The other day I read yet another blog about the “old” ways vs. a “new” way to do data protection. The post was from startup Rubrik. Each post I read like this has great intentions. They start out with a bunch of new buzz words – the new one they talk about in this particular post, which I hadn’t heard yet, is “SLA Domains”.
As you read blog posts by companies who provide data protection solutions, you start to see a pattern. They all start out trying very hard not to use words like, copy, move, retention, backup, recovery, etc., but as you read on, about ½ through, all these words start to rear their ugly heads. It’s a pretty interesting patter. And by the way, there are over 100 companies that provide “data protection” solutions these days, so while the market is highly competitive, at the end of the day, all these products do the same thing and in order to compete, you have to try to sound different, but as you see, it is difficult.
It reminds me of a quote from Henry Ford, “If I had asked people what they wanted, they would have said faster horses.” Simply put, there is no innovation if we don’t begin to think differently.
The fundamental problem with data protection today is that each solution creates another copy of your data and moves that data all around your enterprise. I am not saying that you don’t need a copy of your primary data for protection, but what I am saying is that you don’t need to rely on the same old architecture, backup software, backup appliances, or tape to have a good data protection solution.
Here is what I find interesting. All of today’s enterprise storage arrays take snapshots. In addition, most all of them can create some form of replica off the array. These are storage services you have paid good money for. What are these data copies use for? Then, why does IT feel like they need to create another copy, using some other software and move that data copy to another storage platform or appliance? This is costly from a CapEx perspective and OpEx from a management perspective.
One fact remains constant, as data continues to grow, users are struggling to meet their data protection windows. Given that data isn’t expected to stop growing – IDC says 35% to 50% per year depending on your industry – how will IT continue to keep up? How much money can you throw at the issue to make another copy of your data, and to what end? I am sure all of these 100+ solutions for protecting data are good, but are they necessary?
Another interesting factoid from Gartner is, by the end of 2016, 20% of enterprises, up from just 7% today, will abandon traditional backup hardware and software for snapshots and replication. In theory this is great. IT has these snapshots and has policies to create lots of them, the question is, HOW are they going to be able to use them for data protection or even other use cases?
This is where an in-place copy data management solution comes into play. While copy data management may sound like a buzz word/phrase, like ‘deduplication’ it actually does what is says. It manages the copies of data you are creating with your existing snapshot technology using the public API’s. If clients can better manage their snapshots, and have a catalog of them, they can now actually use them for a number of business purposes.
Think about this example. When a user says to IT, “I need you to restore some data I lost.” What does IT say? The first thing they do is cross their fingers and say “I hope I have that data in a snapshot!”. And typically they do. The challenge is, they have no solution that tells them what snapshot has that data. So what does IT do? They go to their backup catalog, in their catalog there is a list of all thier data, and perform a recovery which is a time consuming arduous process. However, if you had a catalog of all of your snapshots and replicas and could search them, like your backup catalog, you could instantly mount the data and drive your RTO’s to near zero. And by using snapshot technology, your RPO is as often as you take a snap, so you can get more granular with your RPOs driving better business resiliency.
So, now if your using snapshots with a full catalog for operational backup and recovery – why would clients need another piece of software (backup software) and or hardware?
I do agree that backup is broken and finding a new way to perform efficient operational data recoveries for IT is a must. It is important to know that clients have already paid for the tools to be able to improve your situation, they just aren’t taking advantage of them. Buying more software and hardware isn’t helping your cause. In fact, it is hurting it.
So while VC’s are pouring money into ‘new age’ ‘backup’ products, they are only helping the industry to compound the data management challenges IT has, and not helping them to get ahead of their problems. It’s time to think differently about data management.