Over the past few years, corporate thinking on whether to move data and applications to the cloud has evolved into figuring out which data and applications to move to which cloud.
The public cloud computing model, with its flexibility and price savings — compared to the cost of maintaining and updating internal data centers — has been proven. As one piece of evidence, note the recent estimates by market research firm Gartner that worldwide spending on public cloud computing will grow a healthy 23 percent this year to $332.3 billion, up from $270 billion in 2020.
There are a few reasons for this new acceptance of cloud, some relating to the fact that there are now several solid public cloud options that offer quality computing, networking, and storage options. And for most businesses, choice of providers is a good thing.
It means, for example, that a large company can leverage one vendor against another to get better pricing and other terms, especially when it comes to finding a cloud for brand-new built-for-the-cloud (aka “cloud native”) applications. More choice also means businesses can pick different clouds for different workloads or application types.
This is important because, while all clouds offer the same basic services, not all clouds are built the same. During the client-server era of computing, most businesses used several technology stacks: One or perhaps two vendors for desktop applications, another one or maybe two for server-side database and enterprise applications; and another one or two for networking technology.
Today, many will follow that bifurcated deployment model in the cloud era as well.
It’s also useful for technology buyers to have at least the threat of working with another vendor in their pocket if their legacy vendor starts adding costs or other constraints. Last year, a Bain & Co. survey of a few hundred CIOs found that fear of cloud lock-in is real; about two thirds of the respondents said they would like to spread their cloud usage across several providers for that reason alone.
For many corporate buyers, the advantages of using two or three clouds for the applications best suited to run them outweighs the added complexity that managing multiple vendors brings.
Steve Mullaney, CEO of Aviatrix, a company that helps large businesses simplify the management of multi-cloud deployment, partners with all the major cloud providers. He has seen interest in multi-cloud soar recently as deployment decisions rise to C-level executives who know they need to modernize operations across the board to stay competitive.
Companies are “mandating a level of agility from their IT teams that naturally leads to working with multiple public clouds,” he said. “If one cloud has the best AI, then use that cloud for AI. If another is best at database, then that’s the cloud for databases.”
As Hurwitz & Associates analyst Dan Kirsch said on an eWeek CIO Panel in July, companies have to start their plans by understanding where their data and workloads are now.
For businesses, the next step is to assess which of the public cloud offerings is best to run existing workloads, including the sort of complex finance, inventory, and manufacturing software that pre-dated cloud computing—but that also pay the bills—and put them on the cloud infrastructure that will run them optimally.
About the Author:
Barbara Darrow is Senior Director, Oracle Communications