Jupyverse may take a different path than JupyterHub for a multi-user scenario?

[damianavila]

From this comment (and below): jupyter-server/team-compass#11 (comment)

We might scale Jupyverse differently

Interesting...

so that it remains the only server.

serving what... what would be the entity you are spawning? Kernels? Something else?

And @davidbrochart suggested continuing the conversation here 😉

[welcome]

Thank you for opening your first issue in this project! Engagement like this is essential for open source projects! 🤗

If you haven't done so already, check out Jupyter's Code of Conduct. Also, please try to follow the issue template as it helps other other community members to contribute more effectively.

You can meet the other Jovyans by joining our Discourse forum. There is also an intro thread there where you can stop by and say Hi! 👋

Welcome to the Jupyter community! 🎉

[davidbrochart]

I mean that we will probably not scale Jupyverse by launching a server for each user, but by using other techniques such as replication.

[Wh1isper]

I have had some practice in this area, similar to Deepnote, where splitting the kernel, file system, and RTC into individual services.

As an example of file system, in my product, I implemented a standalone file system that mounts a whole block of user storage (remote path like nfs://notebook) and then creates a separate space for users under it (like nfs://notebook/user1) to provide file services directly to users when they enter our page without relying on Jupyter Server's back-end functionality. When the user uses Kernel for computing, only the user's directory needs to be mounted in the Kernel container, which is easy to implement in K8S

In this respect, jupyverse and FPS would bring a better architecture in a cloud environment (a true WEB IDE architecture), where users would be able to edit code directly without starting any compute resources (as they would locally with tools like Pycharm), and if a stateless front-end like JupyterLab could be implemented, and then jupyverse's multi tenant (file system, remote kernel), a DeepNote-like product could be implemented