The complexity of today's company's IT departments must provide the required information technology infrastructure to support corporate growth and change.
Enterprise data integration is essential for companies that have multiple databases, multiple data types, multiple computing platforms, multiple applications and want it all to come together.
IBM is currently producing monthly measurement packs for customers to service level agreements and needed a data warehouse to meet the following challenges: multiple data sources and platforms, small learning curve, accessible with various query/OLAP tools and completed within three months.
The data warehouse solution has to extract, transform, validate and move all this data into a scalable, efficient, fast warehouse engine. Sybase's specialised warehouse engine Adaptive Server IQ forms the backbone to this solution across the entire enterprise.
"There are four steps in the data arehouse process: design, transform and move, store and access. The purpose of data warehouse design is to increase productivity through graphical representation of DW models and automated generation of Data Definition Language scripts.
"This provides a 'blueprint' for the data warehouse and ensures data consistency through importing existing operational information and re-using of metadata throughout DW construction.
"Other reasons to embark on a detailed design phase are to preserve data integrity manage the DW metadata for accurate multi-dimensional modeling, to interface with DSS tools and to reverse engineer the structure of existing data sources. The design tool providing this functionality and used for IBM was Sybase's Warehouse Architect," says Estelle de Beer, Business Intelligence product manager, Sybase SA.
There are several business benefits to a data warehouse:
Analyse large amounts of business information quickly
Enables flexible business queries any way they are needed
Allows economic management of business information
IT staff are relieved of the ongoing generation of reports for the entire organisation
A vital part of the building of the warehouse is the ability to source the date, transform and load the data in the correct format. Getting this underlying infrastructure right is the make or break of a warehouse project.
Sybase's ETL option 'DataStage' transforms or moves data from the source platform to the Sybase Adaptive Server IQ database where it is accessed via the Internet using various query/OLAP tools. IBM will use Seagate/Info as the OLAP front end for their customer access.
The IBM project is still in the development and implementation process. " We are at the first stage and hope to complete the project at the end of February 2000. The initial implementation of a data warehouse takes some time. Sybase consultants had to gather requirements, understand our line of business, design the warehouse model, train our staff and assist with implementation," says Louis Kleynhans, IT specialist contracted to IBM.
Share