Maximum Availability Touts *noMAX Install at Fonterra
October 28, 2008 Alex Woodie
When the large New Zealand dairy company Fonterra upgraded its System i servers recently, it decided to implement a data replication system along the way. The company and its systems integrator, EDS, recently completed the HA setup, and Maximum Availability, which provided the replication software, says the installation was one of the smoothest ever.
Backed by more than 11,000 New Zealand and Australian dairy farmers, Fonterra is the world’s largest exporter of dairy products. Each year, the company produces about 3.7 billion gallons of milk, bringing the company about $7.5 billion in annual sales. With such a large operation, it’s not surprising that Fonterra relies on the IBM System i server to power its business software.
Recently, Fonterra’s Australian subsidiary decided to upgrade its System i server. The subsidiary supported more than 500 Australian users from a System i server based in Melbourne. However, the new server would be based at company headquarters in Auckland, more than 1,200 miles across the Tasman Sea.
According to Maximum Availability, the company did not want to risk the potential downtime associated with performing a full system save, flying the tape across the rough sea, and then restoring the Australian subsidiary’s data and applications on the new server. So instead, the company looked to Auckland-based Maximum Availability to provide a high availability solution.
Maximum Availability’s flagship product, *noMAX, was installed and replication began between the old System i server in Melbourne and the new System i Model 570 in Auckland, on a 10 MB/ps communication line. Fonterra’s 100 GB database–which included 100 libraries, 50,000 objects, and the IFS–were kept in lock-step for four months as the company prepared for the big switch over.
The day finally came for Fonterra’s Australian subsidiary to go live on the new i570 in Auckland. Switchovers, of course, are the Achilles’ Heal of HA products. Problems that remained unnoticed while users are on the comfortable old production machine can suddenly pop up when the roles are reversed and users log onto the backup machine for the first time.
But the trepidation was all for naught in Fonterra’s case. The EDS technicians came in and performed a role swap in *noMAX, and in less than 10 minutes the new i570 became the production machine, relegating the older server in Melbourne to backup status.
The role swap went perfectly smooth, with no tape saves or restores required, according to Simon O’Sullivan, senior vice president with Maximum Availability. In fact, it might have been the smoothest role swap ever for EDS technicians in New Zealand, he says.
“Clearly there is no need to suffer extended outages during upgrade hardware migrations or when moving to a new location–even if the new location is thousands of miles away,” O’Sullivan says. “… *noMAX technology can replicate data in real-time and ensure that go-live is outage-free and without the risk of data loss.”