The CSIRO is boosting its use of public cloud services to support the computing needs of its researchers at 55 sites Australia-wide and its facilities overseas.
Under its Cloud Connect program, the CSIRO is initially putting in place ‘AARNET-speed connectivity’ between its locations and Amazon Web Services’ data centres, CSIRO’s chief information officer, Brendan Dalton told CIO Australia.
“We already have 10Gigabit connectivity between all of our sites and my hope – but I would need to test this – is that is exactly what we're [deploying] in terms of connectivity ... as a first cab off the rank with AWS,” he said. “We will also be utilising other cloud service [providers] over time.”
Dalton and his IT team are effectively helping researchers broker cloud services when they have “small timeframe, high compute requirements” and when they need high levels of security.
The CSIRO has also started working with Vault Systems and Sliced Tech, which in March became the first cloud providers to have services approved by the Australian Signals Directorate’s CCSL program for use with classified government information. Dalton said the CSIRO is also looking into using Google’s cloud services under the Cloud Connect program.
CSIRO business units such as health and biosecurity, Data61, and Australian Animal Health Laboratory work with research and corporate partners and demand a high level of security, said Dalton.
The highly secure infrastructure offered by these government-certified cloud providers is vital for CSIRO researcher Denis Bauer and the agency’s transformational bioinformatics team.
They have created VariantSpark, a machine learning method that mines genomic information to detect disease genes. Bauer said recently that as some diseases are caused by changes in the genetic code, genomic data provides crucial information for medical research and in developing treatments for patients.
“Genomic data requires a lot of security because it’s very personalised and needs to be encrypted,” Bauer told CIO Australia.
Bauer and her team have also built a new software tool for computationally-guided genome engineering on AWS. GT-Scan 2 is a software tool which improves the effectiveness of a molecular engineering technology called CRISPR-Cas9, which can be programmed to recognise and edit specific locations in the genome by pattern-matching unique sources of DNA. This technology may revolutionise cancer treatments.
Dalton the CSIRO’s data volumes are so vast which does present a cost issue and the agency’s internal IT service provision “to put it mildly is quite cheap in comparison to commercial [cloud] offerings.”
“So I do have that conundrum to deal with but our strategy over time is to move away from tin and like a lot of people, use cloud services where it’s the right thing for us to be doing,” he said.
In addition to brokering cloud solutions, the CSIRO has an internal strategy around getting more technical support staff to work with researchers to help them realise how they will benefit from new and high performance computing technologies.
“My team has a role in assisting researchers to better understand the possibilities of technology. Other people are focused on other things so part of my strategy is to have the people translate technology and how that might be applied to different research scenarios,” he said.
Follow Byron Connolly on Twitter: @ByronConnolly