Automation know-how Outsourcing meets automation Organize to automate Grid on the job Danger in the new data center

Outsourcing is as old as the hills, and on-demand computing (or utility computing, if you prefer) is becoming a better understood concept in enterprise computing.

Put the two together and you get a fascinating opportunity for IT organizations to "transform the way computing power is consumed," says Jeff Kaplan, managing director at THINKstrategies, a research firm. "Traditional outsourcing was about handing over enterprise problems to a vendor, without any expectation of transforming the operation. With on-demand [outsourcing], that expectation is there."

On-demand outsourcing has the potential to help companies meet today's IT goals: a direct, measurable link to business goals; the ability to change quickly; and increased flexibility in staffing.

Vendors pile in

From the beginning, IBM has been closely associated with pay-by-the-drink computing. This extends to the outsourcing arena. IBM's Strategic Outsourcing Flexible Support Option, introduced early this year, lets companies outsource data center management on a pay-per-use basis. In a departure from previous IBM on-demand outsourcing, customers need not transfer their data center IT assets and employees to the outsourcer.

Here's how it works: IBM's Global Services unit uses its Universal Management Infrastructure (UMI) - a blend of systems management; software deployment and configuration management software; architecture workflows; and outsourcing methodologies - to manage the client company's data center remotely from an IBM facility. First comes an assessment of the customer's data center to map the infrastructure. Next, IBM consolidates and standardizes the data center environment, with the goal of simplifying management. IBM then uses UMI to link the data center to its remote management site.

From that point on, IBM manages the data center, primarily via automated tools. In this fashion, the customer has the option to outsource only parts of its data center - applications, storage devices, servers, networks or some combination thereof. Pricing depends on the scale of the job, the number of components under IBM management and service use volumes; fees are negotiated monthly or quarterly.