Today we’re excited to announce the public beta of our fully-managed and hosted offering of Pachyderm: Hub.
When my co-founder Joey Zwicker and I started Pachyderm nearly 6 years ago, we set out to make data science better. Pachyderm’s current feature set provides collaboration on large scale data workloads, but it’s confined to companies with the resources to manage a Kubernetes cluster. We’ve felt for a long time, that the last part has been holding us back. That is, of course, until Hub.
What is Hub?
Hub is the connective tissue for Pachyderm. It enables you to get a Pachyderm cluster, on-demand, for personal, individual use, or to be shared by a team, without touching Kubernetes. It’s the quickest and easiest way to get to explainable, repeatable, and scalable data science.
Currently, Hub is totally free, but for the time being, only supports 4-hour temporary clusters. To start, each user that signs up will get access to a fully functional Pachyderm cluster that has 4vcpu’s and 15Gbs of RAM. Rest assured, larger node-sizes, GPU’s, and permanent clusters capable enterprise workloads will be available early next year, and we’ll be sure to keep you posted. These are the early stages so we only recommend that Hub be used for demos and proofs-of-concept until we’re able to get the bugs worked out. Over time, we’ll expand Hub into a full, large-scale data collaboration platform. We want Hub to be to data, what Github was to code.
If you’re interested in learning more about the origin of Hub from an engineering perspective, checkout our deep-dive post: (Kubernetes as a service) as a service
What’s our vision for Hub?
Hub is the natural evolution of Pachyderm; it’s something we always planned to build when we had a strong enough foundation and the resources to build on it. We’re starting small with a beta, but just around the corner is something much greater. However, we won’t be able to get there without your honest feedback. Together, we’ll cultivate Hub to become the easiest way to connect and collaborate with other data scientists - inside or outside company walls - and become “GitHub” for data science.
As with any platform, early users like you will be critical in steering it and defining what the platform ultimately becomes, so please don’t be shy with your feedback, good or bad. And most importantly, thank you for your support, we hope to stop by to say hi in our slack channel.Try Hub For Free