There are so many fascinating things about the history of high-tech but none more so than how the advancement of innovation renders very recent technology obsolete -- no matter how compelling it seemed just a few short years ago.
It's also fun to watch the predictions that follow the advancements and how they often prove to be anything but accurate. The 2013 New York Times bestseller "The Big Switch" by Nicholas Carr is a great example. The premise of the book was that computing was finally going the way of electricity in that the cloud would provide us with a method to buy computing as we do electricity by hitting a switch.
I never bought into the simplicity of the argument because electricity does only two things: It powers lighting and systems, be they mechanical or otherwise.
Carr also used the history of electricity in his argument, pointing out how businesses at one point manufactured their own electricity by locating next to rivers or windy locations but later bought it from more efficient power plants that specialized and could produce power for less.
Computing technology in comparison does an untold number of things that add complexity and a need for customization that electricity has never had. So while cloud technology is a huge advancement, it is not the last. Now we are beginning to see how newer advancements will challenge the hype.
Hyper-converged infrastructure and software-defined networking are changing the cloud dynamics and legitimizing the argument that on-premises may actually prove to be more cost-effective than pure public cloud deployments.
As CRN has recently reported, some of the greatest minds in high-tech have begun making the argument that software-defined networks and hyper-converged infrastructure drive down the cost of business and in many cases result in decisions to do more on-premises than less.
In my opinion, this is a result of how computing is really about solving specific business problems and that means needing a degree of customization, which is not something that comes from flipping a switch. At least not yet.
When you look at a hyper-converged stack that reduces the cost of hardware and the software-defined networking that puts more of what used to happen at the hardware level into the software, there are enormous cost savings.
Add in the fact that most CIOs today desire to have more control of their company's technology, not less. And then add the constant need to tweak, improve and customize the end results of technology spending -- and the argument that on-premises will be totally extinct dissipates.
Consider this statement from a recent CRN interview with Michael Dell: "What we have seen when you automate and modernize the infrastructure, software-define everything, and move up to the platform level, is that for the predictable workloads -- which are for most companies 85 percent to 90 percent of their workloads -- an on-premises solution is much more cost-effective," said Dell.
Is he correct about 90 percent? Who knows, but he is likely in the ballpark. If so, it changes everything over the long run.
This kind of change and its pace is typical of this industry. Luckily, it's also the reason why politicians have never been able to keep pace and impose the regulations that would surely stall future advancements and allow other unfriendly countries to catch up to our technology centers.
Thankfully, by the time the politicians get a handle on the technology and try to regulate it, we've moved on to the next thing.
BACKTALK: Make something happen. Robert Faletra is Executive Chairman of The Channel Company. You can contact him via email at firstname.lastname@example.org.