|
|
ETI and The Complexities of Database Mapping
A major drug company (who wishes to remain anonymous) needed to migrate and convert
the stored results from more than 100 individual drug studies. They wanted to be able to
re-examine this data to gain more information from them. The data was stored in flat files
and systems using the network data model in more than 100 different database applications
on platforms such as TurboImage and SAS. The target platform was Oracle.
The drug company first paid a vendor more than $300,000 to develop an application
to convert the databases. Not one database was converted. The transformations were too
complex and the environments too diverse. Then the company discovered Austin Texas
corporation ETI (Evolutionary Technologies International). ETI's product 'ETI·EXTRACT'
automates the generation and execution of data interface programs as well as capturing
the metadata.
The project required some of the most complex mappings that had ever been done by ETI·EXTRACT.
For example, one mapping required one and a half pages of English text to describe it.
"It would have been inconceivable to do this without ETI·EXTRACT"
said a senior MIS manager. Using the tool, the drug company generated 13 million lines of
executable and well-documented code. The company estimated that it would have taken 325
man-years to produce the code without ETI·EXTRACT. Instead the project was completed with
8 man-years effort, giving a 40-1 productivity improvement over hand-coding.
|