The Australian Research Data Commons’ (ARDC) Storage and Compute Infrastructure Summit was held on Monday 21 October 2019 in Brisbane, and co-located with the ARDC Data and Services Summit and the eResearch Australasia conference. It brought together a combination of project participants and other eResearch infrastructure providers to share their learnings from the 10 Discovery projects funded by the ARDC.
The delegates discussed Australia’s major eInfrastructure challenges and opportunities around the technologies, techniques, and services that underpin a national research data commons.
Key issues, challenges, and opportunities discussed at the meeting included:
authentication and authorisation – including practices and techniques that ensure data is safe, trusted and secure and at the same time accessible and useful for researchers and other appropriate groups of people
machine Learning – questions addressed included what makes a useful viable data set to enable machine learning and what are the attributes of the systems and platforms to support them
skills – the importance of being able to map out the skills needed and the most appropriate training providers, including the type of training required to enable people to take full advantage of the systems in place, particularly relevant for research involving sensitive data
High-Throughput Computing (HTC) – there was general agreement amongst the delegates that further development of ideas and solutions in the context of a Research Data Commons would be beneficial, with the ARDC to look at specific use cases to support different groups
GPU Computing – the role GPU’s should play within the ARDC cloud infrastructure and which of several resourcing models should be adopted
support for the Nectar Federation model – there was strong agreement that the federated model should continue in the future as well as support for the adoption of a more consistent and compatible technology mix across partners. This included discussion of governance models and the pros and cons of various models of support.
Delegates discussed the ARDC’s role in helping to tackle these infrastructure challenges and the role of storage and compute infrastructure in driving international research competitiveness. Given that the ARDC’s core business is within the Data and Services and Software and Platforms themes, the focus for the Storage and compute theme is on supporting and enabling our core business to thrive. The ARDC will continue to invest in the ARDC’s Nectar Research Cloud to ensure researchers can work seamlessly across different sites with an evolution of our resourcing model to emphasise investment in innovative leading-edge infrastructure over less differentiated cloud resources.