Australia and New Zealand are ahead of the international state of play when it comes to planning for grid computing, according to a survey commissioned by Oracle.
On an index measuring the average place of organisations on the “adoption lifecycle” for grid computing, Australia/New Zealand organisations rate 3.9 as against an international average of 3.2.
They also lead on a “foundation readiness” index, rating 7.3 against 6.9 for organisations worldwide. Paradoxically, however, Australia and NZ rate lower than average on knowledge of and interest in the technology (6.1 against 6.3).
“The adoption lifecycle index is an indicator of the degree to which interest and intention are being turned into tangible action,” the survey report explains.
“The foundation index measures the degree to which organisations have standardised and consolidated within their IT infrastructures [as a basis for grid computing]. The knowledge and interest index is based on measurement of understanding (self-declared) together with assessments from IT professionals of the level of benefit offered by virtualisation and grid related technologies.”
According to the survey, conducted for Oracle by UK-based analyst Quocirca, our 3.9 adoption rating puts us and the Aussies fourth in the world for grid computing behind the US (4.3), Germany (4.3) and the Nordic region (4.7).
Oracle nominates the NZ Public Trust as one New Zealand organisation that is effectively using ‘grid computing’ with Oracle’s Real Application Clusters 10g software. Three of its core applications, data warehousing, document management and CRM, have so far been moved into grid mode.
Public Trust’s principal database specialist Simon Brock says “The new infrastructure has allowed us to cost-effectively scale our processing capacity in line with future growth, while delivering significant improvements in the reliability and performance of our document management system.”
But Brock and fellow database specialist Rod Orr are hesitant to describe the Public Trust’s rack of six HP servers as a “grid”. “We don’t [care] whether it’s a grid or a cluster really,” says Orr. As long as it gets the work done for the number of users that need it that’s all that matters, he says.
The term “grid” has connotations of a service like electricity supply, says Brock — a utility provided from a large number of unknown sources and mixed uniformly into an undifferentiated resource. “I’d describe ours as [being] more like a kitchen with six chefs. I take an order in and give it to the chef [who] looks like he’s least busy.”
A major advantage gained by the move to an Oracle grid/cluster computing solution is the time saved on Public Trust’s month-end processing, says Brock.
“One of the month-end reports, summarising timesheets, previously took over 29 hours to run but now takes an unbelievable 72 seconds. Aggregated data summaries, known as materialised views, can now be created four to eight times faster than they used to be.”
But the change in architecture has been accompanied by an upgrade in application software and processor power and it would be hard to disentangle the factors in the net performance improvement and say how much was due to the clustering, he says.
In Australia, insurance company NRMA reports notable performance improvement from grid-enabling its call centre and membership applications.