What do you do when a global liquor vendor discovers that a change in office means drastically-reduced room for hosting a datacentre? Virtualisation, that’s what. MATTHEW SAINSBURY reports.
Pernod Ricard is a global liquor and wine entity with major labels including Absolut Vodka, which it distributes worldwide. The local branch, Pernod Ricard Pacific (PRP), also represents a number of local brands including Jacob’s Creek wine. The organisation employs over 2000 staff and has an annual turnover of $1.2 billion.
It’s also a regionally-spread organisation. The head office – where marketing and the company’s CEO operate from – is in Sydney; however the main operational office is in Adelaide.
An IT-related challenge for PRP sprung up when it relocated its Adelaide office. Where it once had an entire floor available to host a datacentre, the new location was far more limited – now, PRP could only access 30m2 of space and a maximum power input of 32kW. It called in Novell partner, AkurIT, to help come up with an appropriate solution.
Limited resources
AkurIT director, Steve Ferguson, said virtualisation was a clear solution for the migration of the environment to take account of the more limited resources.
“Typically, it [PRP] would have infrastructure for a development and testing environment, user acceptance environment, and production environment, so it had three lots of infrastructure for each of its applications,” he said.
“What we were able to do is virtualise a lot of those servers into a virtual farm so we had a number of different clusters and we could just virtualise those applications. We were able to cut down the amount of hardware significantly.”
Less hardware means less space used in PRP’s now minimalist datacentre. The entire solution consisted of virtualised infrastructure on IBM BladeCentre hardware, running Suse Linux Enterprise Server with built-in Xen virtualisation. The Xen environment hosts guest instances of Suse Linux Enterprise Server and Windows Windows Server.
It was a typical rollout for AkurIT, but one that has proven benefits and makes customers happy, Ferguson said. It was also relatively pain-free, allowing a smooth implementation of the solution.
“We had some issues around support from a hardware perspective, with things such as driver support,” Ferguson said. “It was just a matter of getting all the componentry to talk and work together, but apart from that it was a reasonably straightforward process from a migration perspective.”
Making a good recovery
AkurIT also deployed PlateSpin Orchestrate to manage the new virtualised environment, and PlateSpin Forge to assist with disaster recovery. This deployment added a number of other benefits to PRP’s new environment, Ferguson claimed.
“One of the issues we had before was the disaster recovery,” he said. “That side of things was at an operational place in the Barossa Valley, where PRP had issues with bandwidth, as well as having to replicate the infrastructure we had in the production datacentre it had across the disaster recovery site.
“One of the main benefits of PlateSpin is it’s an appliance-based solution where we can virtualise up to 25 workloads onto a single appliance, so what we ended up doing is taking the hardware they had at the disaster recovery site and putting it back into the production datacentre and reusing that in a virtual environment.”
The best part of this, from the customer’s perspective, is that the solution allowed PRP to test its disaster recovery more regularly and with greater ease. Prior to the engagement with AkurIT, testing the disaster recovery environment would cause a major production outage – an issue the new technology resolved, Ferguson claimed.
“One of the benefits of Forge is you can actually enable your DR service and test it while your production environment is online – that’s a major benefit for them,” he said.
The end result is that PRP’s datacentre was consolidated to fit within its new, stringent floorspace requirements. At the same time, the solution is reliable and scalable – something that gives any enterprise a sense of security and long-term opportunity. The only issue came from getting third-party vendors to support Novel Suse Linux as a platform, Ferguson claimed.
“A lot of the stuff PRP had running before was certified to different flavours of operating systems – whether it be Windows or Red Hat, or something else – so some issues we had were around getting the vendors to recognise Suse Linux as an enterprise-level Linux platform. We had to work in the back-end with the vendors to try and get them to certify the products.”
In the end, PRP was a happy customer, Ferguson said, and a repeat one. Following the initial rollout, AkurIT has maintained a solid relationships and provided ongoing services and support.