The trillion-dollar question that most of the IT industry is trying guess the answer to these days is what percentage of application workloads will ultimately wind up running on public clouds.
Analysts at RBC Capital Markets are forecasting that in its next quarter Amazon Web Services (AWS) will generate $3 billion in revenue, a 42 percent increase over the same period last year, with an operating profit of $522 million.
Microsoft in its most recent financial quarter — off a much smaller base of cloud business — reported a 116 percent increase in Microsoft Azure revenue for the three months ending September, while the Office 365 business grew 51 percent.
One of the primary reasons there is so much growth in the usage of public clouds is that they are simple to invoke. By making extensive use of application programming interfaces (APIs), the public cloud is essentially the preeminent software-defined IT environment of our day.
Movement to the public cloud
But as impressive as the growth numbers are, they still represent a fraction of a multi-trillion dollar IT industry. Nevertheless, it’s also becoming apparent that providers of IT infrastructure are becoming more concerned. Just about every one of them is starting to emphasize that the total cost of ownership (TCO) for stable application workloads over a multi-year period is substantially less on premise than it is in the public cloud.
The challenge they face, of course, is that the cost of acquiring cloud services is perceived to be much less risky. If an application development project winds up failing, the cloud service that supported it can simply be turned off. Just as importantly, anything running on a public cloud can be treated as an operating expense that can be fully deducted. Best of all, it only takes a few minutes to fire up the cloud service. When business and IT leaders talk about IT agility enabled by the cloud, they are approaching the discussion from both a technical and financial perspective.
IT infrastructure providers are clearly feeling that pressure. Every one of them has a software-defined infrastructure (SDI) initiative underway that promises to turn IT infrastructure deployed on premise into “infrastructure as code” that can be programmed using APIs. At the same time, IT infrastructure vendors are getting much more aggressive about leasing to counter the financial appeal of the cloud. For example, Dell EMC has a new payment option that converts servers and storage deployed on premise into an operating expense. Alcatel-Lucent Enterprise (ALE) is putting a leasing program that enable IT organizations to rent networking equipment all the way down to the port level.
Preparing for the future
Analysts at 451 Research clearly expect these and other moves to increase the appeal of deploying OpenStack on premise. The research firm is predicting that from a revenue perspective instances of OpenStack running on premise will overtake cloud instances of OpenStack by 2020.
This battle over workloads in the cloud is of keen interest to managed service providers (MSPs) because the cost structure associated with managing software-defined workloads is substantially different. Today the software-defined advantage is with the public cloud. But MSPs will need to support both deployment models.
Where their profits come from in the future will be heavily influenced by the degree to which those services can be programmatically delivered in a way that simultaneously reduces their costs while increasing the size of the total available market. In short, the faster application workloads become software-defined inside and outside of the cloud, the better off MSPs will become.
Photo: Lindsay Henwood