New Taverna Release

Back to index page

Previous article: News in brief

Next article: Bringing order to service assembly

A new version of Taverna was released in December 2008 by the myGrid team – one of the three OMII-UK partners. The release comprises the Taverna 2.0 Workflow Engine and an experimental workbench interface, which showcases some of the engine's new capabilities. Taverna 2.0 is set to continue the success of its predecessor: more than 1200 downloads were recorded less than a month after its release.

Taverna 2.0 uses a re-engineered, scalable, high-performance and extensible enactment engine. It natively handles data-reference management so that data is shipped around the engine and can be cached to disk without intervention from the user. The Engine has been developed to solve real-world problems, such as the need to handle large amounts of data, and the need to iterate over, combine and collate data sets. These features have been designed to overcome problems faced in the diverse domains in which Taverna is used, such as the neurological data handled by the CARMEN project, the medical information processed by the caGrid project, or the resources used by text and data mining groups.

The performance of Taverna 2.0 has been improved. The overall effect on the execution time for a workflow can be significant (preliminary tests show a five-times improvement over Taverna 1). Workflows for the Taverna 2.0 Workflow Engine can be written using either the 1.7.1 Workbench or the experimental Taverna 2.0 Workbench. A Taverna 2.1 Workbench is due to be released in March 2009. It will benefit from a more user-friendly workflow interface, whilst retaining access to the Taverna 2.0 Engine’s capabilities.

http://www.mygrid.org.uk

Taverna 2.0 features at a glance:

• Handles large numbers of large data items.
• Extensible: it is already being extended to run services on caGrid.
• Includes implicit iteration over items in a data set.
• Allows configuration for combining and iterating over sets of data.
• Pipelines data for faster workflow execution, with services called and workflow outputs produced as soon as possible.
• Copes with erroneous data, preventing it from affecting the rest of the workflow run and allowing traceback of problems.
• Can run both Taverna 1 and Taverna 2 workflows.