New applications and technical platforms are being developed every day. That helps boost industry competitiveness, but many of these software-based solutions don’t last. The Kubernetes container-orchestration system is one platform that is both surviving and thriving.

Despite the platform’s complexity, Kubernetes remains popular among both technical experts and novices and has done well where many similar systems haven’t, according to Janet Kuo (pictured, right), a software engineer at Google LLC, and Bryan Liles (left), a senior staff engineer at VMware Inc.

Kuo and Liles spoke with Stu Miniman (@stu), co-host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, and guest host Corey Quinn (@QuinnyPig), during the KubeCon + CloudNativeCon event this week in Barcelona. They discussed what Kubernetes can do for both people that are and aren’t information technology experts, why Kubernetes has done so well when similar platforms have failed, and why the Tiller in-cluster component’s removal from Helm is a big deal (see the full interview with transcript here). The interview was condensed for clarity, but the full interview is below. (* Disclosure below.)

Quinn: Why has Kubernetes started thriving where other projects had failed trying to do the same things?

Kuo: I think the biggest thing about Kubernetes is the really strong community and ecosystem. And, also, Kubernetes has the extensibility for you to build on top. We’ve seen people building frameworks and different open-source platforms on top of Kubernetes. So other people can use higher layers, stacks on top of Kubernetes. It’s very powerful, because everyone can just use the Kubernetes-style API to manage almost everything.

Liles: The secret of Kubernetes … is that it’s a platform for creating platforms, so Kubernetes really is almost built on itself. Kubernetes extends itself with the same semantics that lets users extend it. Janet talked about the CRD’s custom resource definitions. It’s the same mechanism that Kubernetes uses to add new features, so whenever you’re using these mechanisms, you’re using Kubernetes. This is the tool kit for creating your solutions. Kubernetes is not an endpoint; it’s a journey.

Miniman: Talk about the people who aren’t quite ready to step onto the Kubernetes ship.

Kuo: One of the sub-projects that I own is the workloads APIs. So, basically, you’re using Kubernetes declarative APIs to run different types of applications — workloads. For those who are just getting started, maybe you just start with a stateless workload — that’s the easiest one. And then for people who are looking to contribute more, I encourage you to start with small fixes. Maybe fix some documents or do some small PRs and build your reputation from there. Start from the small contributions, and then build all the way up.

Quinn: One of the big announcements was that Tiller is no longer going to be a part of Helm. What is Tiller, and why is it perceived as less than awesome?

Liles: When Helm was introduced, it had this thing called Tiller, and what Tiller did was it ran a cluster-wide level to make sure it could coordinate software being installed along with Kubernetes-named spaces or groups. What happened is that this was the best vector for security problems. You had this root-level piece of software running, and people were figuring out ways to get around it, and it was a big security hole.

Now, what you’re going to get by default is much more secure, and I think that’s the most important piece. The community really loves Helm, and now they have Helm with better defaults.

Here’s the full video interview below, part of SiliconANGLE’s and theCUBE’s coverage of KubeCon + CloudNativeCon. (* Disclosure: The Cloud Native Computing Foundation sponsored this segment of theCUBE. Neither the CNCF nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE