Cassandra

The topic of NoSQL isn't complete without discussing columnar storage. In Cassandra, the data is stored and read in columns instead of rows. Each column is stored separately with the same row offset related to the table. We will study offset in the Data Modeling subsection:

Figure 5.50: Keyspace and column family internal structure

Cassandra is a mixture of structured data, consistency, high scalability, and no single point of failure, and also has a powerful column family design. This particular database was developed by Facebook in 2008 and is horizontally scalable. Imagine you run a multinational company. You have thousands of employees, but you want the information of all the employees at the London location. In such a scenario, a traditional relational database will scan each row, parse it, and compare the u_location column value against the query. This is extremely time-consuming.

On the other hand, the columnar store only requires columns...