IT as a Change Agent: It’s Time to Look Inward, Starting with Storage
For decades, IT has been considered a change agent for the organization, but there is a dirty little secret that we in IT aren’t always willing to confess: we are sometimes terrible at taking our own advice. What I’m getting at here is that while we in IT have been more than happy to push other business units to change, we have often been unwilling to look inwardly at our own operations to do the same. Our reluctance to change is sometimes for good reasons and sometimes for bad ones, but either way, not changing can have serious long-term ramifications.
Let’s take a look at our data center environments. For many, long days, nights, and weekends are consumed with care and feeding of this critical business resource. Data centers have become overly complex, which makes them expensive and tough to manage. At the same time, the business wants to maximize the investment in this expensive resource, so storage administrators are pushed to tweak everything possible to achieve this goal.
There are other needs around storage as well. The first is to address a major problem, which is this: IT has often purchased four or five years worth of storage all up front and then slowly — and hopefully — grew into that capacity purchase. From a pure economics perspective, this is crazy. There’s a good reason that widget manufacturers don’t build a bunch of widgets and have them sit on a shelf. It wastes capital; customer demand may shift; or widget 2.0 might get introduced while you still have 18 months of widget 1.0 puppy laying around. In other words, a just-in-time methodology enables you to react quickly to changing demand and avoid tying up a bunch of capital. IT needs to be able to consume storage in sips rather than bites.
The second need is to massively simplify storage. There is often far too much work to be done in getting an entire storage system up and running in a way that works well. At the same time, the business needs storage that is fast enough to support application needs, has enough capacity to store everything needed, and stays highly available. Oh, and it can’t cost an arm and a leg anymore, either.
It’s time to change. There are simply too many storage options available on the market today to continue to do the same thing we’ve always done. We need to buy storage differently and it must easier to manage than legacy storage.
Today, I’ve been watching Coho Data’s Andy Warfield talk about exactly this issue, which is one that’s near and dear to me. Coho is working to help customers rethink how they procure and manage storage. Whereas legacy storage forces you to buy a lot of disks up front that you may not need today, Coho, thanks to its scale out nature, helps you to take a sip-sized approach to the problem. You buy what you need. When you need more, you buy another shelf. It’s not a pure OpEx purchase since you still need to buy a shelf of disks, but it’s far closer to cloud-like consumption-based economics than a lot of legacy storage.
When do you need to expand, the additional capacity is automagically added to your storage pool, which is front-ended by a software defined networking switch that makes the expansion process absolutely transparent to the front end. Moreover, by being so easy to use, storage administration needs are reduced, helping to solve the complexity problem.
I’ve always liked what I see coming from Coho and am happy to see that, in addition to introducing an all flash array earlier this year, the company is adding what they term as “fit and finish” features to their scale out storage platform, including integration with Active Directory and syslog support.