Showing posts with label Data Warehouse. Show all posts
Showing posts with label Data Warehouse. Show all posts

Thursday, October 9, 2008

IT as a companion to business process flow...

Typical an organization will have many business processes one for each purpose. Some of the main process are sourcing, customer relation management, finance, human resource management etc. Most of the ERPs provide these functionalities as a part of them. Not all vendors are good in all the processes. Some are leaders in some process. For example Oracle is good in CRM, Hyperion is good in Finance Planning etc. Apart from these processes, there are processes in most of the Organizations, that can be automated. It is for this reason we see many custom Java based applications that cater to each of these processes. All business processes follow a certain order. At every point there should be an application that caters to that business process. This application can be a ERP module or custom Java application. Such a seam less solution will allow the company to capture information at each point and leverage the data.

Once we have data in applications, it can be transferred from it to a data warehouse using standard ETL tools. A reporting layer on top of the Data Warehouse can allow users to see trends. For example a report on customer-related-data from a CRM, will show the customer trend. Reports can be a set of standard corporate reports available for all users or adhoc reports of each user. Value of information will reduce with time. To have more valuable information, Data Warehouse needs to be refreshed more frequently.

Tuesday, September 23, 2008

Life cycle for Right-time Data Wareshouse:

Ideally everybody should start with a traditional Data Warehouse. This stores data of historic value. The delay between source data base and Data Wareshouse depends on the refresh freequency. This delay can be reduced by having Data Warehouse refreshed more freequencty. Freequency of refresh should be decided by the business requirement. Eventually one arrives at Right-time Data Warehouse.
There are tools available in the market that help in building a Right-time Data Warehouse. Infosphere from IBM will help in a building a Right-time Data Warhouse.

It's Right-time and NOT Real-time Data Wareshousing...

People often use 'Real-Time' to indicate quick turn-arround time. In reality, a real-time system has a turn-arround time to the order of nanoseconds. It is good to have a Real-time data warehouse. However due technological limitations we can not have a real-time data warehouse. It is apt to call a Data warehouse with less dalay as a Right-time Data  Warehouse.