• Stenson, Matt
• Watkins, Damien

Abstract

This document looks at potential linkages between two scientific workflow systems developed by CSIRO: the Workbench and Workspace products. As discussed below, the Workbench tool in its current form will not be fit for purpose in the medium term (5-10 years).The Workbench is a workflow tool based on the Microsoft Trident product. The Water for a Healthy Country Flagship has made significant investment in a series of tools to support hydrologic modelling, and aid in the management of modelling scenario’s and the capture of provenance chains as part of the Workbench. The Workbench development has been driven by demand for more transparency in integrated modelling. The demand for transparency has been especially great where the results of the models form the basis for decisions.Unfortunately Trident is no longer supported by Microsoft, and the current Trident designer has a remaining usable lifespan of 5-10 years without significant investment.The Workspace product is also a scientific workflow tools that has been developed within CSIRO’s CCI Division. It offers some overlap in functionality with the Workbench and could provide a useful migration path from Trident for the specific Workbench functionality developed over the last few years.Both tools support a growing demand for management of complex and integrated scientific modelling task. The nature of modelling normally requires the generation and execution of many model runs, each with their own input files, processing steps and output files. Scientists interpret these output files and publish the results (knowledge). This process of data to knowledge is in essence a scientific workflow. Increasingly the results of such modelling are coming under intense scrutiny around the lineage, i.e. what data, modelling code, parameters and expert input contributed to their generation. This is known as provenance. Scientific workflow tools greatly simplify both the management of the modelling process and the capture of the associated provenance information. This document recommends that: • the current Workbench tools be redeveloped to work under Workspace; • experimentation in Workspace begin in the 2014/2015 FY to build a lightweight wrapper to call the Workbench activities; and • by 2015/2016 all new projects requiring workflow support would be developed in Workspace. This is contingent on upgrades to Workspace: • Improved provenance system • Handling of Python and raster operations through GDAL

Topics

• impedance spectroscopy