The Cloud Changes Things, But Not Everything
October 8, 2018 Timothy Prickett Morgan
One of the slowest things in the world is the pace at which risk averse enterprises change their information technology. We have been talking about computing on demand and utility computing for more than two decades, and Amazon Web Services made the cloud a real thing with the launch of its Elastic Compute Cloud, or EC2, service back in March 2006. Fast forward 12 years, and EC2 is the foundation of a vast business that made $21.2 billion in sales in the past 12 months, and it has utterly transformed the way people want to configure IT infrastructure so it can be consumed in a cloudy fashion.
But that does not mean that everything is going to move to the cloud, which really means the public cloud, and sometimes people forget that.
That said, the public clouds are still growing at a pretty healthy clip even if the growth has slowed, as it inevitably does with any new technology that is widely adopted. According to the latest market statistics from Gartner, sales of cloudy infrastructure, platform, and software services are projected to hit $175.8 billion in 2018, up 21 percent over the $145.3 billion in sales of all kinds of cloudy stuff in 2017. Looking out further, Gartner is projecting that growth will slow to 17.3 percent in 2019, with sales hitting $206.2 billion. This is against a backdrop of core IT spending of $1.3 trillion. In round numbers, public cloud services in their many guises comprise only 15 percent of the budget.
Here is the breakdown in sales for 2017 and projections from 2018 through 2021 that Gartner provided across these categories, which is illustrative:
For one thing, most of the cloudy stuff people buy are online versions of applications, what buying software delivered over remote infrastructure through what we used to call Application Service Providers back in the day. You know, before Salesforce.com was a name anyone knew. In many cases, these cloud application providers (SaaS suppliers) run their own infrastructure in co-location facilities, so it is arguable that this should be included in any discussion of public cloud sales. It’s cloud, to be sure, and yes, many of the SaaS suppliers do indeed run their applications on top of one of the major public clouds like Amazon Web Services, Microsoft Azure, or Google Cloud Platform. What is interesting is that the growth for public cloud services is highest in infrastructure as a service, or IaaS, which many had thought would be less interesting over the long haul when people shifted to platform services with a higher level of abstraction that a mix of compute, storage, and networking services that IaaS is comprised of. But of course, these PaaS and SaaS services depend on and drive IaaS, and in many cases, you have to buy the IaaS yourself to suit your own scale and pay separately for the PaaS and SaaS services. I guess we all forgot the accounting would be done that way.
The other interesting thing to note here is that all of these five segments of public cloud services as segmented by Gartner are growing nicely; some more than others, of course, but the overall market will double between 2017 and 2021. It is hard to predict what will happen beyond 2021, but the growth rates are slowing, so it will take more than five years to double again – unless there is some outside force that compels people to stop investing in their own hardware and software and run a datacenter. A big recession can have that effect, and it has been more than 10 years and many say that we are overdue for some kind of market correction. I tend to agree with that sentiment, but I see no reason to rush into a recession.
In a separate report, Gartner put out another interest set of statistics, showing the market penetration of cloud services across four categories (security services, for some reason, were not part of this set of numbers). Take a look at how the public cloud percentages are growing over time:
The share of cloudy sales of business process outsourcing and application software of the total spending on these items (including on-premises software plus the cloudy stuff) are starting to hit their saturation point, which should mean a wave of consolidation as these companies reach their natural levels providing these services. System infrastructure (compute, storage, and networking) will grow steadily and seems on track to be about a fifth to a quarter of the market by maybe 2024 or so, if all trends persist, and infrastructure software is rising to be about a fifth of the total addressable market.
This is nowhere near 100 percent. While some customers are all-in on the public cloud, this does not mean that all customer budgets will be into the public cloud even if all customers do some public cloud as well as some on-premises stuff.