As domination goes, it’s hard to surpass Amazon Web Services (AWS). According to recent Gartner data, AWS now offers 10X the utilised cloud capacity of the next 14 IaaS and PaaS providers... combined.
For those paying attention, that’s double the dominance AWS established last year.
More ReadingPublic cloud storage spend to double in two years – reliable sourcesStuck in Amazon's web tentacles yet? You will be soonOpen-source Linux doesn't pay, said no one ever at Red HatBrace yourselves: Public cloud is coming. When? Er... soon, possiblyYou need one cloud computing provider - and this is why
And while public cloud spending remains a rounding error in the broader scheme of total IT spending, it’s growing at a 29.1 per cent compound annual growth rate, threatening to up-end legacy IT vendors.
Which is just as it should be. Cloud spending is on a tear precisely because traditional vendors have failed to deliver the convenience that enterprise buyers increasingly demand.
This leaves us with AWS assuming the role that Microsoft played for years. In the words of Hewlett Packard’s cloud chief Marten Mickos : “AWS is to the era of cloud what Microsoft was (and is) to the era of client/server.”
According to Gartner, in 2014 the absolute growth of public cloud IaaS workloads exceeded the growth of on-premises workloads (of any type) for the first time. Including all forms of cloud (IaaS, PaaS, SaaS), IDC expects the cloud-computing market to more than double by 2018 to $127bn.
The primary driver of all that growth? Convenience.
As RedMonk analyst Stephen O’Grady posits: “Convenience trumps just about everything” when it comes to cloud adoption. With developers playing an increasingly important role in IT procurement decisions, appeasing them has become Job #1. Developer appeasement starts with open source and often ends with public cloud computing.
And it’s only going to get worse. Or better. Depending on where your company is investing.
Consider that IDC pegs data centre capacity growth at 750 per cent through 2019, much of it driven by increasing IoT applications. But guess what? Those data centre investments are being made by AWS, Microsoft, and Google, not your enterprise data centre.
As IDC notes: "The agility and scale required in IoT deployments will ensure that much of that data centre capacity ends up residing in service provider [cloud] data centres."
Or take other big data investments. AWS data science head Matt Wood told me in an interview that the experimentation required in data science fits poorly in the hard-wired world of traditional IT or even private cloud:
Those that go out and buy expensive infrastructure find that the problem scope and domain shift really quickly.
By the time they get around to answering the original question, the business has moved on. You need an environment that is flexible and allows you to quickly respond to changing big data requirements.
Your resource mix is continually evolving — if you buy infrastructure, it's almost immediately irrelevant to your business because it's frozen in time.
It's solving a problem you may not have or care about any more.
"Obscenely expensive" public cloud
While you’d expect an AWS executive to sing ballads to the public cloud, his thinking makes intuitive sense. Big data is driven by variable workloads, requiring highly elastic infrastructure. A paean to the apologists.
But not all workloads are elastic, of course. As Red Hat CEO Jim Whitehurst recently told me: “The [public] cloud becomes obscenely expensive at scale.”
While cost isn’t the only consideration, Whitehurst continued: “A number of enterprise customers tell us that apps that don’t vary a lot in usage are significantly cheaper to run in their own data centre than on the public cloud.”
While AWS is great for early test and dev, and perhaps even essential, he reasons, once the kinks have been worked out of a workload it’s generally better to move that workload to a dedicated private cloud to improve performance and lower costs.
It’s a thoughtful response, and almost certainly true in some instances. Etsy chief technology officer, Kellan Elliott-McCrea, for example, tells me that the marketplace leader has dramatically improved performance and utilisation by dropping its public cloud services in favor of a homegrown data centre.
Unfortunately, many other IT buyers discover the exact opposite. Gartner polled 140 companies running private clouds, and discovered that 95 per cent of them have significant problems:
Despite this hit (err...fail) rate, some keep pimping private cloud’s staying power.
Take, for example, CenturyLink Office of the CTO executive Ed Saipetch’s defense of private cloud. In response to my contention that private cloud is “dead” (compared with public cloud’s growth) Saipetch counters: “Private cloud... won’t die or be dead. Just like mainframe and tape.”
Wow. I’m not sure I’d ever want to point to tape drive and mainframe longevity as a reason to believe in the health of a market. Sure, people still buy both, but I don’t think anyone outside IBM wakes up in the morning feeling excited about mainframes.
Not that I’m judging them.
Then there’s VMware vice president Matthew Lodge, who chides me that “all public cloud is now 20 per cent of all virtualised workloads.” In other words, still a sliver of the market.
Which sounds great, until you factor in Gartner analyst Thomas Bittman’s analysis of VM growth in the cloud, the fastest growing market. In the public cloud VMs have grown 20x in the past few years, while private cloud VMs have managed just 3x growth.
In other words, 20 per cent of all virtualised workloads growing at a 20x clip means ...er... many, many more public cloud VMs to come in the very near future.
Let’s go back to the numbers. AWS is now ten times the utilized capacity of its next 14 competitors combined. That’s amazingly big.
Microsoft, for its part, is sprinting to catch up, and by all accounts is making some headway, at least against the rest of the cloud pack.
Meanwhile, with the public cloud purveyors dramatically lowering pricing on a regular basis (Google just slashed prices by nearly 30 per cent), Whitehurst’s cost argument risks losing resonance.
But ultimately, the shift to public cloud is about one number: servers you need to worry about.
Which number, increasingly, is “ZERO.”
A recent OpenStack survey found that 77 per cent of surveyed IT organisations believe reduced time to deployment of services is the most important performance indicator for an OpenStack project. But if that’s the goal, the public cloud is almost certainly the safer bet.
As Envoy founder Larry Gadea declares of AWS Lambda:
FYI, this is crazy. We just did a backend call with full-on backend logic without setting up any kind of instances. No worrying about starting a new not-even-free-anymore Heroku Dyno. No choosing machine speed and memory and hard drive sizes. It just works, effortlessly, and scales enormously.
Or as 451 Research analyst Donnie Berkholz’s says: “The future? How about moving away from the need to worry about servers at all?”
That’s the world of public cloud. It’s why developers love AWS (and increasingly are bonding with Microsoft Azure and Google Cloud Platform). It’s why the public cloud keeps winning. ®