In 1865, the English economist William Stanley Jevons published “The Coal Question,” a book with a prosaic title that contained profound implications. Jevons set out to establish the size of England’s coal reserves, a critical question for industrial and naval power. During his research, he stumbled upon a curious paradox: As coal use became more efficient due to the advent of better quality steam engines, coal consumption rose rather than fell.
On the face it, this seems counter-intuitive. Improved efficiency leads to greater productivity which should lead to a resource savings. In fact, however, it meant just the opposite: As coal power got cheaper, people found more and more things to do with it. In the end, people consumed so much more coal at the lower price that the total amount spent on coal rose rather than fell. This phenomenon came to be labelled the “Jevons Paradox.”
Computing is, of course, a premier example of this paradox. Despite the enormously lower cost of computers that’s occurred over the years due to Moore’s Law, total spending on computing has steadily risen. The past few years have been an exception to this regular progression – most of the server companies have seen essentially flat revenues, meaning that shipments failed to increase enough to offset the lower cost of processing. Boiled down, this means that enterprises haven’t been installing enough servers to grow the market.
Has hyperscale led to hypersales?
This past quarter, however, broke the recent pattern – server revenues were up over 18 percent, year-over-year. According to Gartner, total server numbers grew 13 percent, with particularly strong growth in the so-called hyperscale segment – those servers used by large Web properties and cloud computing providers.
This growth spurt indicates that public cloud providers are experiencing a “Jevons Moment.” Users are flocking to cloud computing due to its lower cost, compared to previous computing options (on-premises or colo hosting). And you can expect that this is just the beginnings of gigantic growth spurt in server shipments, and consequently, in cloud computing revenues.
The mix of server types in the numbers indicates the transition we’re seeing as enterprises begin shifting increasing portions of their application portfolios to the cloud. We’re still in the early days of this sea change, with this quarter’s report providing just a hint of what’s to come. Jevon identified his paradox early in the massive adoption of energy consumption, the end results of which can be seen today. Energy is treated as a low-cost service applied throughout every aspect of society – something absolutely unimaginable 160 years ago.
[Related: Looking back at the year in cloud computing]
I believe, however, that the growth of cloud computing is likely to outstrip even the growth implied by Jevons Paradox. The paradox focuses on what happens when a good or service becomes less expensive to use, i.e., when you use less coal for a given amount of work, you’ll do more work and use more coal. It really says nothing about the good or service being easier to consume. And it’s cloud computing’s consumption ease – commonly referred to as agility – that will drive cloud growth (and thereby server shipments) way beyond what might be expected due to the Jevons Paradox.
Frictionless provisioning – faster, cheaper
This consumption ease is vastly underestimated even by those who espouse agility as the primary driver of cloud adoption. Compared to the lengthy, weeks- or months-long painful process that typifies infrastructure provisioning in legacy environments, cloud provisioning is virtually friction-free: from initial request to instance availability is mere minutes.
Use cases that were never even considered in the past due to the hard work associated with the provisioning process become commonplace in a cloud world. I’ve worked with many companies that never did real quality or load testing, because resources simply weren’t available, but are now doing both in the cloud, simply because it’s possible. Likewise, companies are now doing much more experimentation and prototyping because they can quickly access resources (and just as quickly discard if the trial doesn’t work out).
Last quarter’s jump will come to be seen, in retrospect, as the inflection point where computing’s nature changed. Simply put, we are on the cusp of an enormous transformation as computing resource provisioning transitions from expensive and complex to cheap and low-friction.
What are some likely outcomes we’ll see from cloud computing’s “Jevons moment?” Here are a few:
- Growth of server and component shipments way beyond what we’ve seen in the past. Just as the 20th century came to be seen as the time of the automobile, the 21st will be the century of the computer.
- Massive, massive data center construction by the biggest providers. We will see tens of billions of dollars invested over the next decade as providers struggle to keep up with the demand for cloud computing.
- Innovation and disruption across all aspects of our society and economy. I’ve written extensively about what IDC calls the Third Platform, but suffice it to say here that the Platform’s power will go well beyond upending market segments like the taxi industry into areas like health and medicine, human performance, energy, education and many others. The most important implication of the Third Platform is that change is coming to every part of life; depending upon your position and perspective, that offers amazing opportunity or frightening turmoil.
- Incredible pressure within IT organizations to change their natures and operations. If you work in an enterprise IT organization and aren’t thinking – right now – about how to improve your overall processes 10X you are asking to be replaced. Your CEO, board of directors and customers won’t care about your challenges: if you can’t make your processes match the speed of provisioning, they’ll find someone who can.
- A redefinition of jobs and work. The IT cliché of the moment is “full stack developer” – someone who understands all the software components necessary to build an application. What the future will call for is skills capable of meshing rethought business offerings with the IT underpinnings necessary to bring them to life. The move to design thinking reflects this redefinition; the difference in the near future is that design thinking will be everyone’s responsibility, not just that fancy outside agency.
One of the amazing things is how much our society has changed since Jevons first did the analysis that led to his paradox. In the mere span of 150 years, life changed for humans more than it did since they first emerged hundreds of thousands of years ago. The change that this next Jevons Paradox will bring will vastly exceed that of the original one. It’s going to be a wild ride.