CBA CIO champions utility computing

Commonwealth Bank chief information officer Michael Harte has flagged utility computing as the next frontier in the bank's battle to lower IT costs.

The country's biggest bank faced increasing pressure from international competitors who could utilise the cost advantages of distributed systems, according to Harte, who was speaking about CIO level challenges at a business lunch in Sydney yesterday.

These competitors, for example HSBC and Virgin, were large, well-capitalised corporations that had "far better distribution systems" and "far cheaper computing models", said Harte. "Their platforms are global, ours are local."

Harte pointed to the costs of scale advantages companies like Google, with its thousands of servers dispersed around the world, could offer in distributed processing under a utility computing model.

Utility computing has been likened to grid and on-demand computing in that computer resources are provided on a when needed and pay-per-use system. That is customers, for example a bank, would pay an outsourcer for the resources used rather than own the infrastructure themselves. Distributed computing is one service offered through utility computing, as processing tasks can be spread across a multitude of systems and optimised to great effect.

While distributed and utility computing typically made use of large, commoditised infrastructure to offer cost advantages, Harte said the bank had trouble reducing the high cost of ongoing maintenance to systems.

Utility computing has been championed by industry critic Nicholas Carr -- author of Does IT Matter? Information Technology and the Corrosion of Competitive Advantage -- and Oracle chief Larry Ellison, who have both argued for the commoditisation of IT in order to lower costs. However, opponents of the argument say it devalues the strategic role of IT.

Citing Carr and Ellison's views, Harte said he believed the processing and productive -- or utility functions -- of IT should be run "as close to zero as possible" in order to reduce costs.

"As soon as we can commoditise all of that utility function we can actually relocate investment closer to the customer and provide a much richer product functionality and get a return on that much faster.

"But that's again the trade-off between how much money you can pour into keeping the old machines going, versus investing in new logic and putting more [investment] closer and closer to the customer," he said.