Trouble is brewing in cloud land.
Maybe AWS chief Andy Jassy started it all, deriding the private cloud as “archaic” and not really cloud at all. Then Red Hat CEO Jim Whitehurst told me that public clouds like AWS become “obscenely expensive at scale” -- which, of course, sent public cloud advocates into a frothing rage.
But not an unstudied rage: As CSC’s Simon Wardley, Google’s Miles Ward, and others argued in response, public clouds like AWS are not, nor need not be, more expensive than private clouds. The opposite is true, they maintain.
Expensive mistakes
Embedded in Whitehurst’s argument is that while public cloud can work for disparate workloads, nonvarying workloads are best “hard-wired” in a private cloud. If you know the compute/storage requirements, why not run those workloads on dedicated local resources? A modest capital investment ends up being a lot cheaper than long, drawn-out, public cloud operating costs.
That may be true for customers who run AWS without thinking, argues Gartner analyst Lydia Leong, because they fail to collect savings that are available with minimal investment of time. If it’s more expensive, maybe they haven’t optimized their AWS pricing.
But a deeper problem is that such “nonvarying workloads” may be more imaginary than real. At the least, she continues, “‘will never change’ is often different from an operator's perspective than from a dev's perspective.”
Piling on, Brocade executive Bryan Beal notes that “private cloud is obscenely expensive if not being run at 100 percent capacity all the time.” The dream of optimal resource utilization in the data center is more fanciful rather than real.
Perhaps Google's Ward offers the bluntest assessment: “Whitehurst is incorrect. I do lots of this math, and I've yet to see it pencil out in favor of buying.”
While Ward obviously has a built-in bias for public cloud, he has made his calculations public for anyone to rebut.
Regardless, “Private [cloud] remains little more than a glorified virt[ualization],” CloudWeavers CTO Carlo Daffara insists. As such, it fails to unshackle enterprises from their hardware in the same way that public cloud does -- with lesser effect, whatever the cost.
In the future, when all’s well
Much more is at stake than a game of riposte and parry. Sure, the back-and-forth heckling superficially resembles a Monty Python argument clinic:
Man: Aha. If I didn't pay, why are you arguing? I got you!
Mr. Vibrating: No you haven't.
Man: Yes I have. If you're arguing, I must have paid.
Mr. Vibrating: Not necessarily. I could be arguing in my spare time.
But this argument over the relative value of private and public clouds is really about the future of enterprise computing, with $3.8 trillion at stake. Although the voices in favor of the private cloud are loud and insistent, CSC cloud researcher Simon Wardley correctly characterizes the market: “Every year, for last five plus, I hear the same story: 'This is the year private gets big, it's the future!'”
But that future never seems to arrive.
Instead, the future is very clearly moving toward public clouds, with VMs growing at 20-fold rate, while private cloud VMs manage only three-times growth, according to a Gartner analysis.
Why? One explanation: “New stuff tends to go to the public cloud ... and new stuff is simply growing faster" than the traditional workloads that currently feed the data center, including data centers of the private cloud variety.
This may mean that while CIOs are telling Whitehurst that their “apps that don't vary a lot in usage are significantly cheaper to run in their own data center than on the public cloud,” developers within those same organizations are almost certainly finding agility (and cost savings) in the public cloud.
The numbers don’t lie -- and those numbers favor public cloud.