Divergent Data Converge in Data
Warehouse with Web Interface

November 8, 1996

By Jeffery Kahn, [email protected]

The World Wide Web is essentially an electronic commons. Whether you have a Mac, PC, or workstation, whether you are a computer scientist or computer illiterate, the Web can provide information that is easily accessible to all.

Taking advantage of this electronic commons, the Laboratory has launched a project to bring a disparate plethora of administrative and institutional data to the desktops of employees.

IRIS -- the Integrated Reporting and Information System -- started out as a demonstration project. The Information Systems and Services group, headed by Carl Eben, set out to show that it is not just theoretically possible but actually feasible to make administrative data available on the Web. Sanctioned by Deputy Director Klaus Berkner's Management and Information Systems group, the project was launched in December 1995 by a team including Rose Bolton, Mark Dedlow, and Esther Schroeder. By July, what had been conceived as an experiment and prototype was so successful that it was brought online.

Using the IRIS web site, employees now can access and find information contained in the Ledger, Property, Purchasing, Procard, Training, Account Authorization, and Accounts Payable databases. In the future, IRIS will bring additional information to the Web.

Dedlow described IRIS as a project conceived for the "grassroots user."

"This project makes it easier for people to do their job," he said. "IRIS allows the average person, who is not a computer expert, to tap into databases that had been difficult to access. And it helps that same person find information within these databases that had been difficult to hunt down."

For the user, the beauty of IRIS is its simplicity. For the IRIS team, this simplicity is a mark of their success.

Here as elsewhere, institutional data resides in a digital Tower of Babel. The ziggurat is built using different computer languages, different database systems, and different hardware. IBM mainframes and Sun servers, SQL and COBOL-based programs, FOCUS and Oracle databases coexist at Berkeley Lab, but cannot communicate with one another.

For the IRIS team, the challenge was to take this chaotic mix of information and unite it within a single data infrastructure. The pioneers around the world who now are attempting such feats call this infrastructure a "data warehouse."

Says Dedlow, "The data warehouse coalesces and consolidates all these disparate data sources into a single, uniform environment. Many of these data sources are legacy systems, as old as 20 years, and lack uniform relationships. One may reference an individual by an employee ID number and another reference the same person by name only. The challenge is to extract or add the common essential elements from all sources."

Creating a data warehouse is the invisible first step in delivering this information to employees. It is the second step, the creation of IRIS, that brings this information to the user. IRIS extracts information from the data warehouse and makes it available via the Web. It responds to requests for information by instantly creating, on the spot, hypertext linked Web files.

Dedlow says that IRIS was as challenging a project as the underlying data warehouse. That's because using the Web to allow users to tap into sophisticated relational databases means charting unexplored territory.

Systems analysts who work outside the Web have a rich set of tools for extracting information and displaying it to users. Claris Corp.'s Filemaker is a good example. Filemaker database developers use a point and click graphical interface to build their applications. But in the world of the Web, the rapid development tools either don't exist or are very primitive.

"We had to invent how to do this," said Dedlow. "A year from now, I imagine that Web developers will be users of Web tools. Today, we're inventing our own. That's what IRIS required."

IRIS and the data warehouse are in their infancy. In the months ahead, additional sets of data will come online.

For instance, right now to access purchasing information, you need a purchase order number. Soon, you can check purchasing information out by a search using a vendor or buyer name. Other data will be added to the system such as stores order information, travel information, and accounts payable information that is updated daily.

The Information Systems and Services group believes this project represents a great leap for institutional data and those who must have access to it. It's a leap that takes data, translates it into information, and delivers it to the customers' desktop.

Access to IRIS is for Berkeley Lab employees only and is controlled by requiring that users log in with their unix account login and password. IRIS can be found on the Web at http://www-iss.lbl.gov/ and you can contact Mark Dedlow at [email protected] for further information.

Search | Home | Questions