IBM expands corporate search ambitions

By the end of the year, Big Blue intends to release an update to its corporate information-management tools, which are designed to bring order to potentially thousands of data sources in a company's network.

Code-named Serrano, the product will use technologies including artificial intelligence and data mining to derive more meaning from corporate documents. It will also have a revamped search engine and front-end tool designed to make hunting for company information as straightforward as searching the Web, according to IBM.

Since its introduction more than two years ago, the company's WebSphere Information Integrator (which used to be called DB2 Information Integrator) has garnered about 1,700 clients. The basis of the product line is the ability to query several different data sources at the same time. IBM has added several tools, such as search interface, on top of that foundation.

IBM is targeting information retrieval within company networks, rather than the Internet as a whole, because finding information easily remains a persistent headache for corporate customers, said Nelson Mattos, vice president of information integration at IBM.

"The problem is not the Web. The problem is people getting the information to get their job done," Mattos said. "Why can't you give me the Google for the enterprise? That's really where the problem is."

Among the planned enhancements in Serrano--named after a spicy pepper--are automated tools for spotting patterns across different sources of information. The software will use pattern-recognition techniques, some of them culled from IBM's research labs, Mattos said.

For example, a pharmaceutical company may use the software to automatically find references in company documents to potentially dangerous drug combinations. The software will be able to analyze several documents and flag those relationships, which may help in the drug research process, Mattos explained.

The Serrano software also will analyze the content of documents. That would greatly aid the searching capabilities in the OmniFind edition of the product because search results will not be based solely on keywords, Mattos said. People also will be able to update documents retrieved from searches.

IBM intends to publish interfaces that allow third-party companies to build applications that use the OmniFind features in specific industries, such as intelligence gathering, he said.

Finally, Websphere Information Integrator will introduce tools aimed at software programmers and database administrators for building applications and managing information. Those tools are intended to help developers more easily build applications that tap into both "structured" data, or information stored in tables in relational databases, and "unstructured" sources, such as content management systems for handling documents.