Cloud computing discussions often turn to the future of cloud—even with my relatives. Over Christmas dinner last year, a lively debate erupted over whether or not all company-owned and operated IT would eventually disappear. One in-law who is a CFO of a software startup believes that, because IT was so easy to do in the cloud “with a credit card,” there was no longer a reason for any company to have a computer room. He cited the example of a large financial services firm who was moving all IT to public clouds. But another in-law pointed out that his company’s IT people had concluded that while cloud is good for some things, it wasn’t for all of his company’s IT and countered that the financial services firm cited was only moving front office apps to the cloud. Problem there was latency and security issues—but not at the public cloud site. They worried about latency and security in the network connecting them to the cloud.
I was asked as recently as yesterday whether or not I believed that all IT would eventually wind up in the cloud, a question generally translates to: Will AWS (Amazon Web Services) rule the world? My stock answer is that, when predicting the future of IT there are no absolutes. No matter how dominant, it’s hard—and in many cases impossible—for enterprises to replace an established, business critical IT practice with another, let alone move it in its entirety to a service provider. Mainframes were supposed to go the way of the dinosaur. They haven’t. It’s still a cash cow for IBM because the undeniable fact is that customers keep buying them. And I think it’s the same with the enterprise data center. It will likely be surprisingly resilient.
But back to credit-card-swipe IT.
Cloud is the most disruptive force ever seen in enterprise IT, whether or not you believe in total world dominance. Business users love its immediate access to services that corporate computer departments can’t provide (think Salesforce.com). Startups wouldn’t do IT any other way. These trends are coupled with a growing realization in C-suites that ownership of data offers more intrinsic value than ownership of IT infrastructure. So will this rising tide float all boats?
Actually no. In fact, two very large players who were well positioned to exploit the public cloud opportunity have abandoned those projects: Hewlett Packard and Verizon. HP concluded that they’re resources were better devoted to developing products that support cloud services rather than competing with the likes of AWS and Microsoft Azure. Verizon couldn’t get enough traction with would-be cloud users to make the numbers work—a two-day shutdown last year probably didn’t help.
So as the cloud phenomenon continues to advance, prospective enterprise IT customers should now try to gauge the survivability of a public cloud service provider. Why? Because when a CSP shuts down, customers can have very little time to get their data back. And if they don’t, it’s gone.
Recently published survey data reveals two clear public cloud winners: AWS and Azure. What’s interesting here is that both originated from two very different places. Microsoft Azure was and is a natural fit with Windows applications. And as Microsoft continues its forced march to ever better versions of the Windows operating system, enterprise IT administrators tasked with managing those desktops can understandably conclude that walking those apps over to Azure relieves a lot of user-generated heartburn. We’ll let Microsoft do the Windows migrations from now on. We have better things to do.
AWS is different, perhaps amazingly so. Amazon the online bookseller becomes a dominant force in IT by carving out a portion of its existing IT infrastructure and offering it for rent. Over time, the road to AWS turns into a super highway. However, I think it’s now difficult for someone, no matter how resource rich, to repeat. Verizon has demonstrated that it’s not enough to stand up a public cloud service and sing Oh Come All Ye Faithful.
A more likely road to success in the cloud is the Azure model. Make cloud a natural extension of something you already do in the enterprise. Oracle for example is making a major cloud effort this year as a logical extension to its applications and infrastructure lines of business. On the app side, there is portfolio of services and the Oracle storage group recently announced a cloud data archive that is connected directly to the latest version of its data center tape systems. Another is demonstrated by Equinix who is approaching the opportunity as a networking and colocation vendor offering the ability to create a unified multi cloud environment. Others will surely emerge.
But above all, cloud has to be easy. There’s now a legion of consultants ready to fast track timid enterprise users to AWS. When business users see no difference between an app hosted in-house or in the cloud, then why not let the CSP do the heavy lifting.
Enterprise IT is not a place for absolute predictions. Rather, it is better understood as a series of pendulum swings. Decades ago, it looked like all enterprise IT would be managed by outsourcers like EDS and IBM Global Services. Then executives began to feel that they were losing control and took back their data centers. The same scenario could play-out with cloud. Or we could wind up with the compromise between private and public cloud known as the hybrid cloud. Which leads me to my next most often asked cloud question: “Have you actually seen a hybrid cloud?”Back to Analyst Blogs