Melbourne’s Monash University has deployed what is believed to be the world’s first 100Gb/s end-to-end OpenStack cloud.
The university’s expanded cloud infrastructure is based on Mellanox’s CloudX platform, which includes Spectrum SN2700 Open Ethernet switches, ConnectX-4 NICs and LinkX cables.
Its Research @Cloud Monash (R@CMon) is supporting researchers in their quest to build ‘the 21st century equivalence of microscopes’ to inspect and make sense of large amounts of data to derive insights.
“Our cloud-based eResearch platform must be able to support their initiatives in every possible way, whether it be accelerating next-generation sequencing to the clinic or making our cities smarter and more energy efficient – we need a mixture of scale, speed and flexible integration in our computing,” said Steve Quenette, deputy director of the Monash eResearch Centre.
“In addition, as Big Data becomes increasingly prevalent in research, business and society, the experience we gather from building cutting-edge, high performance cloud infrastructure can provide invaluable guidance to the rest of the world,” he said.
The R@CMon cloud is part of the National eResearch Collaboration Tools and Resources (NeCTAR) Project, a government initiative for Australian researchers. R@CMon initially received $4.5 million in funding from both Monash University and the federal government.
Monash first deployed Mellanox’s CloudX Platform in January 2015 for the R@CMon cloud, setting it up as an infrastructure-as-a-service offering for researchers.
Prior to deploying the cloud, the university found it hard to handle peak research activities and high IOPS (input/output operations per second). Earlier this year, the Monash said it research activities accounted for around 5 petabytes if usable disk storage.
At the time, the university said it was supporting 317 growing research projects, with each one staffed by 100 researchers.
Monash’s senior cloud architect, Blair Bethwaite, said in addition to R@CMon, the university had created numerous ‘virtual laboratories’ for data-intensive characterisation and analysis.
“They are becoming the standard operating environment for the modern day researcher to support general purpose HOC and HTC (including GPGPU capabilities and Hadoop), interactive visualisation and analysis,” he said.
Follow Byron Connolly on Twitter:@ByronConnolly