Data warehouses are complex, clumsy, ugly animals that have thankfully become redundant. What is the use of a centralized warehouse of data when one can easily browse data across the organisation smoothly using various data accessing techniques? Bandwidth is cheap, processing is cheap and storage does not cost anything. So its easier to keep data and information in a distributed fashion rather than collate in the form of a warehouse. Big data is a fashionable phrase in this context. But big data still does not get rid of the ugliness and clumsiness of a warehouses it just adds a framework to the whole madness. What the world needs is an agile, fast and light weight solution to integrate data present in silos and to provide a first layer of abstraction in analytics framework.
Follow
It seems, I am missing something here.
Whole trouble in creation of a structured warehouse pays out while using data for analytic or reporting. It saves a lot of time of data cleaning and also saves from potential errors because of data standardization. Also I believe it is more efficient and cost effective to maintain data at one place.
Hey manish
That’s the standard argument for using a warehouse. But these days there are powerful tools that can be used to about achieve standardization, cleansing and easy access without actually building a warehouse.
See this for more:
http://online.wsj.com/community/groups/innovation-technology-innovation-/topics/data-warehouse-dead
Hey Gopiks,
Data accessing is one part but what about the standardization of data which anyone need to be done at the source at each separate small warehouse? Isn’t it be costly to do it separately everywhere?
Hey
That can also be done using simple systems to clean the data and present it in a standardized fashion. If such such a system does not work there is definitely a gasp and one needs to be built.
Thank You Gopiks
Your blog is very insightful, please keep posting.