Operate/innovate: New directions for shared services

What's Your Question � Ask Our Experts..

In the early 1990s, the buzz in technology circles focussed on the new breed of distributed systems. Widely seen as the way forward, so-called ‘commodity' hardware could be bought at relatively low cost and deployed at speed to support new applications as and when they were required.

Unfortunately, the result of that enthusiasm for distributed systems, at many organisations, was ‘server sprawl' - the rapid proliferation of numerous computing systems that each make their own demands in terms of power, cooling and systems administration.

It's therefore no surprise that many are now seeking to return to a centralised, consolidated architecture, where fewer, more powerful systems are responsible for performing the bulk of computing tasks.

"It makes perfect sense if you're wrestling with issues of total cost of ownership, management overheads, utilisation rates and the environmental burden of powering and cooling the IT infrastructure," says Martin Boakes, a System Z specialist at Logicalis. He reports an upsurge in organisations looking to consolidate and centralise IT services, on mainframe systems in particular.

What makes even more sense in some sectors, he says, is for organisations to join forces and work together on building a ‘shared service' architecture that fulfills the computing requirements for all members of the consortium in one place, he says.

"You get terrific economies of scale if you do this, and as long as all parties can agree on what kind of facilities they need, the kind of service levels they expect and who will be in charge of maintenance and upgrades, it's a very sound proposition."

Nor would they be confined to using just one type of computing platform. Mainframe and high-end Unix systems, he explains, have long offered partitioning technologies that will enable each party in a shared service centre to keep their data separate from that of their partners. And Wintel boxes are fast catching up, as virtualisation technology from companies such as VMWare and Microsoft now enables them to run multiple ‘virtual machines' on a single piece of hardware. This means that most computing platforms can be used by groups of companies to share IT services.

"In a shared services architecture, mainframe computing is suddenly within the reach of organisations that might consider it too expensive a proposition to invest in on their own, but are happy to split the costs with their partners," he says. "But if that's more power than they require, then there's still the other options lower down the foodchain."