DEVOPSdigest

For the past couple years, organizations have rapidly adopted Kubernetes for container orchestration. The technology's upshot is clear: Kubernetes makes it much easier to manage or deploy containerized workloads in both on-premises and public cloud settings. But even though Kubernetes has become a red-hot trend among tech media and influencers, few organizations are actually deploying Kubernetes in production and at scale.

VMware's recent report on the topic paints the picture vividly. This survey didn't even look at all enterprises — it narrowed its focus to only look at organizations that have currently deployed Kubernetes. The data found that:

■ The majority of these companies (60%) ran less than half of their container-based workloads on Kubernetes.

■ 57% were running less than ten Kubernetes clusters.

■ Just one fifth of organizations using Kubernetes in production had 50 or more clusters deployed.

Why are organizations slow to deploy Kubernetes? After all, Kubernetes has deservedly been heralded for its agility and ability to drive greater business value. Operational difficulties are the biggest issue. Simply put, Kubernetes is very challenging to deploy, manage and run at scale. It's easy to manage ten or fewer clusters in a test environment but it is far more complex to run hundreds or thousands of Kubernetes clusters.

Kubernetes is even harder to manage and deploy in multi-cloud settings. The above survey found that 64% of organizations were using Kubernetes on-premises, 42% were using it in one public cloud platform, and just 31% were using it across multiple public clouds. IT typically experiments with Kubernetes in an on-prem or single public cloud setting. But when they try to expand those deployments to multiple clouds — or even grow those deployments in the traditional on-prem or public cloud setting — things quickly go south.

Adopting any major new technology requires IT to gain approval from their higher-ups and sometimes even the C-level. Kubernetes is no exception, and in fact it may be harder to get buy-in than other technologies. Kubernetes is a core infrastructure technology, impacting many different teams, and as a result the CIO must typically give approval before it can be deployed in full production, organization wide.

VMware's report demonstrates this reality: According to the data, in 83% of organizations, multiple teams are involved in selecting a Kubernetes distribution. Compare that to microservices: microservices are generally only employed by and impact developers, so it doesn't usually require approval from the CIO.

Every new technology has a learning curve, but Kubernetes is particularly steep. Again, because it is a core infrastructure-level technology, it requires mastering a lot of difficult new skills. Most organizations still rely on virtual machines to deploy and run infrastructure and their associated applications. While more IT staff have had time over the past five or so years to get their feet wet with container technology broadly, Kubernetes has not been in vogue for so long. And Kubernetes is more complicated and difficult to learn than general containers.

However, Kubernetes offers major benefits. Enterprises should seriously evaluate the technology to take advantage of these benefits. That said, unless you're an organization with advanced technical expertise, such as a software company or financial services firm, you shouldn't try to manually deploy Kubernetes.

So what are you left with? The best route is to try out a pre-made Kubernetes cloud. Luckily, there are a variety of cloud-based Kubernetes distributions out there.

Which should you choose? That depends on your unique business needs. But you should first narrow it down to one that can scale easily, has been designed specifically for Kubernetes and is centrally orchestrated.

There's a reason that organizations are choosing Kubernetes by a wide margin over other container orchestration options. When it comes to managing and deploying infrastructure and applications, Kubernetes delivers agility, speed and major resource efficiency. However, you will only realize Kubernetes' potential if you can wield it correctly.