Online Integration Outline for May 2008

From GlueXWiki
Jump to: navigation, search
  • Scope and overview of online systems
    • almost everything typical of experiments this size, with some important exceptions
    • trigger and data rates
    • number of detector channels, front-end boards, HV and LV channels, crates, hardware controls channels, monitoring hist, computers, network connections, etc.
    • planning for two-person shifts


  • Foundations
    • software
      • cMsg pub/sub interprocess communication system
      • CODA DAQ toolkit - ROC, EMU, ET system, run control
      • open-source standards: CAN, I2C, OPC, SNMP, EPICS CA, etc.
      • manufacturer-supplied software in some cases
      • standard languages and facilities: C++ + STL + Boost, Java, Perl/Python/Ruby, databases, web browsers and servers, etc.
      • custom software to integrate above and for specialized applications
    • hardware
      • custom detector hardware, by JLab, collaborators, contractors, or refurbished from previous experiments
      • custom readout electronics and trigger system, by JLab and collaborators
      • commercial hv and lv systems, crates, computers, networking hardware, etc.
      • commercial sensors, actuators, gas system components, Allen-Bradley PLC, etc.
      • some custom-built commercial devices (goniometer)


  • Trigger/DAQ
    • overall trigger design
    • adapting CODA toolkit to Hall D
    • implementation plans for low luminosity prototype farm
    • event transfer to silo
    • scaling to high luminosity and full L3 farm


  • Controls
    • slow controls
      • architecture - 3-tier, functional view vs implementation view
      • strategy - control loops, PLC usage, CAN, I2C, SNMP, OPC, etc.
      • component choices (tentative)
      • interoperability with accelerator control system
    • experiment controls
      • run control using AFECS, inclusion of non-CODA components
      • generic high-level state machines


  • Monitoring
    • detector
      • via slow controls system
      • via data stream when running
    • event data quality - via (prototype) farm
    • archive/display system - MonAlisa, other possibilities
    • event display


  • Alarm system
    • design goals
    • epics, LASER, cMsg/CLAS, A-B, others


  • Misc
    • networking
    • databases - MySQL or PostgeSQL
    • run and shift bookkeeping
    • e-log - CLAS, Accelerator div, CRL or other public domain, write one yet again, etc.