LHCONE R&D and RNT-WG call

Europe/Zurich
Edoardo Martelli (CERN), Joe Mambretti (International Center for Advanced Internet Research Northwestern University), Marian Babik (CERN), Shawn Mc Kee (University of Michigan (US))
Zoom Meeting ID
97610360540
Host
Edoardo Martelli
Useful links
Join via phone
Zoom URL
    • 17:00 17:20
      Preparation of SC23 demos and DC24 mini-challenges 20m

      Presents: Edoardo Martelli, Tim Chown, Asif Shah, Joe Mambretti, Eli Dart, Bruno Hoeft, Carmen Misa, Xavier Espinal, Garhan Attebury, Vitaliy Kondratenko, Tristan Sullivan, Andy Lake, Phil Demar, Justas Balkas, Jim Chen, Marcos Schwarz, Chris Walker, Ofer Rind, Thomas Araneta, others


      NOTED demo for SC23
      - CERN and FNAL had a meeting and agreed on what is needed 
      - Canarie and ESnet are implementing the 400G path from TRIUMF to Starlight
      - JoeM asks if the VLANs should pass by the SC booth: Edoardo says yes, if possible.

      Packet marketing demo for SC23
      - Eli: ESnet is preparing to collect statistics on marked traffic and show them at SC23.
      - Joe: Yatish (ESnet) can measure traffic passing on production links. There are ESnet production routers in Seattle and Starlight, so measurement could be done there.

      DC24 challenges
      - NOTED: will run a new version that kicks in when links are saturated. It will not make any real modification to the network though, just signaling when an action is needed and later withdrawn.
      - BBR: Carmen has compiled kernels with BBRv3 for CentOS and Ubunto (5.14).  TimC is also doing some tests. There is a meeting scheduled for next week https://indico.cern.ch/event/1329666/
      - Jumbo frame. Tim: there are two sites in UK that are willing to participate. There are other sites interested.

    • 17:20 17:40
      RNTWG 20m

      SC23 status:
      - DTNs: Starlight (ready/have access), KIT (in prep), CERN (ready - connection TBD), UVic (in prep ?),
      - Collectors: ESnet (ready) - will need access from CERN; need to check if firefly netlink data will work (Andy)
      - Dashboards: CERN (in prep) - Grafana/Prometheus to be configured (Marian); 
                                Got access to the InMon dashboard, can try to test it with linux agent (network sflow exports will need to be configured by SCinet)
      - High-touch: visualisation (in prep) - python workbook to be shared (Yatish)

      RNTWG packet marking - list of activities:
      - Collectors developments (needed):
      .    - flowd extension to send netlink data (if only in state:"end" firefly then we also need to research ways to compress time series data)
           - access to GEANT data source to check if we could show utilisation + fireflies side by side
           - determine if we could plot data side by side with WLCG network monitoring
      .    - dashboard to show netlink data coming from flowd 
           - forwarding fireflies to another collector
      - Storage and transfers developments
           - dCache (needs testing)
           - Rucio 
           - FTS/gfal2 (progressing)
           - flowd in EPEL 
      - R&D
           - Use HBH option (instead of flow label) 
                  - Extension header reachability/drop rate study needed   
                  - Find ways how to implement HBH in Linux and eBPF
      - Others:
            - CHEP2023 paper submitted (we should update scitags.org with publications)
        

      RNTWG packet pacing - meeting next week (https://indico.cern.ch/event/1329666/) - BBRv3 

       

    • 17:40 17:55
      AOB 15m
    • 17:55 18:00
      Next meeting 5m

      Marian and Joe may not be able to attend, but next meting is confirmed for the 10th of October anyway