It is a common practice to extract data from a database so that analytics could be performed on it. The word “extract” implies that there is a movement of data from the database to a data set which resides on a disk. This practice worked quite well till Big Data came along.
Assuming a database table has ten terabytes of data, it is not possible to extract this data considering disk space and time taken for extract ion to complete. It would be best to prevent data movement in the first place and perform analytics on the database directly, hence the phrase in-database analytics. Modern data warehouse appliances such as Netezza, Teradata and Greenplum provide in-database analytics functionality.
So what are the advantages of in-database analytics? Obviously, insights would be derived faster and this in turn would assist C-Suite executives make business decisions hopefully before their competitors do. In a retail industry where today’s “hot” item can be tomorrow’s “has-been”, faster insights are crucial.
The second advantage is total cost of ownership (TCO). Due to elimination of data movement, servers don’t have to be upgraded to account for extra space and processors. Since data is analyzed typically by advanced SQL, license fees for expensive analytics software can be avoided.