Plan disaster recovery: Royal Horticultural Society ready for any business disaster

A virtualised infrastructure with vSphere 4 and a Compellent SAN has optimised Royal Horticultural Society's DR process and saved money. Instead of sending virtualised replications across a WAN link, the organisation physically transports data to its DR site.

Now ready for any business disaster gardening charity the Royal Horticultural Society (RHS) no longer uses a wide area network (WAN) link to replicate its virtual machines (VMs) to its disaster recovery (DR) site, but drives them there instead.

After implementing VMware's vSphere 4, the company went from 50 to 35 physical servers and 15 VMs, all of which are replicated at the charity's DR site, located one mile from its primary data centre in case of a business disaster. Plan disaster recovery
Instead of investing in WAN optimisation technologies to send virtualised replications across the wire, RHS uses Portable Volume,a new virtualisation capability included in Compellent's SAN Storage Center 5, to physically transfer data from the company's primary data centre to its disaster recovery site.

Replication with the link did occur, but not at the speed that we wanted. Instead of weeks it now takes days.

Martin Taylor, Converged Network Manager at RHS,

It replaces the WAN between the two sites, so RHS's IT team doesn't have to send terabytes of data across a link.

The society maintains a library of an estimated 200,000 high-resolution horticultural images, some of which are 200 years old, so any kind of business disaster would leave many of these images lost. These were organised into a centralised image database to enable access for all employees, in addition to making them available as copyrighted materials to the general public.

Martin Taylor, Converged Network Manager at RHS, said the biggest concern, and the reason why they planned their disaster recovery strategy, was bandwidth. "We have a large amount of data, due to all our images that needed to be sent to our DR site. Moving all of it on our bandwidth -- our WAN -- means it is tough to do replications. We don't have the resources available to do 'round-the-clock replications this way."

The company claims to have saved an estimated £10,000 a year, which it would have been paying for a 1GB WAN link in London.

"We also use vMotion, but only in our primary data centre, in a local sense, not across a WAN link, as it would take far too long. We plan to eventually go 100% virtual," added Taylor.

Compellent Portable Volume
RHS has a Compellent SAN at its primary data centre in London and a second SAN located a mile away for disaster recovery, in the event that there's a business disaster.

RHS is currently moving its data between sites via a replication strategy that copies its data to the Compellent Portable Volume, a group of disk drives. When a snapshot is placed on the Potable Volume, Compellent's software replicates the delta changes to the DR site. The device is then couriered twice a week to RHS's second data centre. Here, Compellent's software synchronises the delta changes with the device to provide an up-to-date copy.

"The WAN link normally makes all the changes, but we're doing it in a manual way. Replication with the link did occur, but not at the speed that we wanted. Instead of weeks, it now takes days," Taylor said.

Previously, RHS had to recover each server from tapes that were taken off-site every evening. This required the IT team to configure a set of new servers, rebuild all the applications and reload data from the tapes to cushion themselves against a business disaster.

Portable Volume allows the user to pre-stage data, after being encrypted and compressed, onto one or several 1.5 TB external USB devices, which are physically taken to a second data centre site to kick-start replication.

This feature enables RHS to avoid a lengthy online session to transfer the data to its disaster recovery site.

The data centre infrastructure
RHS's IT infrastructure is required to be fully operational all the time to grant members access to its gardens, handle new membership applications, process transactions at the society's shops and ensure suppliers are paid – downtime, as a result of a business disaster, is not an option.

Its infrastructure also plays a key role in the organisation and management of the company's flower shows, including its flagship events at Chelsea, Hampton Court Palace and Tatton Park, which host more than 500,000 visitors.

Taylor said the amount of data shot up very quickly compared with using direct storage.

"You tend to be very hot on direct storage -- you keep a tight check on it. With the SAN we have about 10.5 TBs of data now, but we were expecting about 6 TBs. If we have the resources there for people to use, though, they are going to use it."

After looking into similar products from NetApp, Hewlett-Packard Co., EqualLogic, Sun Microsystems and 3PAR, RHS decided on Compellent for its functionality, to safeguard it against a business disaster, and value for money.

Kayleigh Bateman is the Site Editor for

Read more on Datacentre disaster recovery and security