It's a problem that has dogged systems administrators and technology managers since the very first personal computers were installed in corporate environments. No one outside of the information technology team really cares about IT until it falls over, and when it does, suddenly everyone wants to know why.
It's also a problem with which Tom Freer, manager of information technology at not for profit health-care provider Ozcare, is familiar.
For the past seven years, he's been centralising and consolidating the organisation's IT infrastructure so that the rest of the company can forget it's there.
At the same time, Freer's long-term plan has been to make the IT department a saving rather than a cost centre, which in the short term has meant securing funding for a series of hard to sell projects.
"If the systems go offline, a lot of people just stop working, and it's then that they realise how much they do rely on the computer network and the internet and all their applications," Freer says.
"Unless that happens, it can be hard to convince people to spend money."
The first phase of his plan began in 2002 and involved a three-year server centralisation project in which disparate servers in 20 separate locations were brought into a Brisbane data centre.
Using a mixture of Citrix servers, a Hitachi Data Systems storage area network and a raft of stand-alone applications, Freer managed to cut back on maintenance costs, improve up-time and keep most of his small IT team on the job rather than the road.
But having improved productivity by doing away with the decentralised network, last year Freer found himself having to convince management to loosen the purse strings once again in order to implement a disaster recovery solution: "From a technology perspective we knew it had to be done, but from a business perspective it wasn't considered an issue," he says.
Freer was working as a contractor at the time, and commissioned an independent IT risk assessment and audit to let management decide whether losing access to core systems would cost the organisation.
"Maybe it was a bit sneaky," he says, "but there was a lot of money involved in the project and I needed to present them with a good argument. Certainly now that it's in place, the chief executive officer and management are a lot more comfortable knowing that if something catastrophic happened at headquarters, the rest of the company would keep on operating and we would not lose any data."
But by the end of 2006, the newly installed and backed-up data centre was still experiencing issues.
"We were running out of storage space - and trying to achieve more and more with a fixed IT team of six people," Freer says.
"We had 35 servers, 500 users, storage filling up too quickly, back-ups not working . . . lots of little stuff happening all the time adding up to one big problem."
Freer needed to find a way to expand the organisation's storage capacity, improve the back-up process and replace a series of servers that were nearing the end of their lifecycle.
"If we just went ahead and replaced the servers, we'd still have these support issues, and I knew I wasn't going to get any more staff in a hurry."
Sitting down with pen and paper in hand, Freer set about redesigning the company's underlying server architecture, opting to replace the ageing infrastructure with virtual technology.
And for once he didn't need to ask for extra money to implement the technology he needed.
He identified 12 servers that were running way below their actual capacity and opted to consolidate the applications using VMware software.
The redesign began in February. Freer spent most of March replicating applications from stand-alone servers into the consolidated configuration and back again, and then sent the entire new system live in April.
"The switchover was quite seamless, but we had done a lot of planning and a lot of testing all the way along," Freer says.
"Operationally there has been feedback from management saying that the systems are more available and more reliable, which is a bonus."
Ultimately Freer was able to cut back the number of servers from 35 to 18, although he opted to quarantine the Exchange, SQL and Citrix servers because of concerns about availability.
"There was always a hesitation to go gung-ho and virtualise everything," Freer says.
"The VMware we're using has provided a more stable environment because we are able to seamlessly fail over between our headquarters and the disaster recovery system we'd installed previously.
"Operationally it allows us to do maintenance during work hours."
And the beauty of it was that Freer was able to carry out a project for which he didn't have to justify a budget.
It's something he'd like to keep on doing well into the future.
"Now we're looking at storage virtualisation to enable us to add additional storage arrays when the current platform gets to the end of its life cycle," he says.
Australian Financial Review
© Fairfax Business Media