Sergiy Serdyuk - Fotolia
Liverpool School of Tropical Medicine (LSTM) has deployed Scale Computing hyper-converged infrastructure nodes and saved £80,000 compared to the cost of replacing its existing HP servers and storage area network (SAN) storage.
LSTM has around 600 students and 500 academic and support staff and was running 16 HP Server BladeCenter servers and an HP 2040 MSA SAN with around 30 virtual machines on Microsoft Hyper-V.
But this setup had reached end of life and had become difficult and time-consuming to manage, said IT leader Matthew Underhill.
“It was ageing and struggling, so we had hardware failures,” said Underhill. “Also, it was the traditional server and storage setup, so compute was in one area, storage in another with a dedicated SAN network, with all the layers of complexity that brings.”
“We had to maintain disks, the SAN and its logical unit numbers (LUNs). It was taking an hour to set up a virtual machine (VM) with extra configuration on top, then there were lots of layers when troubleshooting,” he said.
So, Underhill’s team started to look for a replacement solution, with a key requirement being to upgrade the infrastructure to allow mirroring to a secondary disaster recovery site. Its existing data protection setup was a combination of on-site disk and off-site tape.
LSTM ended up with a shortlist of hyper-converged products from Nutanix and Scale Computing.
Read more about hyper-converged infrastructure
- The rise of hyper-converged infrastructure – with compute, storage and networks in one box – seems ideal for SMEs, but is it always a better idea than traditional IT architecture?
- Hyperscale computing and storage are the norm for web giants. Hyper-converged scenarios make it possible for small and medium-sized enterprises to gain the advantages of combined server and storage nodes.
Nutanix was rejected because Underhill was unable to get a quote from the company. “We thought it was a good solution, but we couldn’t get a quote from them. I don’t know if they thought it was too expensive for us or a waste of their time, but it was their loss,” said Underhill.
LSTM has deployed a five-node cluster of Scale HC3 hyper-converge nodes at its primary site with four nodes as a disaster recovery cluster. The primary cluster comprises 40 central processing unit cores and 66TB of storage, with most data on nearline-storage as a service drives with flash drives used as cache.
“The flash cache on Scale is a lot cleverer than some,” said Underhill. “We can specify a workload or use sliders to pin hot blocks to flash.”
Underhill said the key benefits are that the Scale setup is simple and scalable, and that disaster recovery testing is easy. “We can now say with certainty that all VMs are recoverable in minutes. It gives us peace of mind; it’s tested and we have proof.”