Sidebar: Intel Finds Huge ROI in Managing Metadata

Computerworld - While some IT operations are still struggling to help top executives see the cost benefits of managing metadata, such an effort has paid off big for Intel Corp.
After a false start six years ago, the chip maker now estimates that for every $1 it spends on metadata management, it saves $6.
In addition, the company plans this fall to begin combining its enterprise data repository with a code-reuse system that saved it $53 million in development costs last year, said Gregg Wyant, the chief architect in Intel's IT department.
Intel began its foray into metadata management with the launching of an online metadata repository that it built to let developers catalog metadata created in new projects.
However, the company found that it didn't have a process in place at the time to ensure that metadata was refreshed to reflect changes within the data itself, Wyant said.
"Someone scans databases to extract metadata, [but] if you don't have a regular process to examine the content of a repository, look at usage information and ensure the refresh rate is current with data itself, you can run into a situation where metadata becomes out of sync," Wyant said. "The data in the repository will be fresh for one to two quarters and then fall out of sync."
In 2003, the company turned its approach upside-down: Instead of first looking for tools that can grab metadata, the company began creating a process that focuses on when information is refreshed and then aligns metadata capture with that schedule.

Gregg Wyant, the chief architect in Intel's IT department

Today, the enterprise metadata repository contains metadata about Intel customers, products and any program that includes analysis.
And at the beginning of the fourth quarter, Intel plans to combine its enterprise metadata repository with the code-reuse system, which allows developers to check code in and out while tracking how software assets are being used and reused.
"The metadata repository and reuse program combined together will house data for Intel's enterprise architecture," Wyant said. "In the past, when [we] didn't have the information recorded in a common place, you would have several teams doing development not aware that another team was doing the same work."
Intel hopes the combined repository will help cut the time -- which can peak at 30% of a developer's day -- it takes for developers to find information assets, Wyant added.
"I want people to spend more time developing and realizing the value for the information assets than searching for assets," he said. "Usingthe metadata, they can find the right set of information."