Let's talk about BI layers - Series 1 of 3
To deliver effective Business Intelligence , organizations need to progressively design and implement a rugged data warehouse with the right enterprise data flow architecture and not to forget a logical data architecture.
In this 3 part series , I would like to take you through the various layers that encompasses - data sources, core data warehouse, data mart layer, data staging and quality, data access , data preparation, metadata repository layer , warehouse management layer , application messaging layer and also the internet/intranet layer .
Why do we need a logical architecture and is this something that can be designed real quick ? Well, the answer to this is both yes and no.
Typically different parts of large business organizations often deal with different customer ,product, different sets of transactions and so on . So it is needed to logically separate the data model by the different business domains , but the implementation of such a logical model can only take place in a phased manner . This needs to the aligned to the layers of the enterprise data warehouse.
Ultimately , the data architecture needs to be structured to accommodate business needs in the most efficient manner and main factors to consider include entities forming part of these transactions ,hierarchies important to the business and types of transactions the organization engages in.
In the next blog , I would start talking about the layers in detail and lets try and understand how this layer concept also helps in building a good BI governance model and maximize the value an organization derives from its BI Initiatives.
