I’ll admit it -- I’m a huge fan of the Tour de France. Watching the world’s best cyclists thread through small French villages and fly down steep mountain roads is riveting. I consider the Tour to be the most challenging athletic contest on the planet -- a month-long grueling effort requiring hours of maximum effort each day.
One of the most fascinating phenomenon in the Tour is the breakaway -- where a small group of cyclists sprint away from the peloton to achieve victory in the day’s ride. It’s watching a breakaway where one recognizes a core truth: even in an athletic contest that draws together the best riders in the world, there are a few that are far better than the rest. They bring something special to the competition.
Something similar to this is going on today in cloud computing. After a decade of competition, three providers have left the rest of the industry trailing far behind. Today, AWS, Azure, and Google are way out in front of everyone else, so much so that it’s impossible for anyone to catch them.
This raises two interesting questions:
1. How is it that the big three accomplished this? After all, it’s not like offering remote computing services is terribly difficult to understand. Nevertheless, the big three have vanquished everyone else in the space -- so how did they do it?
2. What does this dominance mean for IT users going forward? Put another way, with three big suppliers commanding most of the revenue in cloud computing, what should users do?
It’s Not Just IaaS
I would argue that the NIST IPS model (infrastructure, platform, software) is too simplistic. While many cloud providers viewed themselves as being in the on-demand infrastructure business, the big three viewed themselves as being in the enabling applications business. On-demand infrastructure is a necessary-but-not-sufficient element of the enabling applications business.
So, while aspiring cloud providers were hailing their use of “enterprise” equipment in contrast to the commodity kit used by the big three, the big three were building out a broad portfolio of computing services -- queues, notifications, event management, rich processing alternatives (e.g., very large memory instances), highly scalable managed databases, and more.
Even today, ten years on, looking at most of the offerings of the rest of the cloud providers shows how poorly they understand the business they’re in. Few provide anything beyond standard instance types, block and object storage, and core network connectivity. The really advanced ones might offer load balancing services, but most of the rest just point users to a partner list or suggest they leave it up to the user to implement their own preferred solution. In other words, they completely misunderstand what Clayton Christensen would refer to as “the job to be done.”
Looking forward, the gap between the big three and everyone else is just going to widen. All of them are offering instance types specialized toward machine learning. Google is releasing enormous data sets so users can develop better AI applications. Microsoft is building out an IoT framework to help users create connected devices faster. And Amazon is, of course, creating a rich smart home hub via its Echo/Alexa service.
Users: What About Lock-in?
From the foregoing it’s obvious: the cutting edge of computing over the next decade is going to rest with these three cloud providers. If an IT organization wants to avoid being relegated to also-ran status, it’s going to have to use one of them as a core part of its computing strategy.
That raises the spectre of lock-in, and there’s no shortage of industry leaders warning users about the dangers of being too committed to, say, AWS. Curiously, many of these would-be Cassandras displaying concern for end user vendor handcuffs previously delighted in binding users with enterprise license agreements, audits, and relentless sales pressure. One might say there’s no saint like a reformed sinner.
Whether or not this solicitousness on behalf of users is genuine or pure artifice, it misses the point. If you are an IT organization and want to perform on the leading-edge of computing, the question is which of the big three you’re going to use, not whether to use one of them.
There is enormous pressure on enterprise IT organizations to help their companies become digital enterprises. The benchmark for customer expectations is clear: offer the efficiency and convenience of an Uber or Airbnb. It’s not just Silicon Valley companies that are achieving this, however. Capital One, located in the not-very-Silicon Valley location of Richmond, Virginia is making huge strides toward becoming a digital enterprise. The key isn’t location, it’s intent and commitment.
Five years from now, the lock-in scare will be seen for what it is -- a last-ditch effort by incumbent vendors to prevent customers from grasping the obvious: most of the innovation in the industry resides with the big cloud providers. Companies that don’t want to be left on the dust-heap of history need access to that innovation and will embrace it despite putative lock-in concerns.
When I was first exposed to cloud computing a decade ago, I recognized it for what it is: a profound transformation in the nature of computing. I immediately felt it was the platform for the future of enterprise IT. Nothing in the ensuing 10 years has changed my mind. In fact, I would say we’re still early in the transformation process.