Ballarat Health Services (BHS) has ditched a bunch of whitebox servers and disparate storage and backup devices for a standardized storage environment.
Describing the old IT environment as "chronically under-invested" BHS director of IT, Paul Jurman, said it was rife with disparate legacy systems.
The hospital services 200,000 people within the 48,000 square kilometre Victorian Grampians region and generates $350 to $400 million annually, despite that it has limited access to medical and IT resources compared to its metropolitan cousins.
"Our server infrastructure was characterised by an aging fleet of no-name servers, we didn't have a standard operating system or hardware platform and our maintenance was unscheduled and disruptive to the business," Jurman said.
"Our backup systems were disparate, unmanaged, and incomplete in some instances, and our disaster recovery and business continuity capabilities were minimal."
In this environment, Jurman said support was non-existent. As a result, BHS deployed VMWare servers and Hitachi SAN storage for its primary and secondary data centre sites and consolidated smaller existing data centres and POP into the facilities.
Server Iron switches were provisioned and tested in both sites, while the SAN linked the data centres.
Jurman said the project was sold to executives under risk mitigation rather than cost savings, while the funding was supplied by the $420 million Victorian HealthSmart initiative as a hardware refresh and capital project.
"The project was designed to make systems more reliable, cost less to operate, and mitigate effects of internal and external disaster to minimize disruption to the business," he said.
The hospital's Oracle-based patient administration database and other mission-critical systems are running on the virtualised system with minimal technical issues, while IT has supported clustering it hasn't gone into production yet.
Virtual servers account for 55 percent of the server pool and this figure is increasing with migration rates between 5 to 7 percent per month.
Storage and server availability has risen from 80 to 99.5 percent and scheduled and unscheduled maintenance has been almost eliminated.
"We developed a charge-back mechanism at the start to reduce server proliferation and a loss of economy of scale," Jurman said, adding that BHS technical staff were trained in Unix, Linux and ITIL's change mechanisms.
Project rollout and core system migration was completed within two months, with BHS offering a commercially hosted virtual service through Bullet Proof Networks to utilize the extra capacity and agility; However, Jurman warned that IT must define SLAs through internal network testing before it can be commercially viable.
According to Jurman, IT managers should establish a strong cultural fit through upfront negotiation, and must diligently negotiate software licensing to ensure project success.
He said that virtualization technologies, hardware procurements and implementations are extremely sensitive to hardware compatibility and processor matching.