It’s summertime, and since I’m a work-from-home freelance writer, my once-quiet creative space is now occupied by two teenagers.
Fortunately, I still have relative peace and quiet until about noon, when, despite my own self-interest, I start pounding on doors and rousing the sleepyheads out of bed. After that, I am forced to contend with the steady drone of afternoon cartoons or the latest doings of the Kardashians as I try to create content that is both entertaining and informative.
The other day, it was Spongebob; specifically, the one with the fish hooks and the following warning from Mr. Krabs:
“They dangle down and draw you close with their pleasing shapes and their beguiling colors. And just when you think you found the land of milk and honey, they grab you by the britches and haul you way up high.”
And that’s what got me thinking about the cloud.
Like fishermen, today’s public cloud providers have their hooks in the water hoping to haul in the latest enterprise catch. The bait is price. The Amazon-Google-Microsoft race to the bottom is at fever pitch these days, leaving the smaller, regional players with little option but to match them or close up shop. This phenomenon is not unusual at the start of a new industry, but in this case it has already spawned its own axiom: Bezos’ Law. Cribbed from the widely known Moore’s Law, Bezos’ Law holds that the price of a unit of computing power will drop by 50 percent every three years. The term was coined by AppZero CEO Greg O’Conner earlier this year, and indeed, a look at Amazon pricing over the past six years has seen costs drop from 80 cents per “unit” in 2008 to 41 cents in 2011 to 21 cents today.
What does this mean for the enterprise? According to CIO Insight’s Jack Rosenberger, it will become very hard to justify the continued existence of in-house data infrastructure. At this rate, it won’t be long before terabytes of storage and processing become available for mere pennies per day, giving organizations that embrace full cloud operations a substantial leg up when it comes to profitability, development support and in general devoting more revenue to business growth rather than IT. The question CIOs face today is whether they want to be ahead of this curve or play catch-up to more far-sighted competitors.
But is this race to the bottom as beneficial as it seems? Only if you are a giant cloud provider that can turn a profit on the barest of margins, says Liam Newcombe, CTO of data center modeling software developer Romonet. For the enterprise, it will result in the loss of Tier III data infrastructure to low-cost, commodity service: It’s akin to giving up Macy’s in exchange for Wal-Mart. And since Bezos’ Law is more of a marketing concept than a technical one, like Moore’s Law, its ability to hold up over the long term is highly suspect. Once the small and medium providers have been weeded out, is it safe to assume that the survivors will quickly settle on a comfortable price point that more accurately reflects their overhead? This will still be lower than a traditional data center, to be sure, but the days of rapidly falling prices will be over, and probably sooner rather than later.
The fatal flaw in Bezos’ Law is that it assumes that cost is the only factor that goes into the infrastructure provisioning process. But CIOs must also contend with issues like security, availability, performance and the like, which is why some cloud providers are already looking to jump off the price-cut bandwagon by highlighting the added value that can be had for just a few extra pennies. Rackspace, for one, has started to break down its pricing structure along both infrastructure and support lines. This gives the company the ability to provide an apples-to-apples comparison with AWS and Google, and then draw attention to the TLC that customer can expect for both architecture and operations management. In this way, Rackspace hopes to appeal both to the enterprise executives looking to outsource infrastructure and the dev/ops team looking for a quick resource bump.
The cloud has often been referred to as the culmination of utility computing, but this is a misnomer. Utilities like water and electricity are pretty much the same no matter what the source. Computing comes in many forms and varying degrees of quality and can be employed, reconfigured and distributed in any number of ways. In short, computing will always require a personal touch in order to make it truly valuable.
The cloud is eminently capable of satisfying this need, but it’s up to the CIO to make sure that a tasty-looking lure like low pricing and easy provisioning is truly fulfilling, not simply hiding a hook that turns the enterprise into someone else’s meal.