GeekWire Illustration / Bigstock Photo

Here’s a no-brainer: 2017 will be a big year for the cloud. Cloud computing is an innovation rivaling the advent of client-server, the PC or the internet, and it’s going to enjoy continued vigorous growth in the new year.

But private data centers aren’t going away. Though the essential balance of power within the public-cloud world won’t change much, competition may favor companies that best serve the organizations straddling private data centers and the public cloud — which is to say, most of them.

Here are some of the key cloud trends to watch this year.

Amazon Web Services CEO Andy Jassy. (Amazon Photo)

Revenue will rise sharply for the big public-cloud providers. Forrester is predicting that revenue from public-cloud services, combined with software as a service, will grow at a compounded annual rate of 22 percent between 2015 and 2020, reaching $236 billion. VC firm North Bridge estimated that public-cloud spending alone will grow to $522 billion in 2026, from $75 billion in 2015, a compound annual growth rate of 19 percent. It predicts platform as a service will see the greatest growth (33 percent CAGR), followed by software as a service (19 percent) and infrastructure as a service (18 percent).

Market leader Amazon Web Services is on track to generate at least $12 billion in annual revenue this year, with no signs of slowing down in 2017. Second-place Microsoft Azure in the most recent quarter enjoyed growth of greater than 100 percent year over year, and Azure’s annualized revenue run rate exceeded $13 billion in the most recent quarter. These growth rates may fluctuate by quarter but show no sign of flagging overall.

Azure chief Scott Guthrie speaks at Microsoft Ignite 2016. (GeekWire Photo / Kevin Lisota)

Other major companies, including HP and Cisco, are getting squeezed out by AWS and Azure. But Oracle’s Larry Ellison is making lots of noise as his company vows to challenge AWS and the other three principal contenders in the cloud: Microsoft, Google and IBM.

The cloud will be used for more and more production workloads. Cloud services were initially used mainly for development and testing, but they are coming into much greater use for production workloads, and that trend will continue. Production migrations have tended to range between one and 50 per organization, said cloud pundit David Linthicum, senior vice president at Cloud Technology Partners. In 2017, they’ll range into the hundreds, he predicted. But it will continue to be rare for companies to go all-in on the cloud, abandoning their own data centers.

The list of computing resources available “as a service” will continue to expand as public-cloud providers, especially AWS, expand their offerings, eating into business once held by their partners and by software developers. This will kill some partners and ISVs and cause resentment among others, but that won’t stop the trend.

Private data centers won’t go away.  A great majority of organizations will continue to straddle the line between data-center-based computing and the public cloud. There will continue a desire to keep data on-premises for legal and privacy reasons. Linthicum calls the private-public straddle “the pragmatic hybrid cloud.”

Based on a survey of 1,351 cloud companies, North Bridge predicts that current numbers — 47 percent doing the straddle, 30 percent strongly embracing the public cloud and 23 percent favoring a private cloud — will hold in the new year.

So if they’re smart, public-cloud companies will improve their pragmatic hybrid cloud offerings. Microsoft disappointed some prospective customers with hardware limitations it put on its Azure Stack, which promises the same controls over applications running on-premises and in the Azure cloud. It would do well to broaden that offering.

Machine learning, serverless computing, IoT and containers will all become increasingly important.

  • Machine learning, which is frequently performed in the cloud because of its huge data and computing requirements, is at the apex of Gartner’s hype curve, meaning it’s at the “peak of inflated expectations.” Still, expect to see progress in developing image recognition and voice interfaces.
  • Serverless computing, such as AWS’s Lambda, Azure Functions and IBM’s OpenWhisk, execute code snippets automatically upon request. It’s still a new offering, but it will grow in importance because it lightens the load on developers. The fact that it’s a misnomer (of course there’s a server doing the computing; it’s just that the user needn’t bother about its details) won’t impede its growth.
  • IoT, still ascending the Gartner hype curve toward that lofty peak, in 2017 may enjoy the year everyone’s been predicting, with millions of sensors and other devices coming online. Or not.
  • There may be controversy over whose container-management product will prevail, after AWS introduced Blox this fall in an apparent bid to compete with market-leading Kubernetes as well as Mesos, Docker Swarm and several other products.

The voice interface will become a major focus. The ability of developers to access voice-interaction cloud services will greatly accelerate their use in apps. Amazon holds a clear lead here with its Alexa, having made voice-activated “skills” development kits and APIs available, which has led to a skills library no other vendor has. Apple built up the App Store in the same way, propelling the smartphone industry, but it failed to do the same thing with Siri. On the other hand, Amazon hasn’t incorporated Alexa into a smartphone (yet).

F5 CTO Ryan Kearny

Multi-cloud efforts may persist. F5 Networks CTO Ryan Kearny predicted that the use of multiple cloud services, not just one, “will become the new normal. Enterprises will continue to utilize multi-cloud strategies and will increasingly look to avoid public cloud lock-in. They will seek to regain leverage over cloud providers, moving toward a model where they can ‘lift and shift’ to select and combine optimal services from multiple providers.”

Matt McIlwain, managing director at Seattle-based VC firm Madrona Venture Group, agrees. “Many companies will use a combination of clouds, both infrastructure services and applications, across a mix of workloads with AWS still dominating the market as Azure and Google Compute gain momentum,” he said.

North Bridge’s numbers indicate concern over lock-in ranks second among cloud adopters (30 percent), behind security (38.6 percent) and ahead of privacy, complexity and regulatory and compliance issues.

“Belief that one cloud vendor can meet all needs is simply out of touch with reality and smacks of vendor hubris,” said Julia White, VP for Azure and security marketing, in a post about the biggest cloud trends of the past year.

Julia White, VP for Azure and security marketing, speaks at Microsoft Ignite 2016 (GeekWire Photo / Kevin Lisota)

But I’m not so sure. The counter-argument is that some degree of lock-in is inevitable, no matter which application, operating system or cloud provider is used. The best that can be hoped for may be to minimize the pain of moving. In order to get most out of any application or cloud service, it’s essential to take advantage of all the features it offers, which means increasing customization and decreasing portability. I think IT folks realize this and may be opting to take a chance on a single cloud provider, pouring their efforts into it and taking the chance it will succeed. The demise of Cisco’s InterCloud offering, which sought (among other things) to offer users linkages among multiple public and private clouds, may be evidence there wasn’t enough interest in doing that to support the product.

Hear cloud tech reporter Dan Richman discuss AWS and Azure as part of GeekWire’s year-end podcasts about Microsoft and Amazon.

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline

Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.