Software Project Quarterly Report Jul-Sep 2001
==============================================

WBS 2.2.1.1, 2.2.1.2, 2.2.1.4 Framework, Architecture and Event Model 

Significant developments in framework and architecture during this
reporting period are as follows.

o A major migration from the SRT configuration management tool to
  the CMT tool was embarked upon during this period. Although not
  nominally a direct responsibility, our previous experience with
  CMT has caused direct involvement in this, with the goal being
  to accelerate deployment under CMT.
o An Object Introspection interface and service for the Data
  Dictionary has been implemented. The interface has been specified
  in conjunction with LHCb.
o The Data Dictionary prototype has been updated to be compatible
  with the most recent ATLAS production release (2.0.2).
o Work has begun on on the Detector Description - with the focus being
  support for alignment within the SCT detector, which will be used
  as a testbed for other detector systems.
o Significant pile-up design work was been done, with a requirements
  document being produced and the prototype scheduled for
  initial implementation in early November.
o An ongoing Grid liaison activity is underway, with the goal being
  to aid the integration of Grid capabilities into Athena. A joint
  testbed has been setup in conjunction with Valerie Taylor
  (NorthWest Univ.) to incorporate monitoring into Athena.
o A cut in US-ATLAS funding has caused LBNL to have to move Chris
  Day, a senior software engineer, to another project. This has
  delayed the Python scripting development and raises as yet
  unanswered questions about the USDP work. The plan is to post
  a job opening for a post-doc who can take take over the Python
  scripting development.

Significant developments in event model during the period are
as follows.

o A data access workshop was held at BNL with some Gaudi developers
  where agreement was achieved on a redesign of this Gaudi/Athena
  component to allow significant performance improvements to be
  achieved.
o Development of an improved back-end for the StoreGate was
  completed and it has been installed.
o The raw data event model was designed and a prototype
  implementation was developed.  It will be used by detector digit
  developers in the coming releases. 
o A conversion service for converting the raw digit in the transient
  store to raw event format has been designed.  A preliminary
  implementation was developed and is under review.  We expect it to
  be available for HLT trigger and online software studies soon. 

WBS 2.2.1.3 Databases and Data Management 

A principal focus of the U.S. ATLAS database effort in this reporting
period was the elaboration of a draft ATLAS database architecture for
the event store.  We had proposed in the previous quarter to deliver a
draft document by the end of the summer.  This milestone was met, with
the draft delivered in mid-September and presented in the database
session of the September ATLAS Software Week.  A phone meeting in
early October and an open session during ATLAS Week have been
scheduled in order to gather input from the wider ATLAS community.  An
additional iteration on the architecture will follow, after which a
review by members of the collaboration and outside experts will be
scheduled. Ed Frank (U Chicago) and Greg Chisholm (ANL) led the
development of the architecture document, with input from many.

Care was taken to propose an architecture that was not biased toward a
specific datastore technology.  A two-day meeting between ATLAS ROOT
experts at Brookhaven and the core database group at Argonne was held
to ensure that this goal was achieved.

Significant U.S. effort during the quarter was devoted to adaptation
of the Objectivity conversion service to the new StoreGate data store
model.  This work continues into the next quarter, as the "new"
StoreGate backend will enter ATLAS software releases in October.
Database support for that new backend is scheduled to be provided one
release later.

User support for the Lund program of Atlfast simulation consumed much
more effort than the database effort should reasonably have been
expected to provide, diverting development time into support for a
product that was intended only to be a demonstration that Athena
could, in fact, use multiple storage technologies.  Effort was also
expended on an investigation of relative storage usage in various
technology products.

The ATLAS move from SRT to the CMT release tool has also been a drain
on effort, albeit a necessary one.  The build infrastructure for
generated source files and for schema gathering and management is
necessarily more complicated than the basic compile-and-link model for
non-database code.

Persistence support for changes to transient generator packages to
conform to CLHEP naming conventions and to move generators to
StoreGate were also delivered by U.S. ATLAS core developers in this
reporting period, although such application-level work is not the
formal responsibility of the core database team and in general will be
done by application developers.

A ROOT conversion service making full use of ROOT I/O capabilities was
developed and introduced into the ATLAS software repository.  It is
intended to replace the simple ('blob' based) generic converter model
used for ROOT persistence in the Lund fast simulation program.  This
service supports the new StoreGate backend, and will enter ATLAS
releases in the final quarter of 2001.  Persistent classes and their
converters have been made for some LAr classes and for Atlfast data
classes.  This work is one component of a ROOT event store
infrastructure that will be essential to meeting the ATLAS milestone
of a datastore technology comparison in Phase II of Data Challenge 1
(mid-2002).  Work is underway to implement a ROOT I/O back end for the
ATLAS Data Dictionary.

