E3DDS Stakeholder 20190327

From neicext
Jump to navigation Jump to search

Time/Location

  • Wednesday March 27th 2019, 10:30 - 17:00 EET
  • Meeting room "Debatti", CSC, first floor. Information on how to get to CSC Espoo here

List of Attendees

  • John White (NeIC)
  • Tomasz Malkiewicz (NeIC)
  • Ari Lukkarinen (CSC)
  • Mattias Wadenstein (HPC2N)
  • Dan Jonsson (UiT)
    • Anders Tjulin (EISCAT)
  • Craig Heinselman (EISCAT)
  • Ingemar Haggstrom (EISCAT)
  • Johan Kero (IRF)
  • Teemu Kiviniemi (CSC)
  • Lars Fischer (NORDUnet)
    • Borje Josefsson (SUNET)
  • Helge Stranden (Uninett)
  • Olaf Schjelderup (Uninett)
  • Harri Hellgren (EISCAT)
  • Esa Turunen (SGO)
  • Roy Dragseth (UiT)
  • Jani Myyry (FUNET)
  • Matti Laipio (CSC)

Objectives for the meeting

The E3DDS stakeholder forum gathers representatives from the NeIC EISCAT_3D Data Solutions project, NeIC, EISCAT, EISCAT_3D users, national e-Infrastructure providers and network providers in order to achieve the goals:

  1. To update each other about progress of projects and e-Infrastructure roadmaps;
  2. To intensify the collaboration between the EISCAT_3D Data Solutions project and the providers for network and e-Infrastructures.

The general objective of this meeting is to bring together the EISCAT_3D technical community with national providers of HTC/HPC and networking from the relevant countries. The goals of the meeting include the need to update each other of statuses and time-lines of national providers and the EISCAT_3D activities.

Agenda

  • Meeting chair: Tomasz Malkiewicz, NeIC
  • 10h00 - 10h30 Coffee
  • 10h30 - 10h45 Welcome (Tomasz Malkiewicz, NeIC)
  • 10h45 - 11h10 Status of the EISCAT_3D project slides (Harri Hellgren, EISCAT)
  • 11h35 - 12h00 Status installation to Skibotn site slides (Helge Stranden, Uninett)
  • 12h00 - 12h25 Preparation for EISCAT_3D Karesuvanto site slides (Esa Turunen, SGO)
  • Uninett status slides (Olaf Schjelderup, Uninett)
  • 12h30 - 13h15 Lunch
  • 13h15 - 13h30 Nordic Network Plans slides (Lars Fischer, NORDUnet)
  • Report on Deliverable 1 Slides (Mattias Wadenstein, HPC2N)
  • Report on Deliverable 2 (Ari Lukkarinen, CSC)
  • EISCAT_3D and Space Debris (Johan Kero, EISCAT)
  • 14h30 - 15h00 Coffee plus Finnish fika
  • Progress on Norway National e-infrastructure slides (Roy Dragseth, UiT)
  • Progress on Swedish National e-infrastructure slides (Mattias Wadenstein, SNIC)
  • Progress on Finnish National e-infrastructure slides(Teemu Kiviniemi, FUNET)
  • 16h15 -16h30 Wrap-up (Tomasz Malkiewicz, NeIC)

Minutes

Attendance

    • John White (NeIC)
    • Tomasz Malkiewicz (NeIC)
    • Ari Lukkarinen (CSC)
    • Mattias Wadenstein (HPC2N)
    • Dan Jonsson (UiT)
    • Craig Heinselman (EISCAT)
    • Ingemar Haggstrom (EISCAT)
    • Johan Kero (IRF)
    • Teemu Kiviniemi (CSC)
    • Lars Fischer (NORDUnet)
    • Helge Stranden (Uninett)
    • Olaf Schjelderup (Uninett)
    • Harri Hellgren (EISCAT)
    • Esa Turunen (SGO)
    • Roy Dragseth (UiT)
    • Jani Myyry (FUNET)
    • Matti Laipio (CSC)
    • Carl-Fredrik Enell (EISCAT) via remote

Welcome (Tomasz Malkiewicz, NeIC)

Round of introduction of participants as there were a few new faces.

We have booked Oct 17th 2019 for the next Stakeholder meeting in Tromso NO.

Status of the EISCAT_3D project (Harri Hellgren, EISCAT)

  • Discussion on the testing of the FSRU units... All the testing of the units are done in China by the supplier.
  • The the subset of level 1b data that can be saved offline is still not really known. This depends on buffer and network capacity.
  • For the level 1a data one has to consider a hard target filter. The meteor zone is below 150km altitude. So this should not be affected (investigate?) Possibility of a white-list?
  • The site buildings schedule was discussed... some sites still waiting for permits etc.
  • For the final funding, EISCAT is still completing applications and negotiations. Anticipate to scale back a bit on the initial deployment. Not the full stage 1.
  • There is still ongoing with an idea that CN could fund sites for the later stages (Andoya and 2nd site in SE). But have to mesh with the five-year plan schedule. No progress on the 5 sites deployment.

