The real focus of this era is on managing multi-cluster, multi-distro Kubernetes across multiple environments at the same time.
Translated from Kubernetes Evolves from 'Distro' Bets to Choice at Scale, by Tenry Fu is the CEO and co-founder of Spectro Cloud. He has more than 20 years of experience in system software. After Cisco acquired his previous company, Cliqr, he led the architectural design of Cisco's multi-cloud management and private cloud solutions. When we launched our production Kubernetes research project three years ago, our goal was to track and validate a major shift in the cloud-native and Kubernetes worlds.
We've observed enterprise Kubernetes becoming more and more complex in many different ways. Organizations are running Kubernetes in more environments, deploying more clusters and software elements per stack, and using more Kubernetes distributions across their architectures.
In our latest research release, announced at Kubecon in Chicago, we found that 83% of organizations are using more than one Kubernetes distribution today. Of these, 59% used two to five;9% used more than 10 different distros!This study paints a picture of a rapidly growing and energetic experiment. It also describes a new era in which platform engineering teams are serving a wide variety of stakeholders – managing production clusters across multiple different environments at the same time.
There are many different variants of Kubernetes distributions, or "distros". They could be:
Tied to cloud services, such as the EKS-Distro (EKS-D) used by Amazon. forms part of the ecosystem of platform services used by developers, such as Red Hat OpenShift.
For very specific use cases — Edge is a great example — use lightweight distributions like the MicroK8s or K3S or our own PXK-E, built specifically for small devices. Built on security, with a reduced attack surface, FIPS encryption, and immutability.
Optimized for simplicity, it provides a fully packaged stack, from the operating system and Kubernetes distribution to add-ons and applications.
Embrace openness or come with a strong dependency or "perspective".
Whatever the driving force, the reality is that the more organizations expand their Kubernetes footprint, the more they need to treat distribution as part of the puzzle—not the puzzle itself.
Many individual maintainers, communities, and vendors have worked hard to innovate and build hundreds of Kubernetes distributions. That work is valuable. But there is no single distro for every use case.
Organizations need choice, and they need tools to operate multiple distributions at the same time—the same applies to the operating system and the many software integrations that make up a full production cluster. This era is truly about managing multi-cluster, multi-distro, multi-environment Kubernetes at scale.
Multifaceted realities are the reasons why we built the palette. No platform should have a strong point of view and lock you in. We are committed to solving real-world Kubernetes problems for our customers, regardless of the "style", quotient, and origin of their Kubernetes stack. We've found that our customers do need choice, open standards, and support for their existing environments.
At Kubecon Europe 2021, Ben Beeman of GE Healthcare mentioned a challenge we helped the company solve: he talked about a solution that "helped them deal with all the cloud-native chaos and multi-tier software" (cloud-native open source and commercial integration) and needed to manage the lifecycle of the entire stack, not only providing flexibility in terms of integration choices, but also maintaining consistency. It's not about the distro at all. It's about the bigger picture. While "lock" and "choose" may seem overused terms, they are important. They point to a fundamental need: there is no need to design an "exit strategy" for future investments. And this is especially important in today's ever-changing market.
The reality of today's IT landscape is this: promising open source software projects will fail, as we see with K3OS. **Chamber of Commerce Transforms, Goes Out of Business, or Gets Acquired – Recently, D2IQ, formerly Mesosphere, laid off employees and the future of its Konvoy distribution and DKP platform was in doubt.
Even giants aren't always safe. Not long ago, Broadcom completed its acquisition of Broadcom and laid off thousands of VMware employees, with VMware's partners reporting that outstanding sales and renewals are now pending;Analysts at Forrester advise VMware's customers to "be prepared for the shock." If you've already chosen a distribution, platform, or service from such a vendor, how can you mitigate your risk and move forward with confidence?
For those who are feeling uncertain about their previous options, want to protect their existing investments and extend them into the future, we're here to help.
We recently announced an offer for organizations that invest in D2IQ's DKP and Konvoy Kubernetes distributions to support their environments from start to finish with Palette and make the transition as smooth as possible, including providing accelerated training and attractive cost incentives. Earlier this year, we announced support for virtual machines, with enough functionality in the palette to support production workloads. Our own distributions (PXK and PXK-E) are 100% hardened, well-curated, and 100% Cloud Native Computing Foundation (CNCF) compliant. If you decide to leave Spectro Cloud, you can use them on any platform. Last but not least, our entire platform is built on proven open source technologies, such as the Cluster API and its pluggable providers for cluster lifecycle management, Kubevirt for virtual machine management, and Helm and Gitops for application lifecycle management. You'll get an all-in-one, enterprise-grade platform, but you're not locked into it. We've been working to help organizations that already use any style of Kubernetes in any environment, workload, and stack, acknowledging the need for choice when it comes to innovating and protecting existing investments in the future. Why don't we take on the challenge?