The University of Reading has deployed Nutanix Files to replace an ageing network-attached storage (NAS) infrastructure of NetApp filers and Gluster storage. The software-based product is an add-on to the university’s existing Nutanix hyper-converged infrastructure (HCI) estate, which runs to more than 0.5PB (petabytes) in capacity.
The deployment has allowed it to offer researchers self-service access to a pool of storage and has cut storage administration time to virtually zero compared with the previous setup.
The University of Reading has more than 50 research centres in areas that include agricultural, biological and physical sciences, and meteorology. The university has around 5,000 faculty members and 20,000 students.
“The NetApp and Gluster were old technology with a number of issues,” said infrastructure services manager Kevin Mortimer. “The NetApp was out of support and had a 1Gbps network card, so that was a bottleneck. It was fine 10 years ago, but as compute has progressed it was blocked at the network level.
“With Gluster, the support contractor went bust, so we had to negotiate our way out of that situation.”
Overall, said Mortimer, the university’s existing file access storage was old, slow and wasn’t able to scale any further.
The establishment already had a Nutanix hyper-converged deployment and Nutanix Files – a NAS product that works from the Acropolis Hypervisor (AHV) was added to that.
As Mortimer described it, Nutanix Files was deployed on additional hardware and rolled out via “a one-click cluster join” where storage pools were assigned and data migrated from NetApp and Gluster. Nutanix Files is managed from Nutanix Prism.
The additional hardware comprised 12 nodes built around Dell XC series servers with 80TB (terabytes) of mixed flash and spinning disk HDD capacity in each. That’s not far off 1PB of NAS capacity, and that’s not accounting for data reduction that has boosted usable storage by about 16:1 on average.
The key benefit for Mortimer – apart from gaining nearly a petabyte of NAS storage – is that the onerous management overhead that came with the ageing NetApp and Gluster legacy environment has all but disappeared.
“We don’t ever look at it,” he said of the Nutanix deployment. “We don’t have to actively monitor performance. The researchers are very demanding. They want the fastest performance possible, but they only want it for an hour a week and previously we spent time watching it for bottlenecks.
“The key benefits are around performance,” said Mortimer, who also spoke about the increasing use of self-service to allow researchers to provision storage for their projects via Nutanix Calm.
The university has also added Nutanix Frame desktop-as-a-service to provide its corporate users with access to business systems during Covid-19.
Three years ago, the University of Reading consolidated around 600TB of data on legacy storage to Nutanix in a deployment that cost up to £400,000 at the time, but which was expected to save £500,000 in avoided VMware costs.
Nutanix has been a pioneer of hyper-converged infrastructure, in which compute and storage are brought together in nodes that can be built in scale-out fashion into clusters.