Now we know how to set up and run a Cassandra cluster, so let's take another step and discuss the architecture of Cassandra. Cassandra is a highly available, distributed NoSQL database. Cassandra clusters can be linearly scalable depending on our load requirements. Cassandra has no single point of failure, as it allows us to control the trade-off between consistency and latency based on our requirements. In a Cassandra cluster, there is no master or slave—node all nodes are equal. This way enables Cassandra to read from any node and write to any node unlike some traditional relational databases. Cassandra partitions data among all nodes in its cluster; this partitioned data can be looked up using a key called a partition key. Cassandra replicates data among its cluster nodes to become highly available. We can configure data replication based on our specific requirements. In this chapter, we'll discuss more about these features of Cassandra.