File and Object Storage

Video: Learn how the University of Queensland uses IBM Spectrum Scale and Elastic Storage System (ESS) to Improve Human Life

By Peter Basmajian posted Tue January 12, 2021 05:24 PM



David Abramson, Director, Research and Computing Centre, University of Queensland
Jake Carroll, CTO, Research and Computing Centre, University of Queensland

David Abramson  00:07

If you're doing research in the 21st century, in anything from humanities through to physics, you're not doing it without computing. We have research across the whole spectrum. So whether it's genomics, whether it's microscopy, images, electron microscopy images, it could also be from scanners, MRI scanners and pet scanners and the like. So they are all generating exponentially more data. There's no area of the university, which isn't touched. The amount of science it's been done is going up. And so that's generating more data. And just the complexity of the problems we can solve is going up dramatically.

Jake Carroll  00:47

We're actually using finite element method and finite element analysis, to work on ultrasonic wave propagation in the human skull to temporarily traverse the blood brain barrier, to be able to more effectively deliver drugs to reverse the effects of Alzheimer's disease. This is a challenging workload, and it's ultimately going to have an impact on human life.

David Abramson  01:09

The University of Queensland is a large research intensive University and we have 1000s of researchers. So from our scientific instruments, we're just seeing exponential growth in the amount of data that's coming in, you need to process it, you need to have computational models that run against it. And so the challenge for us is bringing all of that together in sort of one seamless, seamless way. So not just the volume, but the number of datasets they're capturing, can they find that data again, that experiment they did last week. So we're really trying to simplify those workflows for people who would rather be doing something else, like research. So one of the things we're looking for is, is a uniform storage architecture in which we render that data in multiple ways.

Jake Carroll  01:51

We selected IBM Spectrum Scale, and specifically the ESS building blocks, for a number of reasons performance, engineering capability, and RAS.

David Abramson  02:02

Spectrum Scale, it allows us to actually have that unified access mechanism, regardless of how I want to use the data. And that's really powerful.

Jake Carroll  02:13

An important benefit that the ESS building blocks have created for the University of Queensland is time to discovery. There are experiments that have been done and research that's being undertaken, which on our previous platform would have taken 19 or 20 hours. We can cut that time down to five or six hours. We have recently implemented IBM Storage Insights so that we can gain a better understanding of data growth, performance and a cohesive understanding of our entire infrastructure. Some workloads that we've got running at the moment include everything from cancer detection, melanoma detection and characterization through to blood pathology, workflows, and even Drosophila fly and motor neuron connectivity circuits. These are all complex workflows. They use a lot of bandwidth. We want to be able to innovate and do interesting things. IBM is a company that wants to do those things with us


1 comment



Sun January 17, 2021 11:43 PM

IBM Spectrum Scale is perfectly suited for this type of use cases. It would be interesting to add IBM Spectrum Archive to this solution.