Big Data and RDBS (Relational DB). Do They Fit?

[WWGD]

Hi All,
I am having trouble seeing how Relational Databases (RDBS) can be used in the world of big data.
The inflow of data seems to be way too fast for the database to reflect what is going on at a given
moment. I understand this issue is supposed to be addressed by data warehouses. Is this correct?
If so, what kind of databases are used by data warehouses?

 

[Borg]

Hi All,
I am having trouble seeing how Relational Databases (RDBS) can be used in the world of big data.
The inflow of data seems to be way too fast for the database to reflect what is going on at a given
moment. I understand this issue is supposed to be addressed by data warehouses. Is this correct?
If so, what kind of databases are used by data warehouses?

NoSQL databases like MongoDB, Hadoop, etc.

 

[jedishrfu]

Data warehouses use the same kinds of databases as "big data". Usually, information from the corporate database is extracted and organized into smaller data warehouses useful for doing analysis and forecasting of trends.

As an example, the corporate database (amazon) might contain information on customers, their purchases and wishlists, what they looked at... whereas the datawarehouse might not care about the customers buying behavior but only the products that were purchased when, where and how but not who and from that forecast what stores need what inventory.

 

[WWGD]

Data warehouses use the same kinds of databases as "big data". Usually, information from the corporate database is extracted and organized into smaller data warehouses useful for doing analysis and forecasting of trends.

As an example, the corporate database (amazon) might contain information on customers, their purchases and wishlists, what they looked at... whereas the datawarehouse might not care about the customers buying behavior but only the products that were purchased when, where and how but not who and from that forecast what stores need what inventory.

Thanks. So it seems like a division of labor where corporate finds patterns and the warehouses deal with day to day?

 

[QuantumQuest]

RDBMS have themselves become more fast, reliable and secure - you can take examples of big vendors like Oracle, but there is no silver bullet in this case too, so as pointed out by jedishrfu some well - balanced combinations of both RDBS and NoSQL are used in practice, in order to have the most balanced workload. The exact "how" varies according to company's needs.