A lot of places use Jupyter Notebook (and JupyterHub) to work with private data where security compromises would have serious implications. We have a pretty good track record with security ( to everyone reporting and fixing security issues), but it would be great if we can get a formal, external security audit from a firm specializing in this. This increases notebook security, and acts as a strong signal to many organizations.
This requires funding, and someone to see it through. So,
- How do we find the money for it?
- Who can manage the process? This involves finding a security audit firm, negotiating with them, and seeing it through.
What do you all think?
2 Likes
It sounds excellent!
We would need to provide a good scope for requested security audit. You wrote Jupyter Notebook but I started thinking about kernel network communication, jupyterhub, singleuser servers running jupyter notebook
or jupyter lab
, etc.
Did you have a specific scope for the audit in mind?
Jupyter is currently provided as a service
by most if not all the major cloud providers and also used in multiple large organizations that have easy access to resources that can perform this security audit
/threat analysis
/penetration tests
. Jupyter.org should partner with some of those, which will be beneficial for all involved parties.
Another option might be sponsoring some kind of bug bounty
programs around security
related bugs.
2 Likes
I could ask if our AppSec team has available resources.
I like @lresende’s suggestion. This is a great opportunity for contributions that aren’t code and don’t require a long term commitment. Having the work done as a contribution to the project like all the other professionals who donate their time (aka AppSec team from 1und1 (I think that is who @jhermann works for)) would also be great.