Powering CSIRO EO data analytics in the Cloud made EASI
Dr Robert Woodcock2, Mr Matt Paget1, Mr Ben Leighton4, Dr Alex Held3
1CSIRO Land & Water, Canberra, Australia, 2CSIRO Mineral Resources, Canberra, Australia, 3CSIRO Astronomy and Space Science, Canberra, Australia, 4CSIRO Land & Water, Clayton, Australia
The CSIRO Earth Analytics Science and Innovation Hub (EASI) is a powerful earth analytics platform that speeds up science and helps users analyse data to answer policy and management questions fast and reliably. The platform is underpinned by an open-source core architecture, Python data science and machine learning libraries, Open Data Cube, flexible and scalable cloud-computing (Amazon Web Services), analysis-ready Earth observation (EO) satellite data, and CSIRO expertise.
EASI operates as both a scalable analytics platform for CSIRO research and as an external business model for attracting new clients and partners. EASI has six deployments across regions and stable, development and pilot programs. More than 90 CSIRO researchers and 25+ projects are using the EASI platform.
Current challenges include deep understanding of effective and efficient scalable compute with Dask, to support applications to scale from demonstration-size to continental analyses. The speed of growth and change of cloud-infrastructure and technologies, such as Terraform and Kubernetes, also present an ongoing challenge for platform maintenance and development. We will present our current insights on these challenges.
Through EASI, CSIRO is investing in underpinning advanced technology to support our science and improve our connection to the innovative science and industry sector, both nationally and abroad. This presentation will introduce and update the technology, data services, science and partnership opportunities available now in EASI. We will conclude with forth-coming advancements in EASI data science capabilities including, for example, GPU acceleration.
Dr Robert Woodcock is a Research Scientist and Consultant in Earth Data and Analytics infrastructures with CSIRO Minerals. Dr Woodcock’s portfolio of activities involves national spatial data infrastructure and Earth observation data analysis and integration platforms. He has a focus on ensuring research innovation leads to business innovation.