LCG Tier 2 and DOSAR

1
LCG Tier 2 and DOSAR

• Pat Skubic
• OU

2
Outline

• Introduction to LCG
• List of LHC Tier 2 Centers
• Status of ATLAS SW Tier 2
• Questions for Discussion
• Action Items?

3
(No Transcript)
4
Development Areas

• Applications
• Fabric
• Grid Deployment
• Grid Technology
• ARDA-A Realization of Distributed Analysis for
  LHC

5
Applications
6
(No Transcript)
7
Grid Deployment Organization
8
Grid Technology Activities

• Overall Requirements
• Technology Tracking
• Middleware Development
• Middleware Recommendations

9
Overall Requirements

• Keep a clear view of the emerging requirements
  for the LHC computing including batch and
  interactive analysis. The HEPCAL document is the
  starting point for this.
• Track the grid activities within the experiments
  and help disseminate the experiences gained
• Disseminate the need for high quality
  infrastructure services and work towards their
  technical realization
• Work towards a systems architecture for grid
  technology to support the project

10
Technology Tracking

• Monitor closely the technology evolution and put
  into practice the technologies as they mature in
  production. Work with the experiments to perform
  structured tests will be required. This may also
  involve working with industry and other sponsored
  developments to prototype and test solutions as
  they emerge
• Study the system architecture issues as
  experience is gained in order to steer the
  technology evolution

11
Middleware Development

• Seek strategic advice from the individuals
  driving the multiple projects producing the core
  technologies
• Work towards one set of common and complementary
  projects encompassing the efforts underway. This
  involves participating in the plans for funding
  future projects
• Identify key developments that are required for
  the LHC computing grid that are not developed as
  part of existing projects and promote their
  development.

12
Middleware Recommendations

• Recommend a coherent technology suite to be
  deployed to regional centers. This has to be done
  taking into account the matching with the LHC
  requirements
• Develop a plan for the evolution of this suite as
  new products become available and as new
  technologies mature
• Work towards developing a long term support
  structure for the middleware technology

13
ARDA

• The basic middleware development team. Joint
  European-US team, part of EGEE project, but
  including participation from VDT.
• At least one team in each experiment, that takes
  part in adapting the analysis system of their
  experiment to use the basic middleware.
• Providers of common HEP tools, such as POOL,
  SEAL, ROOT, GANGA and PROOF.
• Users within experiments that will provide
  immediate feedback to the developers.

14
ARDA (cont)

• Service providers from regional centers who will
  be involved in deploying early versions of the
  middleware and analysis systems.
• The participation of teams from other sciences,
  such as those participating in EGEE, is welcomed.

15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
SW Tier 2 Status

• Tier 2 Center
• ATLAS management selected OU, LU, UT-Arlington,
  as a Tier 2 Center (Southwest Tier 2 Center)
• Collaboration started on grid software
  development (PanDA)
• OCHEP linux cluster will be deployed this year
• Efforts will be made to increase NSF funding
• Physics Analysis OCHEP collaboration will begin
  on physics topics, building on Tevatron
  experience
• b tagging
• - t tagging
• top physics
• Higgs/SUSY searches

20
Production and Distributed Analysis (PanDA)
21
Questions for Discussion
22
Questions from Sergio

• How would we fit in the general structure of LCG?
• Will DOSAR VO become a distributed Tier 2 center?
• Which would be our associated Tier 1? Fermilab?
• What should we do to comply with Atlas and CMS
  requirements?
• How will we share our resources between D0 and
  LHC?

23
Should DOSAR collaborate with other Tier 2
Centers?

• CMS Tier 2C?
• Brazil Tier 2 Federation?

24
Should DOSAR participate in LCG software
development or deployment?

• Common ATLAS-CMS projects?

25
Should DOSAR software efforts be part of larger
collaborations?

• OSG proposal?
• PanDA and DIRAC?

26
DIRAC- Distributed Infrastructure with Remote
Agent Control
27
DIRAC (cont)

• It has been designed to be lightweight, robust,
  and scalable. This has been achieved by using a
  service oriented architecture and utilizing
  existing software for key components 99 of
  DIRAC is written in Python, which provides a high
  performance, object oriented scripting language
  for rapid development, and maintenance.
• MySQL database is used for persistency. Services
  make connections to various DBs to manage job
  state, monitoring information, and meta-data
  concerning grid files.
• Condor ClassAds and Matchmaking are the basis of
  the scheduling strategy. This makes the best
  match between an available resource and a queued
  job.
• Jabber Instant Messaging has been used to
  provide asynchronous, real-time, fault-tolerant
  monitoring and control capabilities to DIRAC
  Agents, Services, and Jobs.

28
Action Items

• To be decided after discussion