A review of disaster recovery capabilities has made multinational specialist recruitment company Hudson more resilient in its day-to-day business.
“We have improved Hudson’s business continuity by modifying our disaster recovery strategy,” said Bas Alblas, IT director Europe at Hudson.
“The back-office IT infrastructure of all our offices across Europe has been consolidated into a London datacentre, and we plan to build a mirrored datacentre for complete redundancy,” Alblas said.
Once this project is finished the European datacentre will be fully redundant and Hudson’s infrastructure can survive a complete outage of one of the datacentres.
The disaster recovery review was conducted in the light of a couple of datacentre outages in the past 10 years as a result of external factors. These included a wide area network (WAN) outage as a result of local construction works, a cooling system failure in a datacentre and a water leakage.
Hudson’s view on disaster recovery
As Hudson migrated all local server infrastructure to one central datacentre in the past five years, a disaster recovery (DR) solution for that datacentre started to make sense, to mitigate the risk of single-point-failure. Initially the scope was to have a DR solution for the most critical applications, covering a maximum of 50% of the total user base.
“During the investigation we found out we would be able to create a DR setup covering all applications, able to accommodate all European users,” said Bas Alblas, IT director Europe, Hudson.
“As a result, we were able to get project buy-in/budget without much effort. HP’s solution allowed us to do much more than originally required, with little budget increase. The project also involved replacing all existing servers and network equipment with the newest technology, offering more CPU power, three times more diskspace and an upgrade from 1 gigabit to 10Gb network,” he said.
Virtualising all physical servers also allowed Hudson to improve its backup strategy, improving backup and restore times by moving from tape-based backup to disk-based backups and backing up virtual machines using agentless technology instead of agent-based file level backups to tape.
Hudson, which operates in around 20 countries, chose HP servers and storage to ensure full recovery of all applications, expanding the firm’s existing relationship with the IT supplier beyond printers and desktops.
Alblas said HP had offered the best value for money. “Simple but working solution, preventing Hudson from paying for functionality we do not need or would not use,” he told Computer Weekly.
“HP storage ensures uninterrupted availability for all our users, and virtualising all our physical servers onto HP Proliant blade servers has cut Hudson’s datacentre power usage and floorspace in half,” said Alblas.
A single unified storage pool now supports Hudson’s virtualised environment, business applications and services, at a lower cost and with greater flexibility.
“The total percentage of costs savings is hard to calculate, as it includes simplified management, better server usage with more VMs per host and datacentre costs as a result of lower foot- and power-print,” said Alblas.
“Using a second datacentre is an increase in costs obviously. However, basic analysis convinced Hudson this solution would be cheaper or cost-neutral in the long term, providing more power, diskspace and resiliency at the same time,” he said.
“One other advantage is that we’re much more flexible in choosing datacentres and WAN suppliers, as we can easily move one datacentre to another location without scheduling an outage as, during the relocation of one datacentre, the other will take over.”
According to Alblas, a storage area network (SAN) based on an HP LeftHand Storage system provides 61TB of secure, easily managed storage for Hudson’s business-critical data.
“Thin provisioning has already saved 43TB of physical disk space, and performance can be scaled linearly to grow with Hudson’s business,” he said.
HP Virtual Connect simplifies the connection of the SAN to a virtualised server platform of HP ProLiant BladeSystem c7000 enclosures with 26 HP ProLiant BL490c server blades running VMware 4.1.
“HP Integrated Lights-Out makes it possible to manage the HP servers remotely,” Alblas said.
Hudson’s IT environment is centrally managed through HP Systems Insight Manager. Hardware-level management and automated remote support for the HP servers and storage enable the IT team to maximise system uptime. HP Snapshot software maintains system performance by allowing I/O-intensive applications like zero-downtime backup to run concurrently without impacting the performance of Hudson’s primary applications.
Alblas said HP partner Softcat provided advice and testing for the HP system prior to deployment.