cassandra

Apache Cassandra

Cassandra is a wide-column store database designed for scalability and high availability without compromising performance. Linear scalability and proven fault-tolerance on commodity hardware or cloud infrastructure make it the perfect platform for mission-critical data.

When to Use

• High-Velocity Write Ingestion: Logging millions of writes per second across IoT sensor streams, messaging systems, and time-series metrics.
• Zero-Downtime Multi-Datacenter Replication: Geographically distributed systems requiring peer-to-peer active-active masterless replication.
• Linear Horizontal Scalability: Scaling throughput predictably by adding commodity hardware nodes without cluster downtime.
• Predictable Query Latency: Serving high-concurrency key-value and partition-key lookups under 5 milliseconds.

Quick Start

CREATE TABLE users (
  user_id UUID PRIMARY KEY,
  name text,
  email text
);

INSERT INTO users (user_id, name) VALUES (uuid(), 'Alice');

Core Concepts

Masterless Ring Architecture & Consistent Hashing

Every node in the cluster is identical; partition tokens dictate which nodes own primary and replica data:

[ Node 1 (Token: 0) ] ────→ [ Node 2 (Token: 33) ] ────→ [ Node 3 (Token: 66) ]
          ▲                                                           │
          └───────────────────────────────────────────────────────────┘

Partition Keys vs Clustering Columns (CQL)

The partition key determines physical node placement; clustering columns determine on-disk sorting within the partition:

-- Sensor Telemetry Schema
CREATE KEYSPACE telemetry_data
WITH replication = {'class': 'NetworkTopologyStrategy', 'us-east': 3, 'eu-west': 3};

CREATE TABLE telemetry_data.sensor_readings (
    sensor_id UUID,
    recorded_date DATE,
    recorded_at TIMESTAMP,
    temperature DOUBLE,
    humidity DOUBLE,
    PRIMARY KEY ((sensor_id, recorded_date), recorded_at)
) WITH CLUSTERING ORDER BY (recorded_at DESC);