WLCG Management Board #337

Tuesday 17 Feb 2026, 16:00 → 18:00 Europe/Zurich

513/R-068 (CERN)

16:00 CERN/10:00 EDT/09:00 CDT

URL for the ZOOM call: https://cern.zoom.us/j/66011416092  (the passcode has been distributed by email)

To join by phone (see the dedicated support page):
1) dial your local ZOOM number (e.g. +41432107108) then enter the MB meeting ID  660 1141 6092 then #
2) mute your phone (there is no auto-mute facility)
Note: calls are charged to the caller (there is no callback facility)

Email Distribution List: worldwide-lcg-management-board@cern.ch
Email List Archive: worldwide-lcg-management-board (requires CERN authentication)
Minutes: Management Board Meeting Minutes
Action List: MbActionList
See also the: WLCG Document Repository,  WLCG Web Site

16:00 → 16:05 Minutes and Matters Arising

• On the 28th of Jan, the CB endorsed the two deputy positions, for David and Ale.

• Incidentally, I just received the official charge by the CERN DG (not that operationally it mattered too much).

• Still no news from the CERN legal service for JINR connectivity. As discussed last time, we go as we are. I and Simone had a meeting ~10 days ago with the Director of Meshcheryakov Laboratory of Information Technologies about JINR pledges and network. The ICA expiration date is 27 January 2030. We asked him to provide details of the GPN link (since we know it works, but we have no idea how close it is to its limit).

• The committee for the WLCG-HSF WS in Bologna, Nov 2nd-6th, has been formed. We are going to send a "save the date" soon.

• Agenda for the LHCC in March has been defined. Apart from the standard talks, a touch base on ROOT has been requested by the reviewers. Apart from that, standard Experiments' talks + a touch base on tokens and sustainability.

• Julia has progressed with the “T2 requirements” page. Thanks to her and all the experiments which participated. The link is here. There are still differences we need to understand before making it “public” (even if only on demand, when a new center approaches WLCG/the experiments). Idea is to go category per category and open a thread with the experiments where the situation is unclear.

• We were informed that the ENSURE EU project (HORIZON-INFRA-2025-01-TECH-01, EGI driven) has been approved and will start in due time. 

• "This project contributes new technologies and solutions for reducing the environmental and climate footprint of research infrastructures, tackling sw solutions for reduction of carbon footprint of computing, a collaborative project with various partners including CERN and SKA, and contributing to the technical evolution of accounting"

• For us, personpower on accounting  and power reporting (APEL, AUDITOR, ….): STFC, Freiburg, CERN, Glasgow and Manchester. We should consult experiments on thoughts about accounting but avoid any mission-creep.

• We were also informed that the EU SPE Project (“Strategy for Post ExaScale” - HORIZON-CL4-2025-03-DIGITAL-EMERGING-04) has been approved

• Very small project from the financial point of view. Includes EU HPC centers, Quantum technology providers, industries

• HEP is the only science directly represented

• “The SPE project addresses one of Europe’s most pressing technological challenges: preparing the convergence of High Performance Computing (HPC), Artificial Intelligence (AI), and Quantum Computing and Simulation (QCS) in the context of post-Exascale era.” The concrete outcome would be a “a roadmap for post-Exascale computing in Europe”

• HEP members: CERN and INFN

16:05 → 16:10 Action List Review

16:10 → 16:20 Review of the strategy document

Speaker: Dr Tommaso Boccali (INFN Sezione di Pisa)

WLCG Strategy 2024-2027 - version for MB Feb 17th 2026.xlsx

16:20 → 16:30 WLCG Service Report

Speakers: Borja Garrido Bear (CERN), Maarten Litmaath (CERN)

WLCG-svc-report-260217-v10.pdf

WLCG-svc-report-260217-v10.pptx

16:30 → 16:40 TCB Report

Speakers: Alessandro Di Girolamo (CERN), James Letts (Univ. of California San Diego (US))

Report to WLCG MB #337 on Tuesday, February 17, 2026 at 16:00 (CERN)

Technical Coordination Board (TCB)

We have called for the next meeting of the WLCG Technical Coordination Board on Tuesday, February 24, 2026 at 16:00 (CERN). Indico: https://indico.cern.ch/event/1649097/. Goals of the meeting include:

• Checking on the status of each chapter draft of the WLCG Technical Roadmap (TR) from the various chapter facilitators.

• In particular we would consider taking advantage of upcoming OTF meetings to move the process forward, especially for Chapters 2 (Facility Evolution) and Chapter 5 (Workflow Management, insofar as it relates to facility or service evolution) at the HEPiX Spring Workshop in April.

• Set a timeline and procedure to complete the first rough draft of the TR document

• Discuss how to track progress towards the high-level milestones in the TR.

 

On the last point in particular, we should also start thinking about how to globally track and surface activities that are already happening across WLCG and related communities and that are relevant to the roadmap. We do not want to reinvent everything - we should build on ongoing work and make it visible and actionable. We would like to start something concrete in this direction and get your input on how best to do this in a lightweight but useful way. This has been already discussed related to the milestones, but there are still a few open points - e.g. how/when do we freeze some of them? And update if needed?

