Cassandra

Cassandra is a distributed, wide-column NoSQL database designed to handle large amounts of data across many commodity servers, while providing high availability and no single point of failure. It was named after the Greek mythological figure Cassandra. The project originated at Facebook to support scalable data storage for several applications and was released as an Apache Software Foundation project in 2008, later becoming a widely used database in the NoSQL ecosystem.

Cassandra uses a column-family data model. Tables have a primary key consisting of a partition key and optional clustering columns. Data is stored on disk as SSTables and is written first to a durable commit log, then buffered in memory as a memtable before being flushed. Reads leverage bloom filters and indexed structures, and CQL provides a SQL-like query interface even though the underlying model differs from relational databases.

Architecture and replication: Cassandra relies on a peer-to-peer ring of nodes with no single master. Nodes communicate via a gossip protocol to keep cluster state in sync and coordinate data replication using a configurable replication factor and strategies such as SimpleStrategy or NetworkTopologyStrategy. Data is partitioned with a partitioner (commonly Murmur3), and writes and reads specify tunable consistency levels (for example ONE, QUORUM, or ALL).

Consistency, availability, and maintenance: Cassandra emphasizes availability and partition tolerance in practice and provides tunable consistency levels. Writes are durably logged and persisted across replicas; reads may be served from multiple replicas with read repair and hinted handoff. Tombstones mark deletions and, together with compaction, periodically merge on-disk structures to reclaim space.

Use and ecosystem: Cassandra is well suited to large-scale, write-heavy workloads such as real-time analytics, time-series data, and online applications. It has been adopted by many organizations for high availability needs and massive throughput, including public services and media platforms.