Work on the MySQL-based LAr conditions database continued at a low
level.  The API for usage of the database with test beam data was
incorporated in ATLAS software releases. The databases are filled with
constants for test beam runs and used in test beam analysis efforts.

WBS 2.2.1.10 Distributed Data Management and Processing Software  

Development of the Magda (formerly DBYA) distributed data management
system continued as a joint ATLAS/PPDG effort to provide a production
distributed data management system deployed to users by the end of
PPDG/SCIDAC's first year (end June '02).

Several enhancements were made to the file and replica cataloging in
Magda. Scalability tests of Magda cataloging were done; catalog size
was increased from the current stable count of 160k up to ~400k and
then up to 1.5M cataloged files. After minor bugs were fixed the
system performed well at 1.5M files with a lookup performed on the
entire catalog taking ~30sec. Support for several types of file
collections was added.  A Globus replica catalog loader was developed
to migrate the Magda replica catalog content to Globus and evaluate,
but it remains to be tested. Input was given to the replication
requirements document based on Magda and earlier experience.

Support for file replication between distributed sites was added.
The Globus gsiftp tool is used for replication among US ATLAS
grid testbed sites, while scp is used at the moment between CERN and
BNL. A multi-stage automated process moves a file collection (in the
most complex case) out of a source-side mass store into a cache, over
the network into a destination cache, and into a destination mass
store.  The system has so far been used to replicate ~100GB of ATLAS
simulation data between CERN and BNL, and small volumes have been
replicated to other sites.  Cataloging and replication were extended
to support the Castor mass storage system at CERN.

An 'SQL accelerator' was developed to expedite processing of MySQL
commmands from remote client sites. The mechanism speeds up bulk
catalog operations over WANs by orders of magnitude. With the
accelerator, cataloging 1.5M files over a WAN was shown to be
practical.

Deployment of Magda was extended beyond BNL and CERN to ANL and LBNL,
and partially to Boston University.

Development plans for Magda were coordinated with PPDG, GriPhyN and
the CS projects at GriPhyN and PPDG collaboration meetings in
August. Jennifer Schopf now acts as liaison with the CS projects.  The
description and documentation of the system was improved.  Further
information (the documentation page) and a talk is available at
http://atlassw1.phy.bnl.gov/magda/info The system itself is at
http://atlassw1.phy.bnl.gov/magda/dyShowMain.pl

Near term plans include completion of command-line tools providing a
file access interface to production jobs; tools to monitor 
throughput and gather statistics in a production environment;
ATLAS framework (Athena) integration; further integration of Globus
tools (remote command execution, replica catalog); exploration of
other data movers (GDMP, bbcp); and application and testing in ATLAS
Data Challenges commencing in December. Discussions on the application
of Magda within the ATLAS Data Challenges began during the
period. Development of a DC production scenario for simulation data
using Magda also began during the period.

In other work, prototype scenarios for grid-enabled data access from
Athena were investigated.  Two approaches in particular were explored,
one involving registration of files containing event collections with
the Globus replica catalog, the other involving use of GDMP 1.2.2.
The latter approach was exercised on EU Data Grid testbed nodes in
Geneva and Milan, using the ATLAS fast simulation program Atlfast
running under Athena, with the object database product Objectivity/DB
as the underlying storage technology.  This work was described at the
CHEP'01 conference in Beijing.

WBS 2.2.2 Simulation and Reconstruction Software 

Work was underway during the period to prepare the ATLAS detector
description package for Data Challenge 0 later this year: Reversal of
SCT tilt angles to minimise cluster size; introduction of modular TRT
barrel design; introduction of fully realistic field; optimisation of
cryostat description; update of calorimeter design; and update of muon
shielding.

Several improvements in simulation infrastructure were initiated,
directed principally at readiness for the first two ATLAS Data
Challenges, the first of which begins in November. Development of a
simulation production infrastructure for DC0 began, supporting
processor farm-based production and central archiving of results.
This effort is being coordinated with grid data management efforts. A
new pile-up facility to support up to a thousand pile-up events was
developed. Support for the simulation of heavy ion collisions in the
ATLAS detector was implemented.

