From: Bob Dobinson [Bob.Dobinson@cern.ch]
Sent: Tuesday, July 23, 2002 7:26 AM
To: Giuseppe Mornacchi; serguei.kolos@cern.ch; beniamino.di.girolamo@cern.ch; Benedetto Gorini; reiner.hauser@cern.ch; saul.gonzalez@cern.ch
Subject: Re: FW: AWG ideas
Here are notes I made of the meeting Giuseppe and I had last week about the role of the architecture working group. The list is unstructured and brief. 
  
  a.. Follow a bit from the dectector through to mass store/off line analysis from the point of view of the Sw model and HW components. 
  b.. Go through various operational scenarios, commissioning the detectors and experiment, experiment startup, thunder storms, maintenance etc 
  c.. What should go on guraranteed power supplies? 
  d.. Compile lists of components, interfaces, standard parameters and rates. HW and SW. This can help produce a costing model. 
  e.. Geographic layout of all components of te experiment, this will effect the architecture. 
  f.. Network needs of the on line SW and ROD crates 
  g.. Relationship to off line and GRIDS. 
  h.. Our contribution to the TDR. 
  i.. Fault tolerance, HW and SW, what are we prepared to pay for, what is essential what is considered a luxury? 
  j.. Identify unresolved problems other people should target, eg. rack cooling 
cheers BOB 
Giuseppe Mornacchi wrote: 
   From Serguei... 
  -----Original Message----- 
  From: kolos@pcatd88.cern.ch [mailto:kolos@pcatd88.cern.ch]On Behalf Of Kolos 
  Serguei 
  Sent: Monday, July 22, 2002 3:55 PM 
  To: Giuseppe Mornacchi 
  Subject: AWG ideas 

  Hi Giuseppe 

  Here are some ideas about the TDAQ Architecture group. I agree with 
  the Livio's proposal about a general responsibility if the group - 
  "Coordination 
  of definition of the overall TDAQ architecture including major system 
  elements 
  and their relationships and interfaces". 
  In addition to that here are some more specific issues which in my opinion 
  can be 
  tackled in the context of the AWG: 
      1. the TDAQ operational model including aspects related to changing the 
  run 
          mode (e.g. from physics to calibration), run number, etc. 
      2. the TDAQ Partition operational model including issues related to a 
  splitting 
          and joining partitions online, defining partitions hierarchy, etc. 
      3. the diagnostics and error recovery in the TDAQ system. It seems a 
  distributed 
          diagnostics and error recovery scheme is one of the possible 
  solutions for 
          such a large system as the ATLAS TDAQ. In this case a distributed 
  model 
          shall be defined and agreed among all the TDAQ (sub)systems. 
      4. the general fault tolerance model for the TDAQ. This may include a 
  set of scenarios 
          which describe what happen with some of the TDAQ (sub)systems if one 
  or 
          more of the others encounter errors. In addition those scenarios can 
  specify 
          how to resolve the faults and which (sub)systems are responsible. 
  Here there is 
          a tight reference to the point 3 of this proposals. 

  Cheers, 
  Sergei

-- 
Bob Dobinson 
EP Division, 
CERN, 
1211 Geneva 23, 
Switzerland 

Email: Bob.Dobinson@cern.ch 
Tel office:  +41 22 767 3066  (forwards to mobile after 6/8 rings) 
Mobile: +41 79 201 4236