Is the cloud another fad, promoted only by buzzword happy consultants, or is it the future of IT?
I don’t think anyone really knows. Things certainly aren’t helped by the contradictory and sometimes incoherent definitions of what the cloud actually is.
At the moment not enough organizations have moved into the cloud for anyone to make a determination as to whether or not it’s a really great idea. It takes a few years until after a revolutionary change before people work out whether they’ve improved the process or driven themselves over a cliff. It’s a bit like outsourcing. When outsourcing came along, a lot of buzzword happy consultants could only say positive things about it. Save money! Outsource! Get rid of that pesky bunch of nerds in the IT department! A decade down the track and we hear a lot of organizations that were gung-ho about outsourcing early on have pulled a lot of stuff back in-house. The efficiencies that looked like they were there often turned out to be phantasms.
It doesn’t help that the “cloud crowd” can be maddeningly short on detail when it comes to explain how a cloud deployment will benefit an organization more than a traditional deployment. Sometimes the messaging seems really off. If you’ve been to enough presentations I’m sure you’ve heard the one about “the executive with the credit card who does an end run around IT and buys a cloud service to run what becomes a business critical app”.
(Whatever it’s meant to do, this story is part of the reason that a lot of IT people see the cloud as something that will take their job away as the story is designed to appeal to the exec and not the IT person.)
Whether or not the cloud takes over IT comes down to efficiencies. Cloud providers have vast resources that allow an organization to spin up and spin down capacity as it is needed. You leverage efficiencies as you only pay for that capacity that you use. The cloud provider’s vast resources lead to economies of scale. Just like outsourcing to another country lets you (in theory) leverage the lower wages of workers in that country to get the same work done cheaper, you can, (in theory,) get infrastructure cheaper from cloud providers than you can get in house because the cloud providers are deploying so many servers that it costs them less on a per-server basis.
We don’t know how close that concept is to reality. What works for some organizations won’t work for others and there is more to IT than infrastructure. As organizations have found with outsourcing, you generally can’t tell up front if you’ll save money. Put another way, you’ll only find out after you’ve jumped in the pool whether or not you’re going to drown. I suspect in the next few years we’ll hear stories of failed cloud migrations where organizations end up spending more to achieve in the cloud what they were already able to do in house.
Remember that cloud providers aren’t there to donate infrastructure out of the goodness of their hearts - they are businesses there to make money. This means that cloud providers set their rates based on a calculation on what it would cost to provide a service in house. They make a profit by setting their cost somewhere below this rate, but above the rate that it costs them to provide the service.
The costs of providing in house IT services are decreasing. Technologies such as System Center Orchestrator allow organizations to increase efficiencies locally by increasing the degree to which automation can eliminate repetitive tasks. These efficiencies are going to squeeze the margin between the cost of a cloud deployment and keeping things locally. The more efficient the local IT department is, the less compelling the case is for a whole scale move to the cloud.
So is the cloud another fad, or is it the future of IT? I’ll hedge my bets and say it is a bit of both. I think that enough organizations are going to run into problems “doing an end run around IT” that cloud strategies are really only going to work if there is a technical and economic need. My suspicion is that technologies like Orchestrator and virtualization are going to increase in house efficiencies to the point where using 3rd party cloud providers isn’t a slam-dunk business case.
Follow me on twitter: @orinthomas