Brainstorm an "Institutional FAQ"

JupyterHub
1

The Dask documentation has an excellent page called Institutional FAQ. It covers a few major FAQ items, broken down by different kinds of roles within an institution (e.g. a manager, a tech leader, an IT person).

I think it would be really valuable if we had a similar page about JupyterHub. We can use it as an opportunity to highlight the relative benefits of JupyterHub, its pros/cons, situations where it is useful, and cover some assumptions that may not be correct.

I’ve adapted the Dask questions slightly so that they make sense for JupyterHub, and am posting them below.

  • Please add your own answers to any or all of these questions. I’d love to see what folks in the community think
  • Please feel free to add your own new questions or user personas. I’m starting with what Dask had, but probably am missing some things (e.g. perhaps we should have a “teacher” user persona)

I’ll try to brainstorm my own answers to these questions. Looking forward to seeing what ideas people have!

For all:

# For all

## Is appropriate for adoption within a larger institutional context?

## I keep hearing about Jupyter Notebook, JupyterLab, and now JupyterHub. What’s the difference?

For management / leadership types:

# For management

## Briefly, what problem does JupyterHub solve for us?

## Is JupyterHub mature? Why should we trust it?

## Who else uses JupyterHub?

## How does JupyterHub compare with hosted products, like Google Colaboratory, RStudio.cloud, or Anaconda Enterprise?

For IT / infrastructure folks

# For IT

## How would I set up JupyterHub on institutional hardware?

## Does JupyterHub run well in the cloud?

## Is JupyterHub secure?

## Does JupyterHub provide computing or data infrastructure?

## How do I manage users?

## How do I manage software environments?

## How does JupyterHub manage computational resources?

## Can JupyterHub be used with my high-performance computing resources?

## How much resources do user sessions take?

## Can I customize the look and feel of a JupyterHub?
* Branding notebook server / jupyter lab. Custom error pages / support and help pages

For technical leads within teams or groups

# For Technical Leads

## Will JupyterHub “just work” with our team's interactive computing setup?

## How well does JupyterHub scale? What are JupyterHub's limitations?

## Will our team have to re-write their code when they want to scale to high-performance compute?

## Is JupyterHub resilient? What happens when a machine goes down?

## What interfaces does JupyterHub support?

## Does JupyterHub make it easier for our team to collaborate?

## Can I use JupyterHub with R/RStudio or other languages and environments?
2 Likes
2

In educational institutions, there are also differences in the way things might be used / readily identifiable but different user communities with different user concerns, as well as different requirements:

  • for teaching / learning (24 hour access; bursty use around assessment times; personal access); 1000s-10000s users
  • for research (closed research group access; data security concerns; archiving concerns); 10s-100s users;
  • for internal reporting (interfacing with legacy systems; workflows) 100s-1000s users;

Someone new to requesting a new service and making a case in support of it may well get bogged down in the detail of their own use case / concern, so a roadmap leading someone (a lone academic or researcher) from their specific concerns to the general concerns that are more likely to make a successful case for support might be useful? This might also help clarify how requirements might differ across user communities, which in turn may influence the sorts of question that need asking of them?

1 Like
3

Thanks for opening this! I love it.

Additional questions:

All:

  • I keep hearing about Jupyter Notebook, JupyterLab, and now JupyterHub. What’s the difference?

Technical:

  • Can I use JupyterHub with R/RStudio or other languages and environments?
1 Like
4

A couple more questions on specifics:

  • can I / how do I brand a notebook server / JupyterLab environment?
  • can I / how do I add customised error pages / support / help pages?

In terms of making notebook collections available across an institution:

  • is there any way of publishing a notebook gallery with previewable but non-runnable notebooks?
  • can I search over collections of notebooks?
1 Like
5

Hey all - I took a stab at answering some of these questions here:

Would love feedback of any kind! I hope it’s a useful resource.