Kubernetes Celebrates Its 10th Anniversary: How Is It Doing In The Data Center?

By Drew Robb

Kubernetes turned 10 years old in 2024. The trajectory of its rise to prominence has been phenomenal. This open-source project is regarded by many as the de facto standard for container orchestration. It has even managed to seduce developers who now routinely use it to launch new cloud applications.

“The broad compatibility of Kubernetes is making it the operating system of cloud native computing,” said Omdia Michael Azoff, Chief Analyst, Cloud & Data Center Practice at Omdia.

For example, data center managers are using Kubernetes to create virtual clusters using an open-source tool known as Loft Labs vCluster. It can run inside namespaces and enable multi-tenancy of Kubernetes environments. Its popularity is down to the fact that physical clusters within the data center consume too many resources for monitoring, logging, and security, according to Azoff. A single Kubernetes cluster simplifies operations and reduces resource usage as a great many virtual clusters can be launched that share underlying components.

“More enterprises are adopting vCluster to reduce their cloud bills,” said Azoff.

AFCOM’s 2024 State of the Data Center report surveyed data center managers to find out what kind of workloads they are migrating to the cloud. 26% named custom applications, such as DevOps processes and those built with Kubernetes. This number is expected to rise sharply over the next couple of years.

 Business Benefits of Kubernetes

Cost is a big reason why data centers favor Kubernetes. But beyond containing cloud costs, there are other drivers. A survey of more than 500 Kubernetes users, each with at least four years of experience with it, found that 64% said it made them and their staff more efficient. 60% said developers became more productive when they utilized containers. 37% said it helped the data center to be viewed more as a revenue driver rather than just a cost center.

The study also discovered that 80% of new applications are destined to be built in cloud-native platforms such as Kubernetes over the next five years. This doesn’t signal the end of the data center, however. Many applications will continue to run in house and many data centers will operate in a hybrid cloud-on-prem arrangement.

In any case, cloud applications require massive data centers from which to deliver their services. Expect the hyperscalers, tech giants, and big colocation providers to open many more data centers. Further, some smaller, private data centers will flourish due to being located in an attractive metro area and will pick up business from the major cloud providers who may be unable to obtain real estate or permits in some cities.

According to the survey, organizations prefer the flexibility of deploying applications and data in hybrid cloud environments. 86% of already run their cloud native applications across public and private cloud environments. Large enterprises are far more likely to have a large on-prem footprint. Data centers will survive Kubernetes, just as they did the cloud. But they will need to adapt.

The AI boom, too, is playing into the hands of Kubernetes expansion. With so much attention on adding AI capabilities to existing applications, and in the development of large language models (LLMs), data centers are finding Kubernetes to be a handy way to simplify development and bring AI apps to market faster.

“Kubernetes is no longer an emerging technology,” said Kalyan Ramanathan, Vice President of Marketing at Kubernetes storage and data services provider Portworx. “Organizations now trust it to build applications like real-time analytics and AI/ML workloads.”

But it isn’t only AI. Kubernetes makes life easier for application developers. Two thirds of survey respondents named faster deployment as a primary reason for building applications in Kubernetes, followed by reduced infrastructure costs at 54%. This is particularly the case for data-intensive workloads. 97% of those surveyed said they prefer to run data-intensive workloads on cloud native platforms, 72% plan to build new databases on Kubernetes, and 67% for new analytics applications. 

The Rise and Fall of VMware

VMware has been the darling of the data center for nearly two decades. It grew in popularity as it enabled data centers to escape the burden of having each application running on its own physical server. Instead, VMware opened the door to massive amounts of server consolidation. IT could create virtual machines (VMs) and run a great many of them inside one physical box.

But virtualization is regarded by some as old hat – particularly by a new generation of developers and data center personnel who are well-versed in Kubernetes. Containerization, it appears, is taking over from virtualization as the tool of choice for the data center.

Since its inception in 1998, VMware has changed hands several times. First EMC bought it, then it belonged to Dell, and now it is part of Broadcom. Uncertainty over what Broadcom plans to do with it, as well as a great deal of rumor-mongering and hysteria generated by the trade press (and perhaps some competitors), has caused some to flee VMware. Kubernetes has picked up a good deal of that business.

Some say, Broadcom dithered over announcement of its VMware plans. But don’t write off VMware or Broadcom. Time will tell how well they deal with it. They are quite capable of bringing about a resurgence in interest in VMware, depending on how they make the platform available to others.

In the meantime, the bleeding continues. The Portworx survey found that 65% of organizations planned to migrate at least some of their VM workloads to Kubernetes within the next two years. 80% admitted that that the Broadcom acquisition made them nervous about the future of existing VM workloads.79% added that the primary drivers for migration were scalability, flexibility, operational simplicity, and cost. A handful of organizations stated that they intend to switch from VMs to containers for all new applications.

“The Broadcom acquisition gave a definite boost to the Kubernetes conversation,” said Ramanathan. “Most new applications will be built on Kubernetes going forward as it facilitates hybrid cloud environments and containers can be used for AWS, Azure and on prem simultaneously.”

Data Centers Must Prepare

Like virtualization before it, Kubernetes and containerization are very specific technologies that require know-how and skilled deployment. Those wishing to quickly containerize everything are likely to run into challenges such as maintaining high availability, guaranteeing disaster recovery (DR), achieving platform consistency, and figuring out how they should go about pairing containers with high-performance storage. Some users indicated that existing Kubernetes data management systems lacked some capabilities related to availability, DR, and performance. Thus, data centers should invest in personnel training as well as bringing in outside experts when needed to help them navigate the pitfalls of Kubernetes implementation. Storage, in particular, can be a thorny issue. Portworx is one option.

“Portworx offers a flexible storage solution that automates storage across the application lifecycle, provides data protection and data resiliency, and unifies modern applications across hybrid and multi cloud environments,” said Ramanathan. “It provides a unified platform for VM and container environments, allowing organizations to realize the benefits of scalability, flexibility, and operational simplicity in their cloud native environments.”

With Kubernetes heading into its second decade, it seems likely that it will be part of the data center skills portfolio for some time to come. Data center managers who embrace it will be in a better position to navigate server the needs of users who want to host part of their applications in the cloud and part of it on-prem.