IBM Corp. and Microsoft Corp. are working to pull BI (business intelligence) and knowledge management under a single umbrella, with the goal of providing a new generation of BI that includes both structured and unstructured data.
The end result, the companies say, is to give users better access to more of their data.
IBM’s research arm is at work on software that will make all the information in an enterprise available to employees, according to Anan Jhingran, head of computer science at IBM’s Almaden, Calif., research labs. The new model takes structured and unstructured data, analyzes it, and correlates it, he said.
“Unstructured and structured data are not islands, and we need to bring them together to actually derive useful knowledge,” Jhingran added. “In the future no data type or source of data will be an island.”
Jhingran is calling this initiative a marriage of traditional BI data, which is structured, and knowledge management, in which much of the data is unstructured. For instance, any given employee will have loads of unstructured information in applications such as Word and Excel. Combining that unstructured information with a structured data typically stored in databases creates a more inclusive view of the overall knowledge available.
Another key part of IBM’s strategy is to use the Internet as a data source. That, however, requires natural language processing so that software can understand more than key words in a phrase or string of words. The idea is to make machines realize the difference between statements such as “the interview was great” and “the interview was a great disappointment.”
“The data that sits within the corporate intranet is extremely valuable, and the knowledge on the Internet is valuable. The problem is that the Internet, being free, has a lot of junk on it. We are looking to sift through the junk of the Web and focus on what is truly valuable,” Jhingran said. He continued that IBM will achieve this with scalable algorithms, such as a text-processing algorithm.
Pat Selinger, an IBM fellow on the development side of this effort, said that she plans to have a beta form of the software in testers’ hands by the end of this year. A final version of the software should be available sometime next year, and a second iteration out there by the end of next year as well.
She added that as of now IBM has not decided if the technology originally will be delivered stand-alone or pulled into another product.
For its part, Microsoft’s vice-president Jeff Raikes described an initiative within Microsoft Research that the company is calling Organizational Intelligence, at its Future Forum research event last week, held on the company’s corporate campus in Redmond, Wash.
The basic idea, Raikes said, is to enable applications to communicate more effectively and to produce information that users can then take action upon.
In Microsoft’s plan, XML is the core technology that facilitates the flow of information. “XML is going to revolutionize the way information can be passed around within the organization,” Raikes added.
Using XML that way leads to the bigger concept of Organizational Intelligence, which focuses on making meetings more effective, mobility, organizational sharing of information, business processes, and the ability to view both structured and unstructured data types via what Microsoft calls a knowledge portal, he said.
An Oracle representative said the Redwood Shores, California-based company already has technologies that blend structured and unstructured data.
Oracle iFS (Internet File System), for instance, is the company’s approach to getting at a variety of data types, including flat files, audio, and video files, among others. The representative also said that Oracle has reporting tools for BI, interMedia for parsing unstructured data, and Ultra Search for locating data anywhere its stored. Like IBM and Microsoft, Oracle also feeds the data through a portal for single-interface access.
Mike Schiff, vice president of BI and e-business at Sterling, Virginia-based market research firm Current Analysis, said that the idea is not altogether new, though the current approaches are.
“This is not a radical departure but rather a natural evolution that was never precluded,” Schiff said.
He added that the combining of structured and unstructured data promises to make users more productive as a result of having more and better decision-making information at their fingertips.
“A lot of people are trying to do this. The days of just rows and columns in a database are ending,” Schiff said.