We can all agree that there are three IaaS leaders in the cloud: Amazon, Microsoft, and Google. Right?
It depends on how you look at it. Without question, Amazon maintains a very long lead -- and it’s the only player transparent enough to break out its cloud revenue in detail, with AWS Q2 sales up 81 percent year over year. Microsoft throws in Office 365 revenue, which is not strictly cloud money, but it’s clearly the No. 2 player and now sees the cloud as the most important part of its business going forward.
Yet where the heck is Google? The company doesn’t break out its cloud revenue at all. Microsoft and IBM aren't terribly transparent on this score, either, but according to last week's Q2 estimate from Synergy Research Group, each make more money from cloud infrastructure services than Google. After all, they have the leverage to push big existing enterprise customers to their clouds. By contrast, the vast majority of Google Cloud Platform customers appears to be cloud startups. The biggest success stories are SnapChat and Khan Academy, both built on the Google App Engine PaaS.
You can certainly argue that revenue doesn’t mean much yet because we’re still at an early stage in the cloud. What you should study now -- whether you’re a customer or simply handicapping this race -- is the viability of the strategy going forward. That’s where I wonder about Google even more.
Amazon Web Services' strategy is very similar to that of Amazon’s e-commerce business: more of everything -- more services (third party or native), more server types, more pricing options, more capacity, and on and on. Microsoft for its part is building a bridge between Windows Server/System Center and its Azure cloud to deliver a true hybrid cloud to customers, with very nice complement of open source support and advanced management features.
The latest story from Google, according to a June Computerworld interview with Brian Stevens, vice president of cloud platform, is that analytics will be central to Google Cloud going forward. Sounds reasonable, since Google Cloud Platform offers BigQuery and Dataflow. But Stevens' pitch is also reminiscent of the cloud strategy IBM revealed last fall, which came with a dog-and-pony show featuring big enterprise customers and an impressively broad set of analytics solutions and services.
Earlier this month, Google appeared to send another smoke signal when it joined the OpenStack consortium. InfoWorld’s Serdar Yegulalp speculated that Google might be flirting with a hybrid cloud play, in part because OpenStack has already incorporated Google’s Kubernetes container management system. Maybe so -- but what will be the extent of that interoperability, and how many OpenStack consultancies are really going to point to the Google Cloud?
Moves like these seems like hand-waving, and I find them frustrating. You’ve probably heard the cliché that Google lives in the future and occasionally sends back messages to the rest of us. When it comes to operating a vast cloud infrastructure at scale and crunching ungodly amounts of user data and metadata, who can touch Google? I keep waiting for the company to turn that expertise inside out and make more of it available in the form of advanced services to cloud customers.
For example, although Google was late to the IaaS game, it could conceivably leap ahead if it decided to become the “production container cloud” with all the security, performance, and orchestration that implies. After all, Google got the whole Docker thing going when it invented cgroups and helped ensure that became part of the Linux kernel. But so far, all we have is a beta version of Google Container Engine, which is largely Kubernetes on top of Google Compute Engine. Microsoft Azure also supports Kubernetes and Amazon offers its competing EC2 Container Service.
Meanwhile, little old Joyent got there with a secure Docker cloud before everyone else.
To be clear, according InfoWorld’s own reviews, Google has a fine cloud, easier to use than AWS, with excellent performance and aggressive pricing. It's also kept up pretty well in the feature wars. But there should be more to differentiate such a late entry: Google Compute Engine hit general availability only a year and a half ago, as opposed to 2010 for Microsoft Azure and 2006 for AWS. The Google App Engine PaaS has been generally available since 2010, but Gartner's 2015 Magic Quadrant puts it behind both Salesforce's and Microsoft's PaaS offerings.
As a friend of mine once said, Google is a vast computer science department supported by a giant advertising company. The free version of Google Apps may be firmly embedded in the enterprise, for example, but the company never talks about the money it may or may not be making from the pay version (though Google for Work may change that). The bottom line is that Google has had the luxury to launch all sorts of revenueless initiatives from mapping Mars to driverless cars, but the real business is all about search and crunching petabytes of data to match users with ads.
A recent shot across the bow suggests Google’s computer science department might shrink significantly. The company’s new CFO, Wall Street veteran Ruth Porat, has cited “managing the pace of expenses” as a key goal, sending (rumor has it) ripples of discomfort through Googlers working in noncritical areas. She might also ask: After years of offering SaaS, PaaS, and IaaS to customers, what exactly is the long term monetization strategy?
The enterprise cloud folks at Google better have a good answer. I’d hate to see all that incredible potential to deliver highly advanced cloud services to business customers come to naught.