1 / 15

GRID INTEROPERABILITY USING GANGA

KISTI-CCIN2P3 FKPPL Workshop December 1, 2008. GRID INTEROPERABILITY USING GANGA. Soonwook Hwang (KISTI) YoonKee Lee and EunSung Kim (Seoul National Uniersity). INTRODUCTION. Multiple Grid Infrastructures available EGEE (gLite), PRAGAM, OSG, TeraGrid (Globus), DAISA(Unicore), etc.

eliana-odom
Télécharger la présentation

GRID INTEROPERABILITY USING GANGA

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. KISTI-CCIN2P3 FKPPL WorkshopDecember 1, 2008 GRID INTEROPERABILITYUSING GANGA Soonwook Hwang (KISTI) YoonKee Lee and EunSung Kim (Seoul National Uniersity)

  2. INTRODUCTION • Multiple Grid Infrastructures available • EGEE (gLite), PRAGAM, OSG, TeraGrid(Globus), DAISA(Unicore), etc. • Users have to learn commands or tools for each grid infrastructure • Not easy to use resources across multiple Grids • Our Goal • High-level Tools that enable the use of as many as resources possible from multiple grid infrastructures without having to know the details of each grid middleware • Our Approach • Chose to use the GANGA as a high-level tool for job execution management exploiting multiple Grid infrastructures • Chose to use the WISDOM as a target application

  3. GANGA • easy-to-use user interface for job submission and management • Implemented in Python: Interactive shell or Script files • Use python syntax to submit a job • provide a plug-in mechanism for multiple execution back ends (e.g., Grid, Batch and Local) and applications (e.g., ROOT, GAUDI and Athena) • easily extended and customized to meet the needs of different user communities • Currently, support only the LCG/gLite backend for Grid, not Globus backend ROOT ATLAS GAUDI Athena G a n g a LCG/gLite Backend LCG/gLite Cmd or Lib LHCb PBS or SGE Backend PBS or SGE Cmd or Lib PBS or SGE Local Backend Local Cmd or Lib

  4. New Plug-ins Development in GANGA • Have developed new plug-ins to GANGA to support: • AutoDock applications • GridWay backend • InterGrid backend BioMed LCG/gLite Backend LCG/gLite Cmd or Lib FKPPL AutoDock G a n g a PRAGMA InterGrid Backend GridWay Backend GridWay Cmd or Lib ROOT PBS or SGE GAUDI PBS or SGE Backend PBS or SGE Cmd or Lib Athena Local Backend Local Cmd or Lib

  5. New Plug-ins Development in GANGA (cont’d) • AutoDock Application • Customize configuration for running autodock applications on the grid • LigandSplitter is providedfor splitting ligand-protein docking jobs on the grid • GridWay Backend • Have access to Globus resources • InterGrid Backend • Have access to both LCG/gLite and Globus resources • Intelligent resource selection • Based on load on Grid

  6. Implementation Issues • Authentication • Use VOMS proxy • AutoDock Execution • Autodock shell script developed in WISDOM • Rely on globus-url-copy • LCG: globus-url-copy – OK • Gridway: globus-url-copy – Failed (No job delegation support) • Our approach • “Staging in” technique • LCG: Inputsandbox – size limitation problem • Gridway: Inputsandbox – No size limitation • Ganga provide a solution for LCG size limitation. • Upload large files to SE and download from SE to CE. • Thus, we use inputsandbox with Ganga functionality.

  7. EXAMPLE • One AutoDock Job

  8. EXAMPLE • One AutoDock Job

  9. EXAMPLE • Multiple AutoDock Jobs (for statement)

  10. EXAMPLE • Multiple AutoDock Jobs (for statement)

  11. EXAMPLE • Multiple AutoDock Jobs (for statement)

  12. EXAMPLE • Multiple AutoDock Jobs (LigandSplitter)

  13. EXAMPLE • Multiple AutoDock Jobs (LigandSplitter)

  14. EXAMPLE • Multiple AutoDock Jobs (LigandSplitter)

  15. EXPERIMENT • Performance Evaluation • gLite supports bulk (faster) submission • Splitting jobs may be more efficient than submitting bunches of individual jobs

More Related