By Eric Gregory, Mirantis

By many accounts, this is a multi-cloud age. Enterprise infrastructure tends to sprawl across a diverse architecture, often relying on a variety of cloud providers.

But the promise of multi-cloud was never sprawl. It wasn't dozens or hundreds of services duct-taped into an architectural patchwork. Multi-cloud was supposed to mean the exact opposite: consolidation under a single management plane. Unification, consistency, and ease.

So why do so many teams manage clouds individually, when their organizations supposedly have a multi-cloud strategy?

Halfway to Multi-Cloud

It's not hard to see why enterprises are keen to adopt a multi-cloud approach. Niche cloud services have proliferated to serve organizations' variegated needs, even if coordinating those (ever-accumulating) services can be a challenge. The big cloud vendors provide essential services, but no provider can be everything to everyone-and besides, enterprises have an obvious incentive to avoid the constraints and suboptimal pricing that come with vendor lock-in.

On the whole, enterprises seek out (or wind up with) a diversified architectural portfolio. And that brings us to the multi-cloud approach, which takes a landscape of varied services and providers as a given and seeks to make it all more manageable.

But the obstacles to multi-cloud are non-trivial. Teams crafting their own multi-cloud solution have to ensure stable and secure communication between the various parts of their architecture, which may or may not play nicely with one another out of the box. And in fast-moving industries, these time-consuming obstacles can stall a multi-cloud initiative halfway, leaving teams managing each of their cloud services individually.

The Perils of Individual Cloud Management

You might say, "Well, what's so bad about managing my clouds individually? That's more or less where I started." Perhaps! But unfortunately, "This is how we used to do things" isn't a very persuasive rationale in our industry.

As an enterprise grows, its architecture inevitably grows right along with it. Today, that means that it will come to encompass multiple private clouds, public clouds, on-prem infrastructure, and increasingly, resources at the edge. Managing each piece of your cloud puzzle individually-their integrations and APIs, their idiosyncratic dependencies and security requirements-is a job that will grow exponentially more difficult as the architecture grows, because you need to coordinate not only the new resources, but their relationship to the greater whole. As this work grows, it is liable to become untenable.

The fact is, competitive and growth-oriented enterprises typically can't afford to sit still. For them, multi-cloud isn't an experiment or a "nice-to-have" but a critical piece of an overall strategy that leverages the technical resources required to compete without getting lost in complexity and technical debt.

But if you can't afford to keep managing your clouds individually, that may put your organization in a difficult spot. For many enterprises in many industries, individual cloud management is simply too costly and unwieldy-even if they'd like to stick to the old way of doing things, it just isn't feasible. But we've seen that too many organizations get stuck halfway to multi-cloud...which may leave them thinking they've solved problems that remain very much alive.

So how can you get to true multi-cloud, especially if you're not sure you have the resources to build a solution yourself?

Closing the Distance

Multi-cloud strategies aren't limited to in-house systems. A multi-cloud operations management solution can leverage Kubernetes to create a single management layer, enabling your team to use the services you need, when and how you need them. Though this is a relatively new software model, it's swiftly proving necessary to help many organizations manage sprawling chains of infrastructure, while managing (and simplifying) newly critical technologies like Kubernetes.

We're in a moment where the writing is on the wall. The problem of service and infrastructure sprawl is real, but few enterprises can afford to limit their technological capacity with individually managed clouds. A true multi-cloud strategy is necessary.

However enterprises get there, the object for a multi-cloud strategy is an infrastructure-as-code solution that can encompass and coordinate every element of the architecture. On the way to that goal, it's essential that teams understand what they have accomplished, and how far they still have to go. Simply running multiple clouds individually isn't a strategy-it's an inefficiency to be solved. And we have to see a problem to solve it. So it's worth asking yourself: Are we really doing multi-cloud? And if not, how can we close the distance?

##

ABOUT THE AUTHOR

Eric Gregory is senior technical writer at Mirantis. He is an author and former computer science instructor who writes on Kubernetes, containers, open source, DevOps and more.