Provenance Challenge

Lineage For JOpera

Participating Team

• Short team name: ETH Zurich
• Participant names: Thomas Heinis
• Project URL:

Differences from First Challenge

Lineage for JOpera did not participate in the first part of the provenance challenge. It is a relatively new tool used to disseminate the execution related data stored when executing processes/workflows in JOpera. JOpera is a rapid service composition tool offering a visual language and autonomic execution platform for building distributed applications out of reusable services, which include but are not strictly limited to Web services.

The first part of the provenance challenge executed in JOpera:

Provenance Data for Workflow Parts

Challenge data

The provenance data in Lineage for JOpera is stored in a database (any standard DB will do, we are using Postgres). The files included here are therefore both, plain dumps of the Postgres DB for all parts as well as the data exported into an XML format (also for all parts).

In XML format (The zip files contain the necessary XML files):

• Part 1: part1.zip

• Part 2: part2.zip

• Part 3: part3.zip

And as Postgres dump:

• Part 1: part1.backup

• Part 2: part2.backup

• Part 3: part3.backup

The schema of the data base is described in the DDL systemtables.xml. This file also contains useful comments about the necessary tables as well as about their columns.

If there are any questions about the data, please do not hesitate to contact heinist_at_inf.ethz.ch

Lineage model (Data model)

The data model is directly reflected in the schema of the database. Transformations (simple or complex) work on data items. Complex transformations (for instance the challenge FMRI workflow contain simple transformations (e.g. softmean, slicer, reslice etc.) or complex transformations. Data and control transfers occur between transformations. If there is any data item produced by a transformation which is later going to be used by another transformation, then there is a data transfer between these transformations. A data transfer is also associated with the data item itself. Control transfers occur when two transformation executions are executed after each other, thereby essentially defining the order in which they were executed (or the the control flow). The hierarchy is defined to be the parent-child relationships between the complex and simple and complex transformations.

Model Integration Results

Translation Details

Benchmarks

Further Comments

Conclusions

-- ThomasHeinis - 19 Feb 2007
to top