Status of the E3DDS project (John White, NeIC)

Advice to check out the ESCAPE project (in The ESCAPE project aims at delivering a shared solution to computing challenges in the context of the European Open Science Cloud. It targets Astronomy and Particle Physics facilities and research infrastructures and focuses on developing solutions for handling large sets of data.

Status installation to Skibotn site (Helge Stranden, Uninett)

  • The outriggers only at the Skibotn site. Used for interferometry.... to look within the beam. These are not present at other sites.

Preparation for EISCAT_3D Karesuvanto site (Esa Turunen, SGO)

  • The negotiations with Sami ongoing... want to talk each year. Need to be honest with them.
  • The number of antennas in are the same number as elsewhere (except no outriggers).
  • Also need to remember that the costs of fibre from the nearest "box" to local sites excluded. Advise to add this when digging for the power. Put extra tubes in the ground from day one.
  • Comment from the SE site, there is a new admin of the local Sami people. So some new talks are needed.
  • The VAT situation in Sweden is still open. This is not known yet.

Uninett status (Olaf Schjelderup, Uninett)

Some slides by Olaf Schjelderup from UNINETT are shown. There is a new data centre near Tromso backed by the municipality. Commercial run. Could be a useful for EISCAT_3D? Connected by 100 Gb/s network.

Nordic Network Plans (Lars Fischer, NORDUnet)

It is explained that there is 100 Gb/s TOTAL traffic (NOT duplex) on the ring. Not between the data centres. Upgrading to a 200 Gb/s ring in the future would make the network totally safe for EISCAT_3D to export level 1b data off-sites and to a data centre(s).

The new NORDUnet network works on a "ladder" model. There are NREN-operated interconnects between countries and lateral (E-W) international connections.

Report on Deliverable 1 (Mattias Wadenstein, HPC2N)

The initial operations are planned to be at sites. Not at HPC centres at the start-up. Will migrate of neccessity. The storing data and planning for this will be a later decision.

EISCAT anticipates that there are personnel on-site at the start of the operations. These will be removed later in 2/3 years time. The maintenance of HW a factor on the sites. Also the day to day operations of the services required, but this can/should be done remotely by experts.

Report on Deliverable 2 (Ari Lukkarinen, CSC)

There needs to be a statement on GPU or FPGA black-box solutions in May. Not defined whether the beginning or end of May.

The current network model assumes that all FSRU data can be sent to all SBF processing units (whether CPU, GPU or FPGA). If there are multiple GPUs per box then each GPU can take a fraction of the data. Caution... for the IBM Power9 servers the 6 GPU version needs water cooling! Anticipate 380 Gb/s into each node.

EISCAT_3D and Rockets (Ingemar Häggström, EISCAT)

Clarified that the height of the rockets are possibly up to 500 km.

The SSD buffer space on the sites can be used to keep up with real-time data. In this case for rockets the SSDs may need to to real-time saving of data.

The bandwidth required for the rockets supporting studies could be reduced. If the data was stored on the SSDs then could be read back in "real-time" replay later. Discussed whether the level 1a data can be sent to multiple locations... i.e. could be analysed real-time and also dumped to SSD for later replay and re-analysis? This should be looked at by the project. (Meeting May 29th)?

EISCAT_3D and Space Debris (Johan Kero, EISCAT)

The number of objects goes up all the time as in the number of satellites. Many telecoms satellites being launched. So there is the need to improve the orbits of even more particles.

The computing requirements is that the computing is real-time while the object is in the field of view. This could be refined to make a more precise measurement... The results are needed within a "few" seconds... the objects are typically in the field of view for minutes. There is a catalogue of accepted objects. This filter needs to be applied in the FSRU (before level 1a).

Progress on Norway National e-infrastructure (Roy Dragseth, UiT)

A Data Centre in Norway, operated by Tromso, could be on the EISCAT ring. Need 12 months lead time for data centre increase in capacity.

Progress on Swedish National e-infrastructure (Mattias Wadenstein, SNIC)

The AI plan (WASP) has an aggressive timetable... There are WASP resources planned to be in place in 2019.

Progress on Finnish National e-infrastructure (Teemu Kiviniemi, FUNET)

No comments or discussion.

Wrap-up (Tomasz Malkiewicz, NeIC)

The E3DDS project will now head towards the benefit realisation.

The next E3DDS Stakeholder meeting will be on 2019/10/17 in Tromso. 10h30 local time.

Retrieved from "https://wiki.neic.no/w/ext/index.php?title=E3DDS_Stakeholder_20190327&oldid=10594"