"Data warehousing is just beginning as an industry," according to W. H. Inmon, "the father of the data warehouse." So what lies in the long road ahead? Inmon took a look into his data warehousing crystal ball at Spring 2004 TDWI conference in Boston, Mass. The following are some of his predictions.
Managing really large amounts of data
"The biggest factor in making data warehouses expensive is you," Inmon said. Initially, a data warehouse does not require a large storage expense. But after five or 10 years, the storage costs climb and climb to keep up with data growth. The future will require companies to have a combination of disk and other types of storage to archive unused data, according to Inmon.
Archival processing occurs when you start to have enough data that you notice it has its own lifecycle, Inmon said. Some companies archive for audit purposes, but not for access purposes. "You need to be able to access the data and use it in your corporation," Inmon said. While there are not many vendors currently taking an active or useful interest in archival processing, it's an area that you must pay attention to -- forcing vendors to address it.
Exploring the world of metadata
"People get excited about metadata, but it never rises to the top of the pot," Inmon said. Some issues are that it's not distributed, it's not current and it only applies to technical data. The future will require you to abandon the dead central repository concept, ensure data is current as your environment changes, and develop business metadata, he predicted.
Decision support systems (DSS) applications
"Data warehouses by themselves are terribly unexciting things; it's what you can do with them that's impressive," Inmon said. As new business needs are created -- and regulations like the Sarbanes-Oxley Act go into effect -- it will become more and more important to build applications on top of the data warehouse, making data real to an organization, Inmon explained.
Bridging the gap between structured and unstructured data
"There's never been much interaction between structured and unstructured data," Inmon said -- but all that will have to change as companies have to account for unstructured data, which includes e-mail, .txt, .doc, .ppt, .exe, .pdf and numerous other file types. You'll need to organize that data, meaningfully connect it to structured data and be able to search through it, Inmon explained.
Global data management
"Your biggest challenge is keeping up with change as conditions change around the world," Inmon said, adding that you'll have to consolidate data warehouses around the world and create an adaptive environment, so you won't have to make changes to them all day.
For More Information
Tip: Top 10 new data warehousing features in Oracle Database 10g
Featured Topic: Data warehousing
Ask the Experts: DW/BI expert Ian Abramson