Virtualisation case study: Lakeland Dairies

23 Jun 2010275 Views

Share on FacebookTweet about this on TwitterShare on LinkedInShare on Google+Pin on PinterestShare on RedditEmail this to someone

Share on FacebookTweet about this on TwitterShare on LinkedInShare on Google+Pin on PinterestShare on RedditEmail this to someone

Lakeland Dairies is one of Ireland’s largest dairy processing co-operatives, employing around 650 people and operating out of five sites across the northern half of the island and the UK. The group processes about 1 billion litres of milk annually into a range of dairy food service products and functional food ingredients. As well as catering for domestic consumption, Lakeland exports its produce to international markets and is also a leading provider to Ireland’s Kerrygold brand.


The business had grown organically and through merger and acquisition. Over the last decade in particular, there were three significant acquisitions where three sites with data centres had to be absorbed into the group. This compounded a problem for a small IT department of 10 people that was already trying to manage a dispersed environment that ran across multiple locations.

Like a lot of large companies, Lakeland bought applications to support the business and run on dedicated servers. One-on-one configurations led to server sprawl and inefficiencies in the data centres. “That was the natural order of things but we wanted to try and get more control of the different environments as effectively and as quickly as possible,” says Turlough Farrelly, Lakeland Dairies’ IT manager. “The best way to do it was to consolidate into a group data centre and virtualise the servers.”

With its silos of information and multiple systems, the company’s disaster recovery plan was also flawed. “The time to recover key applications had crept out to five days which is far too long in this day and age,” says Farrelly.


With each merger and acquisition the immediate priority was to bring the new business on board and keep the different systems running. Long-term, Farrelly and his colleagues knew that an effective enterprise-wide IT environment would benefit from running the same systems, managed with consistent control mechanisms.

Working closely with Microsoft and following participation in Microsoft Longhorn academy, the strategic time to make the move was decided as existing hardware came towards the end of its natural life cycle and the business began to outgrow the wide area network. The first step taken was to upgrade the WAN by deploying new point to point radio links across the key sites. This provided the foundation for major system consolidations and centralisation to the group data centre in the company’s headquarters in Killeshandra, Co Cavan. Rather than simply refresh what was there, the company then took the opportunity to virtualise.

It would host the company’s ERP applications as well as business critical systems for milk production and manufacturing. A backup site was established in Bailieborough.


Working in partnership with Microsoft and Origina, Lakeland set about consolidating the entire IT estate, implementing a more effective disaster recovery solution along the way.

Virtualisation software was the cornerstone of the new strategy, providing the platform to run the applications out of the new data centre. As a Microsoft house, it made sense for Lakeland to explore Microsoft products. This was back in 2008 when Hyper-V was new a contender, pitched against the market dominance of VMware.

The R2 release of Hyper-V now includes Live Migration and is now technically on par with the competition. Hyper-V gives us the hypervisor, high availability for both planned and unplanned downtime and disaster recovery all for the price of our Windows licenses, which we would have purchased regardless of who’s solution we went with.

Farrelly had no doubt that VMware was an excellent product but as an investment it would have sat uneasily with his objectives and his budget. “It would have meant having to introduce another software layer into the company and a technology that no one in my department knew much about,” he says. “Hyper-V was immature but quickly gaining momentum. Because we had a big Microsoft footprint and strong in-house expertise, it made sense for us to go with it.”

The migration was made easier by careful preparation and a good working relationship with Microsoft evangelist, Dave Northey. Key members of the IT team had trained up on virtualisation and conducted a number of trials with the technology. By early 2009, the new environment was up and running after three and half months of intense work that started with the virtualisation of non-business critical applications, such as test and development and training environments.

“Once you have concentrated on one system and mastered it, the rest becomes much more straightforward. You can literally take any physical environment and move it over in a very controlled way,” Farrelly says.

The next challenge was the day-today management of the consolidated data centre. Once up and running, the IT team had to get used to a different approach to the infrastructure and applications. “With traditional environments you walk into the data centre, go to a box and see if it’s working. That’s all gone out the door. With one server running multiple virtual environments, it’s a different mindset and a very different approach,” he says.

Lakeland installed Microsoft System Centre suite at the same time as Windows Server 2008 Data Centre, giving them centralised control of the new virtual estate and tools to monitor and maintain it. “You can see very quickly if it’s healthy or not. It was a crucial investment.” he says.


The Hyper-V deployment has achieved a number of prime objectives, including the topline benefits associated with virtualisation – reduced costs, lower carbon emissions and improved IT service delivery. A sprawling environment of 35 servers has been consolidated into two virtualised servers in main data centre, replicated to another virtual server in its disaster recovery centre.

The numbers make a compelling case for even the most cynical business managers:

  • 70pc saving on power and cooling costs
  • Disaster recovery reduced from five days to less than two hours

An integral part of the revitalised disaster recovery plan is the Compellent SAN (Storage Area Network) implemented by Origina. Using SAN-based replication, snapshots of critical data is replicated over the wide area network to the virtualised server in Bailieborough.

Not just a cost-cutting exercise, the move to a single data centre and a virtualised environment has achieved Farrelly’s other goal of simplifying IT management. And by consolidating applications and virtualising the hardware, the estate is much more reliable with reduced downtime. Improved server efficiency is a recognised win for organisations that virtualise and Lakeland is no exception. Old servers ran at 11pc utilisation. The two virtual servers running at 26pc leaves room for future growth and deployments.


Farrelly and the wider Lakeland business have been delighted with the migration to a virtualised environment and the business is already seeing very tangible benefits that are expected to grow further as the technology beds in. It has created a much more agile infrastructure that is future-proofed for fast change. Servers can be provisioned in a matter of minutes rather than days. The success of the project has also been recognised externally, with Origina picking up "Best SMB Solutions Provider" at the 2010 European IT Excellence Awards.