This is not an official committee (yet). If you are interested, please contact haaijer@snt.

SNT has always been a big proponent of open source and self-hosting. Given the current (~2025) happenings in the US, decision makers in Europe and the UT are also starting to share these ideas.

We have been focusing on roughly 4 separate subprojects. Some will be broken out when they get bigger:

For the last ~8 years SNT has been running an oVirt cluster to provide virtual machines to associations at the university and for it's own projects. Since the software is not very maintained anymore, we are looking at alternatives. The research done for this might also help other organisations with the options of VM hosting outside of the big commercial clouds. As of writing, no definite option has been chosen yet. But it will likely end up between Cloudstack, Incus or Proxmox

Some evaluated options:

• Proxmox
  • Very popular, good documentation
  • UI is quite clunky
  • Multiple SNT'ers have experience with this. Some bad some good
  • Missing some nice to haves such as reinstalls, cloudinit scripts
• Cloudstack
  • Less popular but seems to be growing
  • More cloud-based, so end users can make their own VMs on demand
  • We need to look into this more.
• Openstack
  • By far the most features
  • also more cloud-based
  • Hard to maintain due to it's size. Not suggestedd.
• oVirt
  • Using Gluster is preferred, but this seems quite buggy
  • Seems to break in weird ways
  • Not recommended.
• Incus
  • Some SNT'ers are very positive about this.
  • Need to look more into this
• OpenNebula
  • Starting to gain some traction again aswell
  • No experience with it
  • Differences between enterprise and community edition unclear
• Custom libvirt-based
  • Requires development work
  • Not preferred, but possible

TODO: Matrix, Mattermost, Discord, Simplex, UT situation, etc

TODO: Opendesk, LaSuite, MijnBureau

SNT has a workstation GPU available in the room for anyone wanting to experiment with running local LLM's. It can run deepseek etc. at decent speeds.

Local LLM's do not yet have the quality that the big commercial ones do, partially due to hardware limits and partially due to the models. They do however get quite close.

The UT (LISA+EEMCS) are working on procuring some GPU's to set up a local LLM for UT students and researchers. If necessary it might also be possible to lend some unused compute from one of the research groups?

This issue needs work on two levels: Operational and Strategic. Us technies will mostly focus on the operational side, but keep in mind that what you setup will be used by more people than just you. We need something easy to use stable, and open. This is what the decision makers high up care about.

If you are related to a Dutch university and would like a free VM / physical machine to test new projects, SNT can help you!