From: Reiner Hauser [rhauser@fnal.gov]
Sent: Tuesday, July 23, 2002 6:31 PM
To: Giuseppe Mornacchi
Cc: Saul Gonzalez; Bob Dobinson; serguei.kolos@cern.ch;
beniamino.di.girolamo@cern.ch; Benedetto Gorini
Subject: AWG ideas


Hi,

Here's a copy of what I noted down so far. There's quite some overlap
with other people, so maybe that helps us to prioritize.

cheers, Reiner

-----------------------------------------------------------------

- Look at outcome of review and use recommendations as
  starting point for dicusssions.

- Invite mgmt team, subsystem leaders, users via representatives
    to bring up topics
    decide then if action necessary

Architecture/Global Issues
===========================

- Overall system view and interfaces between systems
    - which interfaces are well-defined ?
    - which need proper definition/documentation ?

- Clarify Physics/non-physics runs, use cases
   how do e.g. calibration tasks interface with
   DataCollection/EF ?

- Clarify monitoring requirements

   ROD crate monitoring
   ROS monitoring
   Full event monitoring (with help of DFM ? has L1 + L2 decisions) ?
   Monitoring after EF ?

- Overall fault tolerance

    specify what happens as each component fails
    how does it affects the run
    what is the impact on the overall system performance
    what is the loss in physics

Interfaces
==========

- Clarify storage of data with offline

- DC/EF: E.g. can the SFIs be integrated in the EF nodes ?

Common issues:
=============

- Run Control/Supervision of processors "farms"
    e.g L2PU, EF, SFI, ROS

- Configuration + Conditions + other databases

- Common base libraries
    Maybe this can be off-loaded to the
    people in the "integration" boxes ?

- Testbeds ??
    Make sure that tests incorporate people from OnlineSW + Dataflow together
    Common OnlineSW + Dataflow + HLT tests ?
    Regular integration tests ?

-- 
Reiner Hauser
Email: rhauser@fnal.gov
Tel:   (630) 840 8634
Fermilab, PO Box 500, MS 352, Batavia, IL 60510-0500