Why (public) cloud will eclipse all other forms of IT?

Over the next five articles I’d like to make a set of bold, audacious and somewhat outrageous predictions about the future of cloud and IT. The intent is to spark a conversation and bring together divergent points of view in an attempt to remove biases and blind spots. So, let’s start with an easy one:

Prediction #1: (Public) Cloud will eclipse all other forms of IT.

Depending on who you talk to, and their professional affiliations, you get one of two kinds of responses to this: “Duh!” (This is so obvious, what’s there to debate?) or “Nah!” (We have seen this movie before and on-prem IT will evolve to meet this challenge but never be replaced entirely).

There is some truth to the latter statement – I don’t think on-prem IT will ever entirely go away. I mean, there are customers who are still buying mainframes and RISC based servers. But this technicality distracts from the larger point that all new deployments would eventually move to the public cloud, including on-prem IT as a refresh cycle rolls around. Rather than falling for the emotions let’s explore the reasons behind the momentum for public cloud and how some of the inhibitors would eventually be resolved.

Why (public) cloud?

To me, this is the next stage in specialization that enables companies to focus on their core expertise as opposed to building and operating a data center. There was a time when companies like Sun, HP and IBM owned and operated their own manufacturing facilities. No HW company builds its own servers anymore. Even the large scale cloud operators rely on ODMs. It’s not just the cost of running a DC, but everything else that we have come to accept as routine in IT such as patching, replacing broken HW, refreshing deprecated HW, troubleshooting performance issues etc. While not all of them magically disappear in the cloud, at least it makes it someone else’s problem. No one thinks about installing their own power plant, water purifying plant or cell tower. Owning and managing a DC was a similar necessary evil for most companies. Not anymore. They can focus on their core competency and leave the chore of IT to the experts who have the scale, expertise and budgets to do it well.

The other two reasons often cited in favor of cloud are elasticity (start small, pay only for what you use and scale resources up/down quickly) and ubiquitous access (all you need is a network connection and you can access your IT from anywhere in the world). However, these two benefits could just as easily apply to private cloud, which as I would argue is only a transitionary fix.

Why now?

The vision of IT as a utility that could be accessed similar to a power outlet or phone jack on the wall is both alluring and enduring, but had proven to be elusive over the decades. There are three technology trends that finally came together to make it possible.

1.      Software Defined <Storage, Server, Networking>: this started with server virtualization and has now spread to other parts of the IT stack. Abstracting the underlying infrastructure through software gives tremendous flexibility in both scaling and migrating workloads.

2.      Ubiquitous networking: The 10GbE and increasingly 40GbE makes the geographic distance to the cloud increasingly irrelevant.

3.      Ubiquitous access: The BYOD (Bring Your Own Device) movement coupled with increasingly faster wireless speeds has given customers the freedom to access their information in multiple formats, anytime, anywhere.

Why private cloud won’t be enough?

People often confuse a shared infrastructure for the cloud. While a shared infrastructure is a necessary ingredient, it’s hardly sufficient for a public cloud. Notice in the “why public cloud” section, I didn’t mention cost as one of the factors. I don’t doubt those elaborate TCO models which show that public cloud (despite its PAYG model) is costlier than private cloud. But cost is not the motivating factor. The primary attraction of the public cloud is the freedom from having to own and operate a DC and maintain an IT staff which may not be your core competency. Conversely, from a financial viewpoint, not having to maintain all those heavy duty assets on your books, gives you additional working capital, more predictable cash flows and, best if all, tremendous leverage.

OK, but what about …?

Having talked about the benefits of public cloud and the inadequacies of the private cloud, let’s talk some of the objections. Some of the objections that are often cited against public cloud are: security, data governance and disruption to existing operations.

Security is a valid concern, but as some of the recent hacking episodes have demonstrated, your on-prem DC is no more secure than a public cloud. If anything, you might benefit by “leaving it to the experts”. The public cloud companies have a lot of incentive and lot of resources at their disposal to keep the intruders out by deploying latest technologies and best practices.  

Data Governance is also a valid concern but one that will eventually be addressed through geo specific data centers. It’s only a matter of time.

Finally, the disruption due to migration. This is a tough one and likely the hardest to overcome. Usually, the catalyst here is a compelling event such as license renewal or HW refresh. Upgrading HW or SW is no easy task even when it’s entirely on-prem. With a little extra (one-time) effort, you may forever be relieved from the burden of maintaining your own IT infrastructure.

So, what does this mean?

The emergence and rise of public cloud has profound implications for both the IT landscape (which I’ll touch upon in a future article) as well as the broader economy. In my view, the biggest impact of public cloud is in the democratization of IT. Anyone who has a great idea and a credit card (and not one with a high limit) can experiment and grow without having to become an IT genius. Conversely, companies which in the past have used IT as a competitive differentiator will no longer be able to do so.

What do you think?

Let me know what you think. Like I said in the beginning, the whole point of this series of articles is to spark a debate and hear divergent points of view. I’d love to hear if you agree/disagree with the core assertion or if you agree/disagree with any of the arguments made in support of that assertion, or, even if you have an entirely different take on the whole thing. Please comment below.