IBM EIP Looks to Ease Corporate Content Management
                                                                  September 12, 2000 (9:55 a.m. EST)
                                                         By Andrew Rafalaf, Wall Street & Technology Online

In a time not too long ago, the word "portal" referred simply to those Internet sites like Yahoo and Excite, your windows to the World Wide Web. Today, with the expanded use of Internet technologies within corporate environments, enterprise portals have emerged, offering access to all the data stored within a corporation. A broker, for example, can access from one browser-based screen a client's holdings and investment history, securities information, and related research reports without having to toggle between various systems. That's the intent anyway.

In order to facilitate the construction of these portals, IBM has released Enterprise Information Portal (EIP) Version 7, which intends to ease the integration of numerous data sources into the portal's platform. As companies have grown over time, they've resorted to storing data in all sorts of formats, whether it be located in relational databases or in unstructured data formats like e-mail, audio and video. EIP strives to make all those data types accessible to the portal's main servers no matter the manner in which the data is coded.

"The primary concept behind EIP is to help customers to get to various forms of data they have and to filter it, and then to make that available in a simple and easy method for programming models," ex-plains Teresa Whittle, worldwide content product marketing manager. "The theory is a single point of access to information from various sources."

The first commercially available version of EIP, released late last year, allowed users to connect and search across many types of "unstructured data repositories" like scanned documents, computer generated reports, audio and video. The access to unstructured data is especially important to brokerages that need to maintain a seven year record of all transactions, as stipulated by the SEC. "Not only do they need the transaction data, but a copy of the actual trading screen all corresponding HTML pages," she adds. "And as the amount of data grows and grows, finding the relevant components is becoming more critical."This most recent version of EIP provides users with the ability to mine and parse data found in various relational databases, or ODBC databases, such as IBM's own DB2 or Oracle's 8i. Further, EIP now provides access to CWMD, otherwise known as metadata, that is found in business intelligence reports.

EIP also includes a unified API layer, allowing IT staffs to connect future data sources without having to write an individual API for each one. "As you develop applications and as you need to integrate additional data sources, you don't need to know the language of each source. You just need to know EIP's unified API which will translate behind the scenes to individual data sources," Whittle explains. Data mining and integration systems like EIP have grown more important as the shortage in qualified IT personnel continues. "Our customers are having difficulty finding skilled resources across all the data types," she details. "And, as they grow, acquire and integrate other companies, they are continually faced with new types of data."