The Bioinformatics and DAIT Project BDWorld Exemplar

1
The Bioinformatics andDAIT Project BDWorld
Exemplar

• Shirley Crompton, Brian Matthews (CCLRC)
• Alex Gray, Andrew Jones, Richard White (Cardiff
  University)

2
Overview

• BioDA
• the project
• BDWorld
• the project
• architecture
• resource usage
• BDWorld case study
• the prototype
• early experiences

3
BioDA Goals

• Independent Evaluation of OGSA-DAI
• Is it useful for data-intensive bioinformatics
  GRID applications
• Find out what the community needs
• BioDA Workshop (Dec 2004)
• how they could leverage OGSA-DAI
• Case studies based on BBSRC eScience pilot
  projects
• BDWorld, eHTPX .
• OGSA-DAI Product Improvement
• Feedback to the DAIT Team
• Knowledge Dissemination
• Evaluation Reports
• Publications/Presentations
• Workshop on OGSA-DAI for the bioinformatics
  eResearch community
• September 2005

4
BDWorld Problem Solving Environment (source
BDWorld)
5
BDWorld System Architecture(source BDWorld)
6
BDWorld Example Usage (source BDWorld)
7
BDWorld Thematic Data Resources Key Issues

• geographically distributed and autonomous
• heterogeneous in structure and data standards
• mainly read via HTTP/XML protocols using custom
  wrappers
• SQL queries are limited to the EBI EMBL store and
  BDWorld cache databases
• potentially resource-intensive to harvest
• a single taxa name may resolve into a large
  number of accepted taxon names
• same query repeated on different data collections
• uniform resource access/invocation mechanism via
  BDWorld Grid Interface
• InvokeOperation(ResourceHandler,Operation,DataColl
  ection)

8
BDWorld OGSA-DAI Prototype
OGSA-DAI R5 GDS
3. Invoke wrapper
Wrapper Module
BDWQueryActivity
2. Create GDS and query
6. Download url
7. url
deliverFromURL(url)
5. Download URL
OGSA-DAI Client
deliverFromURL(xsl)
8. XSL transform to BDW format
XSLTransform
9. To WF unit
deliverToURL/GFTP
9
BDWorld Whats next (1)
OGSA-DAI R5 GDS
8. XSL transform to BDW format
XSLTransform
9. Copy output
cloneOutput
10. Update BDW cache
10. To WF unit
Postgres
deliverToURL/GFTP
sqlBulkload
10
BDWorld Whats next(2)
OGSA-DAI R5 GDS
8. XSL transform to BDW format
9. integrate output
mergeOutput
OGSA-DAI Client
10. copy output
cloneOutput
Postgres
11. To WF unit
11. Update BDW cache
deliverToURL/GFTP
sqlBulkload
11
Comments

• What we like
• OGSA-DAI easy to deploy and customise
• Minimum changes to existing wrappers
• Reuse XSLs and XSDs
• leverage grid data transport mechanism
• What we are concerned about
• scaling
• multiple web DB query activities
• xsltTransform memory usage
• synchronisation high fluctuation in web DBs
  response time
• Unorthodox usage not RDBMS/XMLDB/file resources

12
BDWorld(http//www.bdworld.org)The BioDa
Project(http//isegserv.itd.rl.ac.uk/BioDA)