Paul%20Groth,%20Simon%20Miles,%20Luc%20Moreau

1
Paul Groth, Simon Miles, Luc Moreau
2
Outline

• Process Documentation for Provenance
• Power of the P-Structure
• P-assertion Recording Protocol
• PReServs Functionality
• Performance
• Pitch ?

3
Provenance

• The Provenance Question
• Lots of definitions
• Boil it down to a question.
• What is the process that led to a particular
  result?
• How do we answer this question?
• Search through documentation.

4
Documentation

• Process Documentation
• encompasses all other documentation
• SOA based model of process
• Actors communicate via message passing
• Actors make ASSERTIONS to document process.
  Termed p-assertions.
• How to organise these p-assertions

5
P-Structure
6
P-Structure View
7
Benefits

• Domain independent queries
• That are provenance specific
• P-structure is a shared logical organisation of
  p-assertions
• Does not prescribe how p-assertions are exactly
  stored in an implementation.

8
PReP

• Introduces the Provenance Store
• A Separate entity for maintaining process
  documentation
• PReP specifies how an actor can communicate with
  the Provenance Store.
• PReP has a number of nice properties.
• Statelessness
• Idempotence
• Terminiation

9
An Implementation

• What is PReServ?
• A Web Services implementation of a Provenance
  Store
• Implements
• PReP for recording
• XQuery for querying
• Provides libraries and wrappers for making
  applications provenance aware.

10
PReServ Implementation Diagram
Provenance Store
WS Calls
Java Calls
11
Implementation cont.

• Caching mechanism to improve performance
• Berkeley Java Database 2.0
• No setup required
• Completely Transactional

SOAP Msg
SOAP Msg
Dispatcher
Store Plug In
Query Plug In

Backend Store Interface
Java Object Database
Memory

12
Requirements

• Apache Tomcat 5.0
• Apache Ant 1.6.2
• Java 1.5 (1.4 supported with some help)
• Pure Java, tested on
• Windows
• Mac OS X
• Debian Linux

13
Evaluation Deployment

• Protein Compressibility Experiment
• HPDC05
• Workflow runs under VMWare
• deployment consistency
• ease of development
• Workflow is executed on one machine
• PReServ runs on another machine
• Version 0.1.5 of PReServ

14
Record Performance
15
Query Performance
16
Applications
17
Conclusion

• The p-structure allows for domain independent,
  provenance specific queries using XQuery.
• Both recording and query times are linear
• PReServ has a extensible architecture allowing
  for further functionality to be easily added.

18
Download!

• Try it out!
• Download PReServ 0.2
• The AHM release ?
• Released under Open Source MIT License
• www.pasoa.org
• Click software
• Contact us, we will try to help you make your
  application provenance-aware.

19
Configuration

• Redhat Linux 9.1 on VMWare on Windows XP
• Pentium P4 2.8 GHZ 1.5 GB RAM
• PReServ on another machine
• Database backend Berkley JDB
• 100 Mb local ethernet