Apache Cassandra

Apache Cassandra
Apache Software Foundation · Apache License 2.0 · source
Name	Apache Cassandra
Developer	Apache Software Foundation
Released	July 2008
Programming language	Java
Operating system	Cross-platform
Genre	NoSQL, Distributed database
License	Apache License 2.0

Apache Cassandra. It is a free and open-source, distributed, wide-column store NoSQL database management system designed to handle large amounts of data across many commodity servers, providing high availability with no single point of failure. Initially developed at Facebook to power its Inbox Search feature, it was open-sourced in 2008 and became a top-level project of the Apache Software Foundation in 2010. Cassandra offers robust support for clusters spanning multiple datacenters, with asynchronous masterless replication allowing low latency operations for all clients.

Overview

Apache Cassandra is a partitioned row store database, where rows are organized into tables with a required primary key. The system provides a Java-based Thrift and a CQL (Cassandra Query Language) interface, with the latter becoming the primary and recommended way to interact with the database. Its design is inspired by both Amazon's Dynamo distributed storage system and the data model of Google Bigtable, combining Dynamo's distributed systems techniques with Bigtable's column-family data model. This hybrid approach allows it to achieve high write throughput and scalability across many nodes, making it a popular choice for applications requiring massive scalability and fault tolerance.

Architecture

The architecture of Cassandra is a ring design, where each node in a cluster is identical; there is no concept of a master node, which eliminates any single point of failure. Data is distributed across the cluster using a variant of consistent hashing for partitioning and is replicated to multiple nodes for fault tolerance. Key components include a Gossip protocol for peer-to-peer communication, a Partitioner for data distribution, and replication strategies like SimpleStrategy and NetworkTopologyStrategy. For data consistency, it implements a tunable consistency model offering options like ONE, QUORUM, and ALL, balancing availability and consistency as defined by the CAP theorem.

Data model

Cassandra's data model is a schema-optional, wide column store organized around the concept of column families (tables). Each row is identified by a primary key, which can be simple or composite, and rows within a partition are stored in the order of their clustering columns. Unlike a traditional RDBMS, it does not support joins or foreign keys, encouraging denormalized data designs optimized for specific queries. The model is flexible, allowing columns to be added dynamically, and supports complex data types like collections, user-defined types, and Tuples, providing significant modeling power for diverse application needs.

Query language (CQL)

The primary interface for interacting with Cassandra is CQL, a SQL-like language that provides a familiar syntax for users of traditional SQL databases. While CQL resembles SQL, it is specifically designed for Cassandra's distributed architecture and data model, omitting operations like JOINs and supporting specific clauses like `ALLOW FILTERING`. Data definition and manipulation are performed using statements like `CREATE KEYSPACE`, `CREATE TABLE`, `INSERT`, `UPDATE`, and `SELECT`, with secondary indexes available via `CREATE INDEX`. Drivers for CQL are available in many programming languages, including Java, Python, Node.js, and Go, facilitating integration into diverse application stacks.

Use cases and adoption

Cassandra is widely adopted for use cases requiring high write throughput, scalability, and geographic distribution. It powers critical services at major technology companies like Netflix, Apple (for iCloud), Instagram, and Uber, often for messaging, recommendation engines, IoT data, and time-series data. Its ability to handle massive datasets with low latency makes it suitable for applications in telecom, finance (for fraud detection), and retail (for shopping carts). The project's ecosystem includes tools like DataStax, a commercial vendor offering enterprise support and additional tooling, and integrations with big data frameworks like Apache Spark and Apache Kafka.

History and development

Cassandra was created at Facebook by Avinash Lakshman and Prashant Malik to address the scaling challenges of the Inbox Search feature. It was released as an open-source project on Google Code in July 2008 and entered the Apache Incubator in March 2009. In February 2010, it graduated to become a top-level project of the Apache Software Foundation. Significant milestones in its development include the introduction of CQL to replace the original Thrift API, the adoption of the Paxos protocol for lightweight transactions, and continuous improvements in performance and manageability. The project is developed by a global community of contributors and is governed by the open collaboration principles of the Apache Software Foundation.

Category:Apache Software Foundation projects Category:NoSQL Category:Distributed data stores Category:Free database management systems