NT1/SG-Meeting-2018-05-14
NT1 meeting 2018-05-14, 10:30 NordForsk, Oslo
Invited: Mattias Wadenstein, HPC2N; Oxana Smirnova, HEP, LU; Michaela Barth, PDC-HPC
Agenda
- Status MoU
- Data Lakes/Tier-2 storage consolidations
- NGI_NDGF services outside HEP
- Benefit analysis
- Risk analysis
- Starting to sketch a Workplan: Longterm strategy and ambitions
Minutes
Presence
- Mattias Wadenstein
- Michaela Barth
- Oxana Smirnova
Status MoU
Deadline for comments is today. Feedback from Sweden. Informal feedback from Norway on changing the way of pledges. No feedback yet from Denmark and Finland.
Without this MoU the NLCG committee has no authority on regulating the pledges.
Plan: Michaela to send a final reminder today and start calling people by tomorrow.
Data Lakes/Tier-2 storage consolidations
CERN and the experiments are starting to push for consolidation of storage sites (Freeing manpower money to buy more disk to enable the handling the expected amount of High-Luminosity data rates). Also efficiency might increase when going for a federated model.
To address this a new datamanagement project (DOMA) was born as an open WLCG collaboration. All the Tier-1s are supposed to contribute. Matthias, Oxana and Vincent are on the mailinglist, planned start-up meeting is conflicting with NorduGrid conference, though.
Opportunity: Streamlining Swedish Tier-2 and integrating Estonian and Finnish Tier-2 storage, like how we have integrated the Slovenian Tier-1. Maybe even other “nearby” Tier-2s if they are interested. This would increase our weight, increase overall efficiency, and deliver a better quality of service to experiments. Tier-2s which are not big enough might be encouraged to get dismantled, so there is a risk for them to stand outside of this consolidation.
Decision: The SG recommends to strive to take such opportunities for integration where available.
NGI_NDGF services outside HEP
Can NGI_NDGF support services outside of the main focus of HEP and WLCG?
Do we have resources that we should try to find new user communities for? No, our resources are bought by the national LHC computing efforts and belong to them.
If a Nordic site would like to be part of the EGI part of EOSC-hub, they could be allowed to be part of NGI_NDGF for operational purposes, even if they are not in the field of HEP.
Decision: The SG points out that LCG dedicated resources like dCache pools can not be used by other communities through the EGI FedCloud.
Decision: The SG approves that other Nordic sites could be registered through NGI_NDGF.
Benefit analysis
- Contract fulfilment towards CERN
- ATLAS and ALICE publications
- Competence (keeping and development)
- Nordic collaboration
- Bleeding edge: Driving development of Open Source Storage and Computing (dCache, ARC, Rucio, SGAS, etc)
- Contribution to Open Science
- Without NT1 there wouldn’t be a NeIC (past)
- Distributed Operations Best practices (e.g. inspiration for UNINETT Sigma2)
- Use case within CodeRefinery (ARC development, and ops configuration management)
Risk analysis
Decided measure includes prevention and mitigation
WLCG takes a different direction and leaves us on the side would be a new risk One country could leave WLCG MoU
The SG will look at this again next time.
Starting to sketch a Workplan: Longterm strategy and ambitions
Workplan should include also information on how shortterm planning is handled right now.
Ambition: Nordic Data Glacier and Fjords - one of the dozen consolidated storage spaces (Data Lakes) of the future WLCG for the HL-LHC era?
AP on Mattias to make a first draft of such an activity plan.
AOB
IJS in Slovenia and University of Bern should be a collaborator in the NeIC geographic distribution map.
AP on Mattias to update NT1 website with information on all collaborating (no paid staff) institutions.