KubeCon Paris: Launching SpinKube and Shaping the Future of Kubernetes with WebAssembly Integration
Matt Butcher
wasm
spinkube
webassembly
kubernetes
There are a few past KubeCon events that really stand out in my mind. The first KubeCon, small as it was, had that palpable vibe, as we all felt that we were part of a new thing that was about to explode. KubeCon Seattle in 2018 represented that watershed moment. I’ll never forget reading the second Phippy book on the keynote stage.
KubeCon Paris now has a place on that list.
We kicked off KubeCon + CloudNativeCon Europe 2024 with the announcement of SpinKube and its enterprise big sibling, Fermyon Platform for Kubernetes. SpinKube provides a toolkit for installing WebAssembly (Wasm) support into your cluster, and then deploying and managing Spin applications. And Fermyon Platform for Kubernetes adds the engine that powers Fermyon Cloud, making it possible to run thousands of Spin applications on even a modest-sized node.
SUSE announced support for Spin applications in the latest version of Rancher Desktop, while the NGINX team posted about running Spin apps in NGINX Unit. Civo, our favorite Kubernetes cloud, has already tested SpinKube on their own Kubernetes clusters. At AI Day, Radu and I showed how we time-slice GPUs in Fermyon Serverless AI (a feature we’re rolling into Fermyon Platform for Kubernetes) and how we abstract away from GPU/CPU architecture to give developers a fantastic development-to-deployment experience. Meanwhile, multi-time founder and Fermyon COO Tim Enwall shared his acquired knowledge at Cloud Native Startup Fest in his session "Startup Lessons from 25 Years and Five Startups".
Akamai hosted Fermyon on their showcase stage twice on Wednesday. Thorsten Hans gave a developer’s introduction to Spin in his packed showcase. If you missed it, don’t worry. Thorsten and Sohan covered many of the same themes (and also did a show floor walk) in their livestream.
Later, Mikkel joined Nigel Poulton and the Akamai gang for a fireside chat about Wasm, Spin, and edge computing.
But the big moment for us was Thursday morning, when founding engineer Michelle Dhanani took the KubeCon keynote stage with Kai Walter from ZEISS Group and Ralph Squillace from Microsoft. The three of them showed several demos of using SpinKube, including one inspired by ZEISS’ order pipeline, where they handled 10k requests in a slick scale-up scenario. They ended the presentation with the announcement that, together with SUSE, Liquid Reply, and others, we were contributing the SpinKube project to CNCF.
Later that afternoon, Radu and I joined Michelle and Ralph for a session on SpinKube in action. We gave a deeper dive into the architecture of SpinKube, how we achieve higher levels of density, and how Spin applications can run side-by-side with container apps.
Ralph even told the story (which I thought apocryphal) about an Excel C library from 1985 that lives on in Wasm form in the Office.com version of Excel.
What was most exciting to us, though, was the tremendously enthusiastic response to SpinKube. We’re super excited that SpinKube can kick up the density of a cluster by 50x, but we’re equally excited that the same developer experience that can get you from blinking cursor to deployed app in 2 minutes or less transfers 100% to this bold new Kubernetes world.
If you’ve tried SpinKube already, we’d love to hear your story. And if you’re ready to try it out, let us know. Join our Discord, or head straight on over to the SpinKube website.
And if you’re part of an enterprise that wants to explore how to drastically increase the density of compute on your cluster (and thus drastically reduce the cost of operating that cluster), get in touch with us. We’d love to chat!