Difference between revisions of "SRO RTDP Jul. 1, 2024"

From epsciwiki
Jump to navigation Jump to search
(Created page with "The current weekly meeting time is every other Monday at 10:00 US/Eastern === Connection Info: === <div class="toccolours mw-collapsible mw-collapsed"> You can connect using...")
 
 
Line 88: Line 88:
  
 
=== Minutes ===
 
=== Minutes ===
<!-- Attendees: David L., Ayan R., Vardan G., Fabio R., Jeng T. -->
+
Attendees: David L., Ayan R., Vardan G., Marco B., Jeng T., Cissie M.
 +
 
 +
* Marco
 +
** Fabio not available today, but is at JLab and can meet in person
 +
** They have determined that the Alveo boards are not really suitable for steam processing
 +
*** OS used for them is closed and has restrictions on how algorithms are implemented
 +
*** unable to get improved performance over CPU for applications tested under NP-like conditions
 +
*** Looking at alternatives
 +
** BDX
 +
*** Met with Thia last week. Project will move forward but they are waiting on formal statement from JLab on commitment
 +
*** Full prototype (electronics) expected by Jan. 25 @ JLab
 +
*** Japan collaboration is working on pre-approval. They have some of their own software they are interested in using (unclear on what functionality it covers)
 +
*** Electronics and staff to support it will be supplied by Hall-B
 +
 
 +
* Vardan
 +
** Still working on DPPUI. Nothing new to report
 +
** Some work done on Data Lake
 +
*** Multiple options tested: Kafka, Redia, xMsg, Chronicle Queues
 +
**** A demo was shown from of each from slowest to fastest
 +
****# Kafka
 +
****# xMsg (zmq)
 +
****# Redis
 +
****# Chronicle
 +
 
 +
* Jeng
 +
** Looked some into network_exporter for Docker
 +
** May be limited in ability to use in user space. More investigation needed.
 +
 
 +
* Ayan
 +
** Working on ACAT2024 proceedings paper
 +
** Need plots from Dec. 2023 packet capture data
  
 
=== Action Items ===
 
=== Action Items ===
 +
# Finalize ACAT2024 paper

Latest revision as of 13:25, 15 July 2024

The current weekly meeting time is every other Monday at 10:00 US/Eastern

Connection Info:

You can connect using ZOOM Video conferencing (ID: 160 248 4178). (Click "Expand" to the right for details -->):

Agenda:

  1. Previous Meeting
  2. Announcements:
  3. CLAS12 packet capture: May 16
  4. R&D Projects
    • BDX Experiment
    • Configuration
      • JCedit++
      • rtdp YAML
  5. Data Capture and Processing
    • Aggregation/Streaming
      • GlueX
      • CLAS12
      • ePIC
  6. Milestone Progress: (Github Project)
    • FY24Q2
    • M05: Create stream splitter program for EVIO or HIPO data formatted files
    • M06: Create stream splitter program for simulated data in PODIO for ePIC
    • M07: Create VTP emulator using files produced by stream splitter
    • M08: Create controller program to synchronize multiple VTP emulators
    • FY24Q3
    • M09: Determine appropriate schema for all aspects of monitoring system
    • M10: Establish databases for monitoring system using existing JLab servers.
    • M11: Integrate Hydra as monitoring component.
  7. AOT

Minutes

Attendees: David L., Ayan R., Vardan G., Marco B., Jeng T., Cissie M.

  • Marco
    • Fabio not available today, but is at JLab and can meet in person
    • They have determined that the Alveo boards are not really suitable for steam processing
      • OS used for them is closed and has restrictions on how algorithms are implemented
      • unable to get improved performance over CPU for applications tested under NP-like conditions
      • Looking at alternatives
    • BDX
      • Met with Thia last week. Project will move forward but they are waiting on formal statement from JLab on commitment
      • Full prototype (electronics) expected by Jan. 25 @ JLab
      • Japan collaboration is working on pre-approval. They have some of their own software they are interested in using (unclear on what functionality it covers)
      • Electronics and staff to support it will be supplied by Hall-B
  • Vardan
    • Still working on DPPUI. Nothing new to report
    • Some work done on Data Lake
      • Multiple options tested: Kafka, Redia, xMsg, Chronicle Queues
        • A demo was shown from of each from slowest to fastest
          1. Kafka
          2. xMsg (zmq)
          3. Redis
          4. Chronicle
  • Jeng
    • Looked some into network_exporter for Docker
    • May be limited in ability to use in user space. More investigation needed.
  • Ayan
    • Working on ACAT2024 proceedings paper
    • Need plots from Dec. 2023 packet capture data

Action Items

  1. Finalize ACAT2024 paper