BillionPhotos.com - Fotolia
The move saw Nutanix replace traditional server and storage architecture, including an EMC SAN, with huge reductions in server room space and cooling as well as snag-free patching of the infrastructure. The project also saw multiple backup products replaced with CommVault.
The Trust is centred on Chesterfield Royal Hospital, which serves a population of 350,000 and has 550 beds and an IT estate that runs around 260 applications.
At the start of the process, in which it was helped by Chesterfield-based integrator Coolspirit, the trust had about 70 Dell servers each dedicated to an application, with 12 more running two VMware ESX clusters.
Storage was provided by an EMC VNXe SAN and it was here that the key bottlenecks arose as it reached capacity.
“The biggest headache was storage,” said IT technical delivery lead, David Sawyer. “When we first got it it had been purchased for requirements at the time. No-one could know the explosion of servers there would be.”
“We had only about 50 VMs then,” said Sawyer. “In seven years that has gone to 400. As that number has increased storage has been battered, with the app guys wanting more and more.”
Read more about hyper-converged
- We look at the latest in hyper-converged infrastructure – server and storage hardware nodes that can be built into scale-out clusters – and the players in the market.
- Hyper-converged infrastructure products merge server, storage and hypervisor in scale-out nodes, so how can functionality in HCI products help deliver disaster recovery?
At the same time, said Sawyer, the compute side was running out of resources, including RAM on occasions.
“We got to the point where we had to ask, ‘Do we throw loads of money into this and keep expanding it, buying shelves and drives?’ In the end we decided to see what was out there,” said Sawyer.
His team considered a setup from NetApp in the traditional three-tier architecture. “They wanted to come along with a pre-configured cabinet, but we simply didn’t have space.”
The trust eventually plumped for Nutanix hyper-converged infrastructure and now has 400 virtual machines running on 24 nodes.
Did Sawyer have any worries about opting for what was a new alternative to traditional IT architectures? “Yes, we had concerns. It was something completely new but with some research we felt we knew where we were going,” he said.
“The attraction was that we could easily add to it and not create bottlenecks. We had been able to add to the SAN but that created a bottleneck between servers and storage. We decided to take the risk and go down the hyper-converged route.”
Key benefits of Nutanix
Key benefits of the Nutanix deployment for Sawyer have been space saved in the trust’s server rooms and lack of disruption during patching and upgrades.
Previously, patching and upgrades had to be restricted to weekends because of the likelihood of causing unplanned downtime.
“We always expected that something would break,” said Sawyer.” Often it would be ESX. It may have been something we’d done. We’d patch the server and we’d lose a node. Now we know that upgrades will happen without a blip.”
“That means upgrades can happen in office hours without needing to pay for resources at weekend rates and in the knowledge we won’t have anything fall down.”
Meanwhile, the reduction in total hardware has emptied seven racks of servers and Sawyer is looking to reduce the size of the server room. Alongside that around half of the air conditioning units often switch themselves off.
The big benefit for Sawyer is, however, that compute and storage is now less costly and more reliable, and easily added to.
“The number of VMs has gone through the roof, and traditional storage would have struggled, with extra chassis and disks needed,” he said. “With Nutanix, you put a node in, go back to your desk and turn it on; it’s not a project.”