1. Cassandra is a distributed, decentralized, and fault-tolerant NoSQL database that distributes data across nodes in a cluster to provide high availability and no single points of failure. 2. It is optimized for high write throughput and sacrifices consistency in favor of availability. Data is distributed across nodes and replicated for fault tolerance, with tunable consistency levels. 3. Cassandra is best for write-heavy workloads with large volumes of log or time series data where data is accessed by key and queries are satisfied by a single partition. It is not a general purpose database and lacks features like transactions and joins.