Highlights from the Science Gateway Workshop at the eResearch Australasia Conference 2015 by AARNet’s eResearch Advisor, Alex Reid
Science Gateways is the term used (especially in the USA, possibly now here in Australia) for what we have previously called “portals” – they are software systems to provide integrated access to data, analysis tools, and modelling. Typically they are discipline-specific. The NeCTAR Virtual labs (VLs) are the current Australian equivalent.
There is considerable interest in pooling approaches to building Science Gateways globally, and the session on this was opened by Nancy Wilkins-Diehr, Associate Director of the San Diego Supercomputer Center at University of California, San Diego (by video conference). She described a number of successful Science Gateways in the US, as well as details of a bid being made to the US National Science Foundation (NSF) for funding to establish a Science Gateway Community Institute. This would operate as an incubator, extended support team, community engagement, modular layered framework, and workforce development. This way the construction of Gateways could be significantly accelerated.
NeCTAR’s Glenn Moloney and Nigel Ward provided a background to the development of the dozen or so NeCTAR VLs, which were formed around engaged research communities, ie they were deliberately discipline-specific. This approach was taken for several reasons, the main one being that it was critical to their success that actual research groups be at the centre of their development, so they had a stake in their success; another reason was that attempts elsewhere in the world (eg the UK’s early work on Virtual Research Environments) concluded that generic environments did not add much value for researchers. The current breed of research environments (or VLs, or science gateways) add value by enabling research to be undertaken more quickly, at larger scale, by a broader cohort of researchers, and in ways not previously possible or envisioned.
Subsequent speakers in the Workshop described some notable Australian Science Gateways (VLs), some built through NeCTAR projects.
For instance, AURIN (not a NeCTAR project) is the Collaborative Urban Research Environment for Australia project, which has provided integrated access to 1,750 datasets held by 70 different agencies across Australia. This allows unforeseen correlations to be analysed, such as urban air quality and tree locations across Melbourne, links between car ownership and diabetes, alcohol consumption by location, and a wealth of other urban planning analyses. Sadly, funding is now running low, so access to some (costly) datasets, like housing sale transactions, has had to be turned off.
Other projects described included the following:
There was quite a bit of comment about researcher workflows, with some of the more common tools mentioned (eg Kepler, Taverna, Galaxy, VisTrails, Triana & Pegasus). There is quite an emphasis these days in research in reproducibility of research outputs, which becomes very complex in the eResearch environment (need to store and annotate data sources, archive software, record parameters used, etc). To this end, WorkWays, Workflow driven Science Gateways, was described.
Sandra Gesing, a computational scientist at the Center for Research Computing at the University of Notre Dame, USA, also described her experience with Science Gateways, highlighting current solutions and future challenges. There are now many tools to help build science gateways in an agile way, leading to Science Gateway Platform as a Service, SciGaP, funded by the US NSF.
The Workshop concluded with a Panel which focused on the proposition ‘What makes for a successful Science Gateway?”, with the predominant view being that they were successful when they were so valued by the research community that they used them, owned them and funded them.
Apr 15, 2020