Storage Field Day Session #2: MinIO and Kubernetes


Kubernetes seems to be the answer to a lot of questions these days. It should be—it is the dominant way to deliver compute, networking and storage to applications. Its dominance is surprising in a tech world where there is generally room for several players. This was a winner take all market and Kubernetes took it all.

As we note early on in this session, Kubernetes became dominant for a few simple reasons and about a thousand complex ones. The simple ones are as follows:

  1. It brought billions of dollars of intellectual property from the web-scalers to the rest of the world. The technologies that built web-scale also built Kubernetes and were made available to anyone with a hunger for learning.

    You will have to watch the video to get the alliteration magic that happened.
  2. The second reason for the meteoric ascent of Kubernetes was the fact that it was so effective at abstracting the differences between environments—public, private, and edge— that they effectively vanished. On some level, the rise of Kubernetes has slowed the growth of the public cloud and contributed to the growth of the private cloud. It sounds odd, but it is true.

    Kubernetes vaporized private cloud shortcomings. Infrastructure took a backseat to more important tasks such as management, scalability and continuity—granted, that only happened if you could actually take advantage of it. We will probably say it a dozen times in the SFD recaps, but Kubernetes doesn’t do you much good if you are an appliance.
  3. Finally, the rapid rise of Kubernetes is a function of the fact that it is open source. Yes, it allowed for adoption, but also for the creation of entire companies around features, functions and the platform itself. The ecosystem became self-sustaining in a technology nanosecond.

To really appreciate all this, you need to watch AB’s section. It is a masterclass in what it means to be cloud native and how to get the most out of Kubernetes.

When you look at the market objectively, you come away with the realization that MinIO is actually unique. We are cloud native. Our storage can be orchestrated natively with Kubernetes. We are based on RESTful APIs, understand S3 at an atomic level and were built on microservices leveraging Go. We are the performance leader, meaning there are more applications running on MinIO because we support a broader set of workloads.

This is why containerization and orchestration is our default deployment configuration. 62% of our instances are containerized. 43% of those are managed via Kubernetes. We would posit that the rest of the private cloud object storage companies combined don’t get above 10% of their deployments. Our numbers for containerization and orchestration match what you would expect to find in the public cloud.

We really enjoyed putting Slide 12 together. The fact that we are influencing the orbits of Pure, EMC and NetApp tells you something about the gravitational pull of MinIO. It is growing by the week.

Keep in mind, a year ago, no one talked about high performance object storage—it’s table stakes for the marketing vernacular at this point. It doesn’t mean other companies are actually high performance object storage—try and find any published data outside of us and OpenIO (now part of OVH)—but it does mean customers are demanding it.

Why? Because object storage isn’t about archival anymore—it's about primary storage.

We spent some time on an additional point of evidence: VMware. The private cloud juggernaut selected MinIO as a design partner for the launch of their data persistence platform. This speaks to our reach. We have a ton of VMware content on the site so I won’t spend too much time here, but I think it bears mentioning that when VMware chose to demo their data persistence platform they did so entirely on MinIO. There is a reason for that.

So the summary for Kubernetes + MinIO section:

  1. Kubernetes is winning. That means object storage is winning. Block and File are legacy architectures on the cloud. Kubernetes is accelerating their demise.
  2. The rest of the private cloud vendor community is in full reaction mode. This is a terrible place to be when the world is moving as quickly as it is. It is fraught with risk.
  3. The cloud-native ecosystem is consolidating around MinIO. We see it everywhere, from VMware to Kasten before they were acquired. Cloud native apps want cloud native storage and MinIO is the only one offering this outside of AWS, Google and Azure.

Here is the video...