IBM to push 'cloud computing,' using data from afar

The IBM strategy, to be announced Thursday, seeks to exploit the technical work and commercial interest in large data centers that can be run more efficiently, searched for information and programmed from remote locations over the Internet.

This model of Internet-based supercomputing is known as cloud computing because vast stores of information and processing resources can be tapped from afar--by a laptop personal computer, cell phone or other device.

IBM is calling its initiative Blue Cloud. Most of the basic software needed for cloud computing is open source, meaning that the code is freely available and can be modified by users. The hardware used in the data centers is typically many thousands of industry-standard server computers, powered by processors made by Intel or Advanced Micro Devices, and produced by many hardware makers.

But IBM, analysts say, is trying to position itself as a leader in the corporate market for cloud computing, which many specialists regard as the next evolutionary step in information technology. The business strategy, they say, is to sell more IBM hardware, software and services tailored for cloud computing. Starting in spring 2008, IBM will offer versions of its server computers, including mainframes, that are adapted for cloud computing.

The game plan, IBM executives say, is similar to the one the company followed in supporting Linux, an open-source operating system and an alternative to Microsoft's operating systems. IBM's endorsement of Linux, which began in 2000 and included investments in technical development and marketing, sped the adoption of that technology among corporate customers.

"To me, this feels like Linux in 2000," said William M. Zeitler, senior vice president in charge of the systems and technology group.

IBM now has 200 researchers working on cloud technology, and Zeitler said the company had a staged plan over the next three years that would involve a large investment, though he would not elaborate on the amount.

Several customers, including corporations and government agencies, have been working with IBM in pilot projects on cloud computing. Zeitler did not identify the companies, but said, "Large financial services companies are going to be among the first to be interested."

Companies with fast-growing data centers, like banks and securities firms, are facing the same headaches as the large Internet companies, like Google and Yahoo. Efficiency, power consumption and management costs are mounting. And companies of all kinds are increasingly adopting some of the technologies of Internet companies like searching, mobile commerce and communication, and collaboration tools like blogs, wikis and social networks.

In recent years, IBM has championed efforts to make data centers more efficient and to centralize more computing tasks in the data centers, with desktops and devices tapping in. These have had names like "autonomic," "utility" and grid computing.

Those concepts and research efforts have made a contribution to cloud computing. Experts say tools have been added to spread computing tasks across clusters of many machines and to make programming simpler. Advances likely to broaden the reach of cloud computing have often come from researchers tackling the challenges posed by Internet searches.

"In some ways, the cloud is a natural next step from the grid-utility model," said Frank Gens, an analyst at the research firm IDC. "What's different is the Google programming model, and that really opens things up. You don't have to be a Stanford or Carnegie Mellon Ph.D. to program cloud applications."

The software that IBM is packaging in its cloud offering is called Hadoop, running on the Linux operating system. Hadoop is based on an open-source search project called Nutch, and an open-source version of Google's MapReduce software for spreading complex computer tasks across clusters of machines.