For most companies, multicloud and hybrid cloud environments aren’t a choice. They’re just what happens as those companies evolve. So while 451 Research projects that 69 percent of organizations expect to run a multicloud environment by 2019, the reality is that 100 percent are already there. That’s because any company that has set up in the cloud is almost certainly already running in more than one. The reason? Developers.
Multicloud by the grace of developers
Oh, yes, I know that CIOs want to claim credit for having a strategy around hybrid cloud (running public and private cloud workloads) and multicloud (running workloads on more than one public cloud), but these things just happen in a world that can no longer be command-and-controlled by the C-suite.
This doesn’t mean, of course, that there’s zero control of cloud adoption. There’s just not as much as in the past.
For example, as Rishidot analyst Krishnan Subramanian has highlighted, “Multicloud as a [high availability] use case is meaningless, but multicloud as a way to avoid shadow IT (giving developers the cloud services they want) is a critical strategy for enterprises.” As such, he continues, “Going forward, most enterprises will have a multicloud strategy.”
Catch that? Enterprises can’t stop developers from embracing services that make their jobs easier, but they can evolve to offer many of those services on private clouds, not to mention adding official support for public clouds that have services unavailable on the enterprise’s default choice.
For example, at my company Adobe Systems, we’ve built Adobe I/O Runtime to offer developers a way to run their code on the Adobe Cloud Platform and extend it in ways that fit their needs. While directed at third-party developers, it’s also used by our internal engineering teams to build out integrations with our own products and third-party offerings.
Built on the open source OpenWhisk project, it gives us a way to embrace the benefits of an AWS Lambda on our terms. (No, it doesn’t replace Lambda or similar services from other clouds, and that’s not the point. The point is to complement those third-party cloud services while giving our internal and external developers ways to extend our platform.)
The one constant in all this is that developers aren’t going to slow down for IT to provision a server or service. Enterprises that want to keep their developers close must build out and/or sanction the services developers need to be productive.
Either way, it’s going to result in multiple clouds being adopted, with a blend of private and public cloud infrastructure deployed. That’s not strategy. It’s just what happens when developers are in charge. Sure, CIOs will try to appear to be in control, but they’re not. They’re simply not.
The CIO is the last to know
If CIOs were in charge, we’d see the usual legacy vendors—Oracle, IBM, HPE—with thriving cloud businesses. But we’re seeing the opposite. Oracle has brayed loudest about its cloud bona fides, but it has seen its already tepid growth slow further in its latest earnings. Quite simply: The tech-sales-over-golf era is over.
In this new era of developer power, the currency that counts comes down to rich APIs and equally rich documentation and technical content. In a recent survey of more than 16,000 developers, Slashdata asked developers what they value most and compared it to where companies spend the most time and resources in an effort to attract those developers.
The top need for developers? No, not a T-shirt at a trade show (despite companies spending a significant chunk of their budgets on this), and not even hackathons. No, the single-most popular request by developers is for documentation and other technical content.
It makes sense that developers want information to get their jobs done on their time frame, with as little human interaction as possible. It’s the same self-service model that has driven open source software and cloud computing.
In fact, in that same survey, every interaction with people was rated well below any content developers could access on their own. This isn’t because developers are antisocial; it’s because they’re busy.
So, yes to multicloud and yes to hybrid. But not because there’s some grand plan to limit lockin or deliver high availability. Rather, it’s because developers are in charge, and they’re going to get their jobs done however and wherever it makes sense.
Enterprises can facilitate this by bringing some of those services home, but developers will only use them if they’re actually better than what they can get outside the four walls of their firewall.