An increasing number of NHS trusts are undertaking virtualisation projects. Often, these are to address a specific issue, such as disaster recovery. But many spin-off benefits are being reported. Daloni Carlisle reports.
NHS trusts are starting to implement virtualised systems for a range of needs. The organisations that have moved in this direction are anticipating a range of benefits, from improved disaster recovery to better performance, reduced costs and increased flexibility and agility.
Ending server sprawl
Liverpool Women’s NHS Foundation Trust was the first NHS organisation to build a fully virtualised, consolidated and mirrored Storage Area Network (SAN) IT infrastructure. In 2007, this was nominated for a Computerworld Honor.
In its citation, the trust claims huge benefits for this complete rebuild of its IT infrastructure. It has, for example, cut the number of servers from 30 running 50 major applications to five new Dell servers running VMWare virtualisation software.
The trust’s figures show that operational running costs have been cut by 70%, hardware and software costs by 60% and total ownership costs by 64%. Meanwhile, the new set-up has delivered ten times faster access to IT systems for clinical staff and allowed the trust to pilot many state of the art technologies.
What were the major challenges in implementing this? The compete change of mindset required by IT staff? The data migration? No, it was convincing the trust board that the IT department could do it. Other trusts are now following suit, with several virtualisation projects completed and many more commissioned in 2008.
“Trusts were in a holding position until 12 to 18 months ago,” says Alastair Kitching, commercial director of Esteem Systems, which has worked on virtualisation projects in several NHS trusts. “As a result, many are suffering server sprawl and their equipment has become quite outdated. But the NHS Infrastructure Maturity Model (NIMM) means trusts now have the go ahead to move their infrastructure from fairly basic to standardized, virtualised equipment.”
Ready for disaster
While some trusts have taken a radical, system-wide approach others have focused on a particular area such as disaster recovery - and then seen other benefits spin off from it.
Royal Liverpool Children’s Hospital is a good example. It is one of the largest and busiest hospitals in the country - and the largest children’s hospital in Europe. Yet as recently as two years ago, it had no fallback option if something disastrous had happened to its servers.
Last year, the trust engaged ISC to implement a virtualised server environment using VMWare running on six HP ProLiant servers. ISC created a virtualised environment in the hospital and then replicated this in a separate room as part of the disaster recovery plan.
The implementation included a NetApp SAN populated with fault tolerant RAID storage to provide a centralised data system for the servers, while a Quantum tape library catered for backup.
The main benefit, according to ISC, is that if there was an incident - such as fire destroying the building housing the servers - there is a fail-safe in place to keep the IT up and running. But the trust also reports improved reliability of its IT systems and improved system performance. Data security is better, backup simplified and - yes - there have been cost savings too.
Speed and flexibility
Leeds Partnerships NHS Foundation Trust has also gone down the virtualisation route in its disaster recovery plan, this time using CSA Waverley. The solution saw their server numbers cut from 32 to nine, with three newer SQL Servers being recycled into the new mirrored DR system.
Like Royal Liverpool Children’s Hospital, the trust reports a series of spin off benefits. New applications can be deployed in minutes rather than weeks and IT staff can easily create test builds before new services are launched. IT staff no longer have to spend time on manual back up and can instead focus on delivering a better service.
Dorset Healthcare NHS Foundation trust is just now embarking on a similar journey with CSA Waverley -- although the driver here is a critical need to free up data centre space and power. CSA Waverley plans to consolidate the trust’s extensive Windows server estate by two thirds using a HP Blade Server Infrastructure and virtualisation.
The result should be cheaper to run – for a start, the new infrastructure is expected to use just a quarter of the energy of the old one. A further major benefit will be the flexibility to run applications across UNIX and Windows.
CSA Waverley managing director Andrew Boyle, says: “Virtualisation is helping the NHS to stay one step ahead of data and disaster recovery needs. This is another example of how a creative approach can unlock huge value to keep an organisation ahead of its operational needs.”
Picking up speed
Trusts working with Esteem Systems report similar experiences. Last year, West Wales General Hospital in Carmarthen cut its server numbers from 40 to three when it implemented a virtualised environment. Log-in times are down, clinical and clerical staff can access information much more quickly, and the trust has a robust disaster recovery plan in place.
Paul Solloway, IM&T services manager, adds: “Our in-house IM&T department will also benefit from seamless rapid deployment and upgrades of servers which can now be done with zero downtime.” And that means no disruption to clerical and medical staff.
In another project, North Tees and Hartlepool NHS Foundation Trust is building a virtualised infrastructure, this time driven by the need to save costs by reducing power, cooling and heating. It is cutting its servers by one third.
So virtualisation seesm to be the flavour of the month right now. Those looking for guidance should look to NIMM, which is on the NHS access only pages of NHS Connecting for Health’s website. But Ian Masters, UK sales and marketing director of disaster recovery software providers Double-Take, offers a word of warning.
“There is a lot of talk around virtualisation, specifically to do with disaster recovery and business continuity,” he says. “There are some very good tools around but they all require servers to be SAN attached - and that may not be the route that every organisation wants to go in. There are other solutions to some of these problems.”