GriPhyN and iVDGL Progress

1

• GriPhyN and iVDGL Progress

Paul Avery University of Florida http//www.phys.u
fl.edu/avery/ avery_at_phys.ufl.edu
CMS PMG Meeting FermilabApr. 21 2003
2
NSF Review of GriPhyN 1/2003

• Major mid-term review Jan. 29-30
• 5 NSF people, 6 reviewers (3 CS, 3 physics)
• High visibility of GriPhyN largest NSF/ITR
  program funded
• GriPhyN mentioned in many NSF announcements,
  reports
• Oral report uniformly excellent
• Research, E/O, impact, collaboration
• Door opened for supplemental funding for VDT
• Written report not yet made available
• Expect mid-term review of iVDGL late 2002 or
  early 2003

3
Several VDT Releases in 2002-03

• Continuing feedback from users
• Testbeds, individuals, EDG, LCG
• Result much simpler installation, better
  post-install configuration
• VDT 1.1.7 released recently
• Improvements in usability
• Globus 2.2.4, Condor, Condor-G 6.4.7
• GLUE schema
• Fault Tolerant Shell 1.0
• EDGs CRL-Update mkgridmap
• DOE EDG CA Certificates
• Chimera Virtual Data System 1.0.1
• UW-Madison team increased by one person
• Alain Roy (GriPhyN)
• Carey Kireyev (iVDGL)

From NMI next time
VDT support person
4
VDT PACMAN News

• Recent VDT deployments
• US-Atlas Testbed
• US-CMS Testbed
• Non-HEP sites (LIGO, SDSS)
• WorldGrid and all GriPhyN demos at SC2002
• VDT 1.1.8
• Globus
• Very latest GT 2.2 release
• Condor-G
• Condor 6.5.1 (if ready if not, 6.4.7)
• New DAGMan supporting just-in-time grid
  scheduling
• Many robustness and scalability improvements

5
VDT Plans and Challenges

• Integration with NMI for core components
• Globus, Condor
• Timeline for new internal products
• Wisconsin CS products
• New schedulers, etc.
• Other work
• Products from external software contributors
• Demands for hardening and support services
• Easy auto-config for non-expert or lite users
• Longevity of VDT
• Short term GriPhyN supplement?
• Long term NSF Cyberinsfrastructure program?

6
New iVDGL Collaborators

• New experiments in iVDGL/WorldGrid
• BTEV, D0, ALICE
• New US institutions to join iVDGL/WorldGrid
• Many new ones pending
• Participation of new countries (different stages)
• Korea, Japan, Taiwan, Brazil, Romania, Australia,
  
• Russia and China discussions just started

7
US-iVDGL Sites (Late Spring 2003)

• Partners?
• EU
• CERN
• Brazil
• Australia
• Korea
• Japan

8
US-LHC Testbeds

• Significant Grid Testbeds deployed by US-ATLAS
  US-CMS
• Testing Grid tools in significant testbeds
• Grid management and operations
• Large productions carried out with Grid tools

9
ATLAS Simulations on iVDGL Resources
Joint project with iVDGL
10
US-CMS Testbed
Wisconsin
Korea
MIT
Taiwan
CERN
Fermilab
Russia
Caltech
FSU
UCSD
Florida
Rice
FIU
Brazil
11
Commissioning CMS Grid Testbed (2002)

• A complete prototype (fig.)
• CMS Production Scripts
• Globus, Condor-G, GridFTP
• Commissioning Require production quality
  results!
• Run until the Testbed "breaks"
• Fix Testbed with middleware patches
• Repeat procedure until the entire Production Run
  finishes!
• Discovered/fixed many Globus and Condor-G
  problems
• Huge success from this point of view alone
• but very painful

12
CMS Grid Testbed Production
13
Production Success on CMS Testbed

• Results
• 150k events generated in Fall 2002 1.5 weeks
  continuous running
• 1.5M event completed on larger testbed Winter
  2002 8 weeks
• Recently, 300k events generated in Spring 2003
  for testing

14
Chimera Virtual Data System (GriPhyN)

• Virtual Data Language (VDL)
• Describes virtual data products
• Virtual Data Catalog (VDC)
• Used to store VDL
• Abstract Job Flow Planner
• Creates a logical DAG (dependency graph)
• Concrete Job Flow Planner
• Interfaces with a Replica Catalog
• Provides a physical DAG submission file to
  Condor-G
• Generic and flexible
• As a toolkit and/or a framework
• In a Grid environment or locally

VDC
AbstractPlanner
XML
XML
VDL
DAX
ReplicaCatalog
ConcretePlanner
Virtual data CMS production MCRunJob
DAG
DAGMan
15
iVDGL US-CMS ProtoTier2 Centers

• Sites at UF, Caltech and UCSD
• UF 40 dual nodes, 6 TB RAID (GriPhyN match)
• Caltech 40 dual nodes, 4-6 TB RAID (some iVDGL)
• UCSD 40 dual nodes, 4-6 TB RAID (some iVDGL)
• Used for major CMS productions
• Forms current IGT and future DPE
• More purchases planned with iVDGL funds
• 1.4M for Year 1 funds (most not spent)
• Active discussions with Dell, IBM to get deals
• White box vendors also look attractive

16
Creation of WorldGrid

• Joint iVDGL/DataTag/EDG effort
• Resources from both sides (15 sites)
• Monitoring tools (Ganglia, MDS, NetSaint, )
• Visualization tools (Nagios, MapCenter, Ganglia)
• Applications ScienceGrid
• CMS CMKIN, CMSIM
• ATLAS ATLSIM
• Submit jobs from US or EU
• Jobs can run on any cluster
• Demonstrated at IST2002 (Copenhagen)
• Demonstrated at SC2002 (Baltimore)

17
WorldGrid
18
WorldGrid Sites
19
iVDGL WorldGrid and US-CMS

• A US-CMS operated Grid
• US-CMS-iVDGL resources with WorldGrid components
• Enables US-CMS to use non US-CMS resources
• TeraGrid
• HPC farm at Florida gt 1000 processors by 2006
• etc.
• Leverages experience from other iVDGL experiments
• Consultation of WorldGrid developers and US-CMS
  (4/22)
• Consultation with LCG
• Establish flexibility of LCG installation (too
  intrusive now)
• WorldGrid re-packaging of LCG components (PACMAN
  package)
• Preserve local site autonomy!!!

20
An Inter-Regional Center for High Energy Physics
Research and Educational Outreach (CHEPREO) at
Florida International University

• Status
• Proposal submitted Dec. 2002
• Presented to NSF review panel
• Project Execution Plan submitted
• Funding in June?

• E/O Center in Miami area
• iVDGL Grid Activities
• CMS Research
• AMPATH network (S. America)
• Intl Activities (Brazil, etc.)

21
Large ITR Proposal 15M
Will use iVDGL resources, extend GriPhyN
22
UltraLight Proposal to NSF

• 10 Gb/s network
• Caltech, UF
• FIU
• SLAC, FNAL
• MIT, Michigan
• Intl partners

Very Long Base Interferometry
Grid Projects(iVDGL, PPDG, EDG)
HEP(CDF, D0, BaBar, CMS, ATLAS)
Distributed Radiation Oncology
23
GLORIAD

• New 10 Gb/s network linking US-Russia-China
• Plus Grid component linking science projects
• H. Newman, P. Avery participating
• Meeting at NSF April 14 with US-Russia-China
  reps.
• HEP people (Hesheng, et al.)
• Broad agreement that HEP can drive Grid portion
• More meetings planned