Choose timezone
Your profile timezone:
Plenary Meeting Techwatch WG
Description
Initial meeting for the revived HEPiX Techwatch Working Group
Shigeki Misawa
HEPiX Techwatch
Present: Andrea Sciaba, Alison Packer, Andrea Chierici, Andrea Valassi, Charles Leggett, Alison Packer Eric Yen, Martin Gasthuber, James Walder, Michele Michelotto, Shigeki Misawa , Thomas Byrne, Xavier Vladimir Sapunenko, Markus Schulz (minutes)
Apologies:
Harvey Newman Andrea started to point out the main topics: 1) The Mandate of the WG can be seen at: https://w3.hepix.org/techwatch.html 2) Definition of the subgroups. This can be found on the above page. There has been an evolution since the last meeting. 3) We need to collect ideas how the visibility of the working group can be expanded. The mandate form the WG web page was discussed. Martin G. suggests to add the aspect of sustainability, including the production and operations related costs and energy. Xavier: Asked whether it s really needed to exclude DAQ technology? Martin G: There is an overlap with DAQ tech since this also is moving towards commodity hardware. Vladimir: Is this the right organisation of this work? Maybe we should not form a WG, but a forum where we can discuss developments and express our opinion. In the current approach we have to suggest a common result when we write the document. Andrea: This is already the case, we are not suggesting one view. But we have to produce documents to reach the community. Vladimir: These documents are the problem. They limit the way to express different opinions. Andrea doesn’t agree with this. In his view the documents are a collection of findings. The result of the discussions. Martin: If there is a non agreement then this can be documented. Vladimir: Can’t see how we can discuss different views when we produce one document in the end. Andrea lists the communication channels. When we start producing the document then the discussions will start. Vladimir: More space for discussions is needed. Andrea: Mattermost will be opened soon, this will create a forum for discussion. Short discussion on clarification of the usage of Mattermost for discussions only, announcements have to be sent by mail too. Shigeki: Alternate/multiple views should be visible in the documents. If there are alternative views they have to go into the document or they will reach no one. We do two things: 1: collect tech info, which is neutral 2: express value judgements which are based on personal preferences. This has to be showing up in the docs and presentations too. Andrea: Next topic is the presentation on subgroups. Our goal today is to converge on the subgroup definition, guidelines for managing the subgroups and assign people. There should be a manageable number of subgroups. 6-7 can be handles. It would be good if they could have comparable sizes and cover all relevant technologies. A specific technology should not span multiple subgroups. Andrea showed the proposed groups. Guidelines on collaborative tools have been discussed. There will be some use of Google docs, but to reach those who don’t use Google products there will be also PDFs. We are trying to fill the table of subgroup membership. We are looking for volunteers (HEPIX workshop). Andrea V: We are 7 people connected, how do we cover the subgroups? How can this be covered? Andrea S: The subgroups can organise themself (as they did last time). Andrea V: Full WG meetings should be held regularly. Charles: At the end one of the purposes is to help experiment to decide on buying hardware. MSRPs have little to do with the real prices. These depend on the deals they get and rebates for buying bulk. Do we encourage sites and experiments to publish their deals? Andrea S: No, we should put the focus more on the change, not on absolute prices, we can’t be too detailed. Eric Yen: Maybe we should also include the file system? Andrea S: FS is software and it comes most of the time with the OS. If it is crucial to use the software and there is a link to the hardware, then yes. Shigeki: If we look at FSs we can get into a large numbers. We should focus on generic aspects. RAID and how to get reliability in different ways is generic, not bound to specific hardware. For Flash Systems the discussion would be whether you can build one yourself? Xavier: Once the subgroups have been defined they can expand their scope and cover what matters for them, like shared filesystems. Charles: However, all the time the justifications have to be related to their mandate. Andrea presented the following slides: https://docs.google.com/presentation/d/1oxQiyOxWoPRr2RSHo4Ufge5mbQs_el4X_A_lbs0_fNk/edit#slide=id.g2bb2197d414_0_20 Currently the table on slide 8 with the coordinators and contributors is still quite empty. Please add names. For the April HEPIX we need a plan. All WGs are supposed to present for the first time. We have to start to organise and collect information. Martin: Who plans to attend HEPIX in Paris? Michele plans to go, but not completely clear. Alison Packer (STFC) doesn’t know yet. Michele pointed out that we need some first results within the next two months and wanted to know whether this is feasible. Andrea S. confirmed that this should be possible. We will have the next meeting in the second half of March. We have to check whether there is an overlap with ACAT?
There are minutes attached to this event.
Show them.
