup hardware and software runs right. There's definitely a "because that's the way we do it" mindset when someone asks why so many copies are made or so much data is backed up. And there's no denying that changing backup procedures and apps is no small matter, especially if you have a lot of old backup data stashed on tape.
There's some blame that can be shared among vendors and users alike. With the inexorable growth of file data about to put a stranglehold on storage operations in many companies, there are probably a lot of storage pros who realize that keeping everything is impractical but still don't know what data can be ditched. That usually means shops are backing up data that should have been discarded and making multiple copies of useless stuff. Six or seven years ago, a few startup vendors helped define a new storage product category with data classification products that could help cull the goodies from the garbage. But they were largely ignored and the concept of data classification has mostly gone away, leaving a bunch of backup admins trying to figure out what to keep and what to toss.
Since the great recession put the brakes on a lot of planned storage purchases, data storage managers have learned to use their installed storage as efficiently as possible. Some of that efficiency also shows up in backup operations, but not enough to stem the tide of spiraling data stores.
So maybe it's time to rethink data storage and backup. Just as RAID and erasure coding are built into systems and run virtually unattended after some setup and with a little oversight, backup ought to be far more integrated with storage systems and processes. As an external process, backup isn't cutting it; many companies are struggling to complete backup jobs before the morning shift arrives or trying to cope with voluminous, multiple backup sets. And with so much pressure and so little time, it's inevitable that some data -- maybe some important stuff -- is slipping through the cracks. A few companies have even stopped trying to back up everything; they just spin off copies of the critical things and keep their fingers crossed about the rest of the data.
It has to be a lot easier than that. Let's just get rid of backup. The tools and technologies are here, they just need to be better integrated. By now, one would have thought that continuous data protection (CDP) would be an integral part of everybody's data protection plan, but adoption has been slow. On our most recent Purchasing Intentions survey, only 18% of respondents said they're using some form of CDP and 13% said they plan to add it to their backup repertoires this year.
The combination of CDP plus snapshots plus replication could be the wonder drug that cures backup's ills. Rev up CDP, and without doing anything else, your company's data gets backed up not in big unwieldy batches every night, but in little drips and drabs throughout the day. Now what if CDP was built into the storage system's operating system just like other services? You might have to turn a few dials to set how frequently the new or modified data should be scooped up and where it should be sent, but the process would be nearly invisible and cause far less disruption than traditional backup operations.
Of course, even if backup was tightly integrated and largely unseen, you'd still need to monitor it to make sure it's doing what you expect it to do. You need to know, for example, that database tables have been copied in a consistent manner using Microsoft VSS or a similar technology. You'll also want to control what gets copied and when, so you'll need to be able to throttle CDP services appropriately for mission-critical data down to user files.
There are many more details that you'd need to work out, but if a decent data protection management app was tossed into the mix, you could keep tabs on the process and be warned if something jumps the tracks at any point.
Backup is tough, but it's tougher than it should be when it operates in isolation from the rest of a storage system's processes. A number of startups -- as well as established players -- are beginning to address the issue, but the required degree of integration is not yet in sight. But as data continues to grow and storage managers get more desperate for a data protection scheme that works, perhaps storage vendors will pay attention. And maybe then you'll be able to forget about backup.
About the author:
Rich Castagna is editorial director of TechTarget's Storage Media Group.
This was first published in May 2013