Cloud Native ComputingDevelopersDevOpsFeaturedLet's Talk

How Virtual Clusters Improve Operator’s Workflows: Loft Labs @ KubeCon NA

0

This year, KubeCon is a hybrid of both virtual and in-person events. Loft Labs is attending the Cloud Native Computing Foundation‘s flagship conference as a company with four team members appearing in their official booth. They’ll also be giving a talk at the event which will be live-streamed and will also host a virtual booth for those participants who won’t be able to be on-premise. And, as usual, Loft Labs will be available in the CNCF Slack they’ve set up for KubeCon.

One of the benefits of having the virtual component is that it allows more people to attend. Lukas Gentele, Co-Founder and CEO of  Loft Labs, brings this up when he says, “My CTO, for example, is based in Germany. We have employees in Eastern Europe, in India, pretty much all over the place, because we’re building a distributed team and it’s really hard for them right now. It’s impossible to attend a conference because they just can’t enter the United States at this point.” Because of the virtual component, it will be easier for people to participate, watch the talks, and engage with the community.

But how is the CEO of Loft Labs participating? He’s giving a talk called “Beyond Namespaces: Virtual Clusters are the Future of Multi-Tenancy,” which is focused on virtual clusters and how they can be the future of Kubernetes multi-tenancy. According to Gentele, “It’s a really interesting talk, that’s going to be centered around our new open-source project, vcluster, which recently became a certified Kubernetes distribution.” He adds,  “I think that’s going to make it very, very interesting for a lot of folks because it’s a new project and they may not have heard about vcluster or may not have had any hands-on experience. And this talk will really dive into the different use cases for vcluster and virtual clusters in general, and how they can improve operator’s workflows when managing Kubernetes in a multi-tenant way.”

One of the more important features of vcluster is the ability to avoid namespace collision across a larger Kubernetes cluster. Gentele explains, “Just imagine you have an application, and for each pool request you stand up some kind of review application.” He continues, “If you know this application has a database component then you could only stand it up once because the names would be colliding. That’s why we’re using namespaces. You could now have a pool request one namespace with a database and pool request two with a database and they would not be colliding.”

The target audience of Gentele’s talk is advanced users. On this subject, Gentele says, “It’s pretty straightforward with two lines of code to spin up a virtual cluster. But the actual use cases are really more advanced.” He continues, “When you have hundreds of users in a shared Kubernetes cluster, you’re setting up pipelines that deploy thousands of times per week to a Kubernetes cluster and they need these isolated environments.” This topic is typically targeting large organizations that work with shared multi-tenant Kubernetes clusters.

Going back to Gentele’s talk, he believes the takeaway is not a deep-dive into the configuration of virtual clusters, but more a broader discussion to serve as an inspiration for companies to see what virtual clusters can really do. The talk will highlight some interesting use cases and show examples of real-world usage that frontier companies are already using.

This blog post, written by Jack Wallen, is based on our interview with Lukas Gentele. Please check out the full video interview.