Saturday, January 08, 2011
Enterprise Architecture and Historic Intelligence
Anne Thomas Manes of Gartner once declared that SOA is dead. Today, I am declaring that Data Warehousing is also dead or at least braindead...
The data warehouse is dead. Age-old methods of gathering and storing data into centralized warehouses, transforming it into information and generating reports, are inadequate and do not deliver either speed or intelligence to the enterprise.
The cost to create a single data warehouse in a Fortune 100 shop is estimated to be nearly $5 million, with additional yearly maintenance costs running around 20% of that figure. Add to that the complexity of report generation, married with the inability to easily connect disparate information sources, and you have an information technology phenomenon of unprecedented proportions that is not just costly but mostly irrelevant to the day-to-day needs of business managers.
It is important to acknowledge that there will be lots of people continuing to waste money keeping things on life support. Anyone care to guess how many mainframes IBM has sold to customers within the last five years that never had one in the past? Know of any software companies that started to develop a new product offering written in COBOL or PL/1?
Just because something is on life support doesn't make it alive. Pull the plug and let's see what happens...
| | View blog reactionsThe data warehouse is dead. Age-old methods of gathering and storing data into centralized warehouses, transforming it into information and generating reports, are inadequate and do not deliver either speed or intelligence to the enterprise.
The cost to create a single data warehouse in a Fortune 100 shop is estimated to be nearly $5 million, with additional yearly maintenance costs running around 20% of that figure. Add to that the complexity of report generation, married with the inability to easily connect disparate information sources, and you have an information technology phenomenon of unprecedented proportions that is not just costly but mostly irrelevant to the day-to-day needs of business managers.
It is important to acknowledge that there will be lots of people continuing to waste money keeping things on life support. Anyone care to guess how many mainframes IBM has sold to customers within the last five years that never had one in the past? Know of any software companies that started to develop a new product offering written in COBOL or PL/1?
Just because something is on life support doesn't make it alive. Pull the plug and let's see what happens...