A small effort has started to investigate the possibility of
integrating Geant4 into Athena. Although this was the original model,
the Simulation group has decided to develop their own framework
('Fads/Goofy'), decoupled from Athena. This is viewed with great
concern by David Quarrie (chief architect) and US ATLAS management,
and was the focus of significant feedback during the last PCAP
Review. This low level effort has been extremely successful, having
resulted in a level of integration where the existing Athena-based
generators may be used to generate particles, be simulated using
Geant4, and the output hits may be stored in the Athena StoreGate
store. All current sources of detector geometry (ATLAS-standard XML
files, Fads/Goofy XML files and G4 classes) are supported, as are
physics lists and visualization. The Nevis group have acted as one
early adopter, with the SCT and muon systems being used as other
testbeds. Discussions have begun with the ATLAS Computing Coordinator
as to how best announce and broaden this development.

Specialized detector studies continue to utilize the Tier 1 Center
heavily, particularly ATLAS radiation background studies being
performed by Mike Shupe (Arizona), running typically on 35 to 50
CPUs. Shielding design activity is building to a crescendo because the
major shielding pieces soon have to be committed to production. The
activity consists of a lot of geometry editing to look at new options,
and then producing statistics of roughly 50k events per option, fully
tracked down to thermal neutron energies.  The Radiation Task Force
will be meeting next during the October ATLAS Week, and has the goal
of producing a final report on the overall shielding design by
December.

WBS 2.2.4 Software Support and QA/QC 

The US-developed and operated facility that produces nightly builds of
ATLAS software based on the most recent tagged versions of packages
was adapted for the new CMT tool. At present the facility is operating
at the BNL Tier 1 Center and has been useful in fixing problems
related to the transition to CMT.  After fixing CMT problems, the
facility will be moved to CERN computers and used to provide feedback
to developers throughout ATLAS on newly introduced software bugs and
incompatibilities. Development of additional testing and QA software
to be integrated with the nightly builds must await the completion of
the transition to CMT due to limited manpower.

New Atlas software releases are being promptly installed at the BNL
Tier 1 Center, usually within one day after CERN installation.  The
environment setup scripts for users of the Tier 1 Center are being
adapted to CMT. The intent is to complete this work in the next
quarter.

A new 'package manager' (pacman) was developed at Boston University to
support the easy distribution of ATLAS software and needed third party
software components to remote sites. During the period the tool was
deployed to a number of early testers and a support infrastructure for
pacman-based distribution of software throughout the US was developed.
Deployment of the system to users is expected in the next quarter.

Project Management

International ATLAS completed its migration of computing project planning
to XProject. US and International ATLAS planning now share common data
and a common tool, with different 'projections' out of the data 
accommodating differences between US and International ATLAS. CMS and
PPDG also requested and received access to XProject to evaluate it for
their purposes. Minor XProject modifications were made to support
multiple non-ATLAS projects.

Summary of Major Milestones and Deliverables

WBS 2.2.1.2 

  Lund software release was completed (Jul)
  StoreGate-based back end for the ATLAS event data model available (Sep) 

WBS 2.2.1.3

  Database architecture document released (Sep) 
  CMT release tool support for database software (Sep)
  New ROOT I/O service completed (Sep)
  ROOT I/O support for the new StoreGate transient store (Sep)

WBS 2.2.4

  Migration of nightly builds and US release support to CMT (Sep) 

Forthcoming Milestones and Deliverables

Many of the forthcoming milestones and deliverables are driven by the
Data Challenge schedule. Data Challenge 0 (primarily a software
completeness and continuity test involving a few 100k events) takes
place in November and December. Data Challenge 1 (10M events for high
level trigger and physics studies and new software evaluation,
including evaluation of Objectivity and ROOT based event persistency
mechanisms) takes place February-July 2002.

WBS 2.2.1.2

  StoreGate version of Atlfast fast simulation in production (Nov)
  Service management restructuring deployed (Oct) 
  Physics analysis output/binding to JAS analysis tool deployed (Oct)
  Pile-up support prototype (Nov)
  Detector description prototype (Nov)
  DC0 release of Athena (Nov)

WBS 2.2.1.3

  Interval-of-validity based retrieval infrastructure prototype (Oct)
  Full compliance of database software with gcc compiler (Oct)
  Objectivity support for the new StoreGate transient store (Nov) 
  ROOT persistency support implemented in the ADL back end (Nov)
  Migration to Objectivity 6 (Nov)
  Prototype database support for pile-up (Nov)
  Meta information management (bookkeeping) prototype (Nov)
  Conversion service support for input to the detector data service (Dec)
  Collection registration prototype (Dec)

WBS 2.2.4

  Migration of nightly build service to CERN (Oct)
  Activate developer notifications for CMT-based nightly builds (Nov)
  Incorporation of code checker in software builds (Nov)