SoC Committee Meeting

Wednesday 2 Sept 2020, 15:00 → 17:00 Europe/Zurich

Agenda:
- Support of CentOS/aarch64 by CERN-IT
- Date of the next SoC IG meeting
- Possibility of a SoC Workshop in 2021
- Common System-on-Module
- AOB

Minutes:
- Support of CentOS/aarch64 by CERN-IT:
  Open issues:
  * Support for CentOS 7, required by ATLAS. Not sure when ATLAS will be able to migrate to CentOS 8 - mostly related to man power.
  * Packages in addition to CentOS, e.g. EOS, puppet, etc.; need to define which ones are required by ATLAS and CMS; do not ask support from Linux team, but rather for them to work as a single point of contact.
  * Suggest to move the discussion with the Linux team up in hierarchy in order to see what issues IT has, in particular, in terms of resources.
  Action: Marc to send an email to Alex Iribarren, in order to continue the discussion.

- Date of the next SoC IG meeting:
  Several dates were discussed, October 6 seems to be the best proposal.
  Action:
  * Ralf to make an open call for contributions for the next IG meeting on Tuesday, October 6, starting at 14.00.
  * For next year, it might be useful to liaise with the TWEPP organizers, in order to be considered in their planning for "user group meetings".

- Possibility of a SoC Workshop in 2021:
  Mainly three issues:
  * If a workshop in 2021, the question is, can it be a meeting in person or will have to be virtual. This is essential for reserving the rooms, and for people making their travel arrangements.
  * Date of the workshop, if in 2021, should be around May, June.
  * What material for the workshop, in addition to and beyond the series of IG meetings? Some possibilities include vendor presentations, e.g. Enclustra, Trenz, and tutorials, e.g. Continuous Integration.
  Action: Ralf to draft an email with a proposal for a workshop in May/June 2021 and asking if people prefer a presential or virtual meeting; and what contribution they could make, or what they would expect from the workshop.

- Common System-on-Module:
  Serval issues:
  * Discussion on the difficulties to define a common, agreed set of requirements.
  * Time scale of such a project: would have to move quickly in order not to be too late for phase-2 projects.
  * The connector definition is crucial. Adapter cards could be developed if needed.
  * The SoC on an RTM could be an alternative to having the SoC on a mezzanine, e.g. CMS/DTH.
  * Marc suggested to have an interface definition at the level of the signals going over the RTM (zone 3) connector, with the possibility of different implementations.
  * Discussion on the need for a common project to have a group taking strong responsibility and ensuring the long-term support.
  * Possibility to discuss this issue at the next IG meeting.
  Action: Marc to talk to Dominique, Jeroen and other experts to see if an interface definition is feasible.