The first post-Covid KubeCon/CloudNativeCon Europe 2023 in Amsterdam this week did not disappoint, as a sold-out crowd of more than 10,000 professionals turned up in person to learn about all things Kubernetes.

The signs of maturity of this massive open-source ecosystem were everywhere. Some 58% of attendees were first-timers, illustrating the explosive interest in Kubernetes among enterprises. Compared with previous KubeCons, the buzz centered more on running Kubernetes in full production than on assembling the bits and pieces to make it work.

And most important of all, it wasn’t an infrastructure-only crowd, as applications finally earned their place in the Kubernetes sun.

The flipside of a maturing market, however, can be a slowdown in innovation. My goal at the show: Take a look at the most interesting startup exhibitors to see if innovation was indeed on the wane. My conclusion: There’s still plenty of gas in the Kubernetes ecosystem innovation engine.

Generative AI: less than you might expect

With the recent explosion of interest in ChatGPT and generative AI broadly, you might think KubeCon vendors would be rushing out new products that included it.

To be sure, generative AI is on many product roadmaps, but the only vendor I spoke with that had already brought its generative AI offering to market was Roost.ai, officially Zettabytes Inc.

Roost.ai provides a cloud-native, end-to-end testing product that leverages generative AI to generate test cases. Upon a pull request that signals a code change, Roost.ai launches an ephemeral environment, runs test cases within that environment, and then deletes the environment when finished.

Roost.ai works with its customers to hammer out test data specific to their use cases – a necessary manual step to support the generative AI capability.

Innovation on Day 2, aka full production

As enterprises scale up their Kubernetes deployments from handfuls of clusters to fleets of clusters and beyond, they often experience the growing pains that result from conquering new levels of scalability.

Within clusters, Kubernetes handles the rapid scale-up and scale-down of containers and pods, but the clusters themselves lack such ephemerality, making it difficult to manage dynamic requirements at massive scale. Fleet management – the management of large numbers or “fleets” of clusters – is an ongoing challenge.

Elotl Inc. addresses this problem by delivering “just in time” clusters that can scale up and down the way that containers and pods do. With Elotl, enterprises can commoditize their clusters, rather than configuring them individually.

Elotl then federates these ephemeral clusters into a “supercluster” that might be a set of fleets of clusters – and then exposes the supercluster to developers as though it were a single cluster.

Championing open-source projects while building enterprise offerings

As the largest open-source conference in the world, KubeCon is chock-full of vendors that have built their business models on top of one open-source project or another.

Two of the standouts: Diagrid Inc. and Akuity Inc., which are championing open-source projects Dapr and Argo, respectively.

The Distributed Application Runtime, or Dapr, provides application programming interfaces to help developers deal with several niggling Kubernetes challenges, including state management, publish/subscribe, secrets management and a host of other capabilities. By exposing these capabilities as APIs, developers can call upon them as necessary when building applications.

Diagrid brings Dapr to the enterprise with a fully managed Dapr service on Kubernetes, including an open-source control plane that helps organizations set up and run Dapr at scale. Diagrid also extends Dapr’s integration capabilities, making it suitable for multicloud deployments.

Argo is a popular suite of open-source tools for GitOps, including continuous integration and delivery or CI/CD, running DevOps workflows and managing clusters.

As with Diagrid, Akuity offers a managed software-as-a-service version of Argo CD, as well as a platform for scaling Argo. The Akuity platform extends Argo to multiple clusters by placing an agent within each cluster that connects to the Akuity portal and platform. The platform, in turn, leverages the Argo control plane.

The common thread connecting Diagrid and Akuity is how they are extending their respective open-source projects to enterprises that want to run them at scale.

Platform engineering: even hotter than generative AI

Among the startups I interviewed, platform engineering is the hands down innovation center of gravity.

Platform engineering seeks to balance two conflicting priorities: the self-service capabilities that give developers the flexibility they require, and support of the automated operations and rigor necessary to meet the organization’s application development and deployment needs.

Platform engineers, in turn, enable product and development groups to move faster via an internal developer platform or IDP that provides a curated set of tools, capabilities, policies and processes as a single, coordinated framework architected for self-service.

Several startups at KubeCon tackled various aspects of the platform engineering story. Humanitec GmbH provides what is basically a platform for building IDPs – a “platform orchestrator” that provides a configuration engine that platform engineers can use to customize their own IDPs.

As part of its platform capabilities, Humanitec offers dynamic configuration management, automatically generating versions of configuration files that organizations can secure, deploy and roll back if necessary.

Competing with Humanitec is Getport.io Ltd. aka Port, which like Humanitec also helps organizations build self-service IDPs. Central to Port is its software catalog that supports its developer portal. The catalog keeps track of microservices, CI/CD flows, clusters, environments, pipelines, deployments and anything else important to the software lifecycle.

Port also offers workflow automation that supports both service-to-service interactions as well as interactions on its portal interface. In addition, it provides scorecards that help organizations maintain their own standards of production readiness, security, DORA metrics, and other priorities.

Rounding out the platform engineering contingent are Red Kubes B.V., mogenius GmbH and Devtron Inc.

Red Kubes offers a preconfigured suite of open-source projects covering CI/CD, observability, logging, backup and other essential IDP capabilities. This suite is fully integrated and installable as a unit, saving DevOps engineers many months of headache.

Mogenius provides a Kubernetes-specific hosted platform for spinning up self-service environments that include the entire DevOps toolchain. The mogenius enterprise platform is cloud-agnostic, so customers can spin up mogenius environments either on-premises or in the cloud of their choice.

Devtron offers 80% of the DevOps toolchain as a preconfigured set of open-source tools, leaving the remaining 20% for platform engineers to specify and configure. The Devtron approach thus balances the preconfigured self-service benefits of an IDP with its customers’ requirements to customize their DevOps toolchains.

Furthermore, unlike other open-source offerings, Devtron includes SecOps, authentication and authorization, and DORA metrics in its free version (its enterprise offering is still in the works).

Final words

Typically, maturation and innovation are inversely correlated, as larger vendors become less able to deliver new, disruptive products.

The open-source world, in contrast, is better able to achieve both these priorities simultaneously. Even though some individual projects thrive while others fail, the overall open-source community continues to drive innovation, even as projects continue to mature.

It’s no wonder, therefore, that KubeCon is firing on all cylinders. I’m excited to see what else is in store for this conference in the coming months and years.

Jason Bloomberg is founder and president of Intellyx, which advises business leaders and technology vendors on their digital transformation strategies. He wrote this article for SiliconANGLE. None of the organizations mentioned in this article is an Intellyx customer. The CNCF covered the author’s travel expenses to KubeCon, a standard industry practice.

Photo: Linux Foundation