It’s no secret that the use of Kubernetes has been exploding for some time. VMware’s “State of Kubernetes 2020 Report” surveyed 247 individuals actively using Kubernetes. Most have large development teams, with 24% having more than 2,500 developers.
The report indicates that development teams are still the primary decision-makers, at 38% for the use of Kubernetes technologies. Also, 57% are running fewer than 10 Kubernetes clusters, and 59% of respondents are running Kubernetes in production. Finally, only 20% have more than 50 Kubernetes clusters.
What you can glean from this report is it’s still early days for Kubernetes. Although Kubernetes is in wide use, the projects have leveraged it in measured ways. Moreover, a great many Kubernetes deployments are not in the cloud, with the report showing on-premises Kubernetes deployments at 60% and cloud deployments at 42%. This surprised me.
Since we’re more than five years in with Kubernetes, it’s clear that it’s more than a quick tech trend. It’s become a platform unto itself, with a robust ecosystem of technology, and a training foundation that cranks out Kube developers and architects monthly.
So, how do we take things to the next level?
Although we often take baby steps with any technology, we’re at a point with Kubernetes where we should be looking for more aggressive use cases and seeing if Kubernetes can remove the stifling complexity from multicloud. This has been the largest hindrance to its growth if you ask me.
Here are a few problem patterns where Kubernetes should be deployed:
- The big one: cloud federations of multiple cloud brands. This means using a federated Kubernetes deployment (cluster running above heterogenous clouds) across more than a single cloud brand, such as AWS and Google, or Google and Microsoft, to create a platform layer above the specific native-cloud services. This gives developers the ability to build clusters once, as well as containers, and run them across different public clouds whose native services become an abstracted commodity.
- Replacing the transactional applications currently in production with Kubernetes clusters. That should provide better scalability to overall better economics—certainly cheaper if leveraged from a public cloud.
- Access to common services and microservices to finally get to a state of reuse among applications. We’ve promoted the model of service reuse for years, but for most enterprises, it’s been elusive. Kubernetes provides a platform to build and deploy these services and the capability of scaling these services.
All these scenarios come with some cost and risk—there is no getting around that. However, companies that move in more aggressive and innovative directions with Kubernetes will perhaps take their business to the next level: better IT systems to support and expand business, and better customer experiences to take them across the finish line.