Luca Cervigni – Designing the systems for superior data storage

Luca Cervigni is the IT infrastructure architect at Pawsey, and for the past five years has been working extensively with Pawsey’s cloud and storage technologies.  He describes his days as spent “writing notes, drawing diagrams and coding”, but with his responsibilities for systems design, he also works closely with other supercomputing centres (such as the National Computational Infrastructure (NCI) in Canberra and the Swiss National Supercomputing Centre (CSCS)) and major research facilities (such as the European Organisation for Nuclear Research (CERN) and South Africa’s SKA pathfinder telescope MeerKAT) to ensure that Pawsey’s systems stay at the cutting edge.  “Exchanging point of view and building on the experience of other experts worldwide to better understand the strengths and weaknesses of each system is very important in my work, to ensure Pawsey invests in appropriate future technologies, while minimising the risks of implementing new systems.”

His scientific interests span physics and astronomy, and he is no stranger to mega-science projects – he was working at CERN in Geneva when the elementary particle the Higgs boson was first detected in 2012.  “That was one of the most exciting moments of my career,” Luca remembers.  “It is why I always love to support and enable research: I can see that what I do is so important for so many others.”

“Pawsey is arguably the most cutting-edge research-enabling centre in Australia.  We may be known for big projects like the SKA, but we support an extremely broad range of science disciplines.  It would be impossible for many of these researchers to perform their jobs and achieve what they do without Pawsey.  It makes working here pretty special.”

Luca has recently been designing, and is the technical leader for, Pawsey’s new 90 PB object storage cluster.  This is the largest part of Pawsey’s $70 million Capital Refresh apart from the new supercomputer Setonix itself.  The new storage system accommodates not just the sheer volume of storage required, but also considers the speed at which data can be initially accessed, and moved in and out of storage.

“This new object storage will completely revolutionise the way Pawsey users access their data,” explains Luca.  “Our object storage will use S3, a standard Hypertext Transfer Protocol (HTTP) protocol that will allow researchers to access their massive datasets more easily via the internet. It will move Pawsey even closer to its future goals of extremely large but very accessible data collections.”

“It’s been very rewarding to have the trust of Pawsey’s management and be empowered to suggest new approaches and change the way we design our internal technologies as we upgrade,” says Luca.  “HPC of the future will only move further towards complete adoption of containerised workloads and object storage.  It allows us to completely decouple where the data is stored (Pawsey) from where it is analysed and worked with (everywhere the researchers are).”

With Luca’s assistance, Pawsey is already making massive moves in this direction.