y provides telecommunications equipment to GTE as a whole, as well as to other major telecommunications compani
es.
When it realized what a large undertaking data-warehouse development requires, GTE Supply sought help from experienced consultants. It chose Information Builders to supply development services and an integrated set of tools for applications development, data access, data transformation, data migration, and reporting. GTE Supply also used the Incremental Warehouse Methodology (IWM), pioneered by Earl Hadden, a pragmatic approach to warehouse implementation and design.
"The IWM encourages a
phased construction
process that achieves results quickly," GTE's Copeland says. "Companies begin with small, manageable pieces. Each increment provides tangible business value and a demonstrable return on investment."
GTE Supply's first four data marts -- currently totaling 20 GB -- handle purchasing, inventory, forecasting and planning, and measurement. Payback on each data mart is expected within a few months. (The company declines to say how much the four data
marts cost to launch.)
"Our developers are getting more efficient, reducing the cycle time with each data mart," Copeland says. "The first data marts were delivered in about 90 days apiece. By the end of this year, we expect to be delivering new increments in 60 days or less."
If you're a reseller looking to cut data-mart development time, check out packaged solutions from the SAS Institute, Platinum Technology, Prism Solutions, Sybase, or Information Builders. Packaged offerings typically include the database itself, middleware, front-end decision-support and data-mining tools, and data-extraction tools. Some packages include usage-analysis tools and query governors as well.
Taking packaged data marts a step further, some vendors are starting to work with VARs to release packaged data marts tailored for specific vertical markets. Such solutions include not just the bundled warehouse hardware and software products but also data models, data-extraction scripts, and prewritten queries and repo
rts designed for target industries such as financial services, retail, and manufacturing. Such offerings are particularly attractive to customers because 80 percent of the development work has already been done. The last 20 percent? That's the niche that VARs and resellers hold onto.
* Data marts allow
for incremental development and rapid deployment.
* Data marts reduce
dependency on network services.
* Data marts improve
performance by storing data closer to users.
* Data marts use
relatively low-priced hardware and software.
Source: DataBase Associates International
illustration_link (54 Kbytes)
