Actuarial consulting company Hymans Robertson has deployed NetApp storage with HP servers and networking kit in a £1 million project to boost disaster recovery provision, improve backup and equip itself for a push to gain greater market share.
Hymans Robertson provides services to public sector organisations, banks, building societies and insurance companies. It has 600 staff in London, Glasgow, Birmingham and Edinburgh.
Key apps are in-house-developed financial modelling apps and its Vitabank mortality-prediction tool as well as Civica’s administration platform, all of which runs in a Microsoft Windows Server, .net and Sharepoint environment.
The firm had two HP EVA SANs that were around five years old but these were near their maximum capacity in their current configuration and took a lot of housekeeping to maintain, said Barry Smart, IT director at Hymans Robertson.
Behind this was a tape backup regime that saw full backups every weekend. These were, however, becoming onerous with a backup window that often spilled over into working hours on Mondays.
The company carries out annual disaster recovery testing, and according to Smart they were looking at “a disaster on the horizon” with response times in tape recovery that were “less than the business expected”.
Following a business-wide review, the firm started looking to refresh large parts of its IT and disaster recovery infrastructure.
Deploys NetApp, rejects flash disk
After an evaluation that also included EMC storage, (EMC/Cisco/VMware) VCE Vblocks and NetApp/Cisco/VMware FlexPod converged infrastructure as well a HP storage, Hymans Robertson settled on NetApp storage in conjunction with an HP server and networking equipment refresh.
Smart said he rejected the idea of flash drives in the arrays after EMC and NetApp profiled the firm’s workloads.
“EMC and NetApp did evaluations of the EVAs and determined that we don’t require super-high performance. Flash Cache is doing its thing. NetApp has recommended we boost the count of SAS drives to cope with higher end databases causing the SATA drives to struggle,” he said.
The NetApp filers support a Microsoft System Center 2012 and Hyper-V virtualisation environment on new HP blades
Previously it had a recovery point objective (RPO) of 48 hours and recovery time objective (RTO) of 24 hours using restore from tape. Now it is looking at being able to be up and running within a few hours of an outage striking a datacentre.
Rejects Vblocks and FlexPods
When evaluating new storage solutions, Smart’s team also looked at HP storage as well as weighing up EMC/Cisco/VMware Vblocks and NetApp’s Flexpods, which are its version of the converged stack product.
HP storage was rejected, said Smart, because “HP didn’t seem to get what we were trying to achieve and the solution proposed didn’t meet our needs.”
EMC/Vblocks got to the final commercial proposals stage but Hymans Robertson steered away from converged stack products, said Smart.
“We didn’t want Cisco servers; we wanted to stick with HP compute and considered that buying storage and servers separately would present lower risk and better value for money,” said Smart.
“We saved quite a few pounds by going down that route. We felt that to change our SAN vendor was well as our compute and network vendors would have been too much at one time. We could see the advantages [of the converged stack] but we have a small team and would have spent a lot of time building knowledge and skills.
The key benefits for Hymans Robertson have been that they can do away with £50,000 a year in tape backup costs, much improved disaster recovery capabilities, and an improved IT infrastructure with more flexible storage and compute to support business growth.