Rob, Wei,

 Please find a first draft of the agenda below.

 

--

   Michael

 

The concept is based on the idea to divide the meeting into 3 parts

 

1.)     Day 1: Overview talks and discussions

2.)     Day 2: Hands-on with tutorials and demos

3.)     Day 3: Site reports and summary discussions & planning

 

Day 1

--------

Overview talks and discussions

 9:00    LHC and ATLAS Status  (J. Shank)

 9:30    Facilities Status & Plans (M. Ernst)

10:00   ?

10:30   Coffee

11:00   Facilities Integration Program (R. Gardner)

12:00 – 14:00    Lunch

Production & Physics Analysis in ATLAS

14:00   Production & Analysis process and functional requirements at U.S. ATLAS Tier-2 and Tier-3 centers

15:00   Running Physics Analysis on local and grid computing resources – some examples

 

Day 2

--------

For day 2 I propose to cover for themes. The agenda details have to be worked out.

 

Fabric Hardware (it was obvious how interested attendees in Bloomington were in this area)

-          Compute Servers

-          OS choices (32/64 bit, etc)

-          Installation management

-          Disk Subsystems

o        Configuration, resilience, performance, etc

… suitable in the context of Tier-2 & Tier-3 centers,

Sizing of Tier-3’s relative to physics analysis tasks and expected funding profile

 

Fabric Services

-          Batch (Condor, PBS) functionality and configuration examples

-          Storage Management Systems for Tier-2 and Tier-3 centers

o        Xrootd        (Ofer Rind, BNL)

o        dCache      (Neha, FNAL and OSG)

       Hands-on installation and configuration

-          Data Transfer Optimization

o        Report from WG (Shawn)

§         Findings / Analysis

§         Mechanisms to improve the performance

 

OSG Grid Middleware

-          Setting up a CE – Hands-on w/ installation & configuration

-          Installation and usage of client tools

-          The OSG Information System – how it works and what’s in there

 

(US) ATLAS Services

-          Installing and using

o        Distributed Data Management Tools

§         DQ2 and Pandamover

§         LRC

o        PanDA Workflow Management System

o        ATLAS Software

§         Demo for PanDA based installation

-          Running the entire Physics Analysis chain w/ PanDA, PROOF & Xrootd

o        A full demonstration based on a “real” analysis scenario  

 

The hands-on sessions will require some pre-requisites, i.e. some machines at either/and/or BNL, UC, SLAC we can use to do the installations and demos.

 

 

9:00 – 11:20     Site reports

11:20 – 13:00   Summary Discussion & Planning

 

Please let me know what you think and add/change/remove as you see things fit.

 

Thanks,

Michael