Upcoming Open Technical Forum (OTF) Meetings for 2026

• OTF#9 is being planned for March 24th & 25th on various topics such as XRootD monitoring, accounting, and token-based authentication.

• OTF#10 will be co-hosted with the HEPiX Spring Workshop in Lisbon in April on the subject of facility evolution (Chapter 2 of the WLCG Technical Roadmap). There is also a separate talk at HEPiX which is a summary of the Heterogeneous Architecture workshop in December.

• OTF#11 will take place in mid-May (exact date TBD) as a practice talk for CHEP (assuming that the abstract is accepted) on the WLCG Technical Roadmap.

• OTF#12 will take place at the end of August (25-26th), topics TBD

• OTF#13 will likely aim to take place during the XRootD and FTS workshop in Lyon September 14-18, to be discussed with the organizers.

• November 2nd-6th is the WLCG Workshop in Bologna.

Report on the Open Technical Forum Meeting (OTF#8) on Tape Systems Evolution and Performance

The eighth WLCG Open Technical Forum (OTF#8) was held at CERN and on Zoom on Tuesday, February 3rd and Wednesday February 4, 2026 on the topic of tape systems evolution and performance. There were up to 120 attendees in person at CERN and on Zoom over the two days. The Indico page [1] links to a live notes document [2] and the meeting was also recorded. 

On Tuesday afternoon we heard about the technological and market evolution of tape systems, tape deployments and performance at Brookhaven and Fermilab, and dCache tape and CTA evolution plans. Tape was seen as a viable technology for the future the evolution of which is being driven more and more by hyperscalers’ demand rather than “traditional” customers. 

Sites had different approaches to adopting or skipping LTO10, taking into account the high cost by TB relative to LTO9.  There are also different options for tape library deployment (scale-up vs. scale out), with the large sites opting for scale-up at this point. Sites also take different approaches to optimization of tape throughput performance, including hardware choices, striping writes across multiple libraries, optimizing (minimum) file sizes, optimizing writing to tape for faster or prioritized recall e.g., with organized writing, R&D into exploiting tape metadata for improving read efficiency but also preserving file co-location during repacking. BNL and FNAL did not anticipate repacking during LS3. 

On the software side, dCache is exploring QoS based on recall needs and bulk operations on files. CTA has development goals based on simplicity, scalability, and performance, with token-based authentication for the tape REST API in place by DC27. Collaborative efforts between the sites and software developers seems to be key to making advances in scalability, performance, and operational efficiency.

On Wednesday morning the experiments presented their points of view about their requirements, positive experiences, and pain points in using tape systems. The experiments emphasized that while capacity is fundamental, (read/write) throughput is also critical for their operations. There is a clear need for concrete estimates from the experiments of foreseen read/write rates across all workflows e.g., data taking, Monte Carlo, and user analysis, to help sites prepare

Common themes were dealing with tails of recall requests, knowing how to react to them. Prioritization of requests was seen by multiple experiments as a possible mitigation, but tape experts noted that implementing sophisticated prioritization is a significant technical stretch for tape storage (and a broken tape remains a broken tape regardless of the priority with which you want to read it). It was discussed whether improved monitoring could help experiments react to problems as well as additional discussion about tape metadata and how it could improve writing and recall and also help sites to react to experiment requests. 

ALICE and LHCb requirements in Run 4 will be similar to Run 3, but evolution for Run 5 will be significant for these experiments and needs to be considered by the WLCG community in the longer-term Technical Roadmap. The use cases of these experiments differ significantly from ATLAS and CMS, which recall much more actively and concurrently when writing to tape.

For the rest of Wednesday we had feedback from tape sites. The need for clearer policies e.g., for retries and throughput requirements was expressed. A metric for tape read and write efficiency (taking into account mounting latency and actual data transfer times) was proposed and shown that for writing CERN is above 90% while for recall around 40-50%. Better co-location e.g., by using archive metadata, may help us improve this. A realistic goal would be 80% read efficiency.

During the discussion on the CMS recall tests it was highlighted the importance of tape families. Sites not using tape families were performing below expectations, though it was pointed out that writing down concrete expectations for each site would be very helpful. We also heard about the CNAF experience from the hardware and software perspectives (StoRM and GEMSS). Several questions were raised needing further evaluation about archive metadata. The Tier-2 tape experience of ATLAS was also presented, which relieves some of the load on Tier-1 tape in particular for both detector and simulation AOD.

Finally there was a lively discussion about archive metadata, and it was agreed that the topic is important but has many variables and no clear simple path forward. More work is needed in the coming months to define possible improvements.

[1] https://indico.cern.ch/event/1609088/

[2] https://docs.google.com/document/d/1tRsEAeNjpAyRUAfuS5uYrP3otCOmq_febIrl-s-zvnU/edit?usp=sharing

16:40 → 16:50 Extension of the mandate of the "Job Allocations and Handling" WG

Speakers: Antonio Perez-Calero Yzquierdo (Centro de Investigaciones Energéticas Medioambientales y Tecnológicas), Julia Andreeva (CERN)

20260217_WG_MandateReview.pdf

16:50 → 16:55 AOB