Skip to content

db-capacity.md

Latest commit

 

History

History
43 lines (33 loc) · 3.15 KB

db-capacity.md

File metadata and controls

43 lines (33 loc) · 3.15 KB
Raw

Here's a comparison of how much data different databases can handle, along with their typical use cases. Keep in mind, real-world performance depends heavily on hardware, configuration, indexing, and schema design, but databases often have practical vs theoretical limits.

🔢 Comparison Table: Database Data Handling Capacity

Database Data Handling Capacity (Theoretical / Practical) Notes / Use Cases
MySQL TBs (up to ~256TB per table with InnoDB) Web apps, CMS, eCommerce
PostgreSQL TBs (up to 32TB per table, unlimited rows) Data analytics, GIS, financial apps
MongoDB PBs (max 64TB per document store theoretically) NoSQL, unstructured data, JSON-like
Redis Depends on RAM (usually up to 1–2TB per instance) In-memory cache, fast key-value store
SQLite ~281 TB (single .db file) Embedded apps, mobile apps
MariaDB Similar to MySQL (InnoDB → up to 256TB/table) Same use cases as MySQL
Cassandra PBs+ (horizontal scaling, billions of rows) Distributed, IoT, time-series data
ClickHouse PBs (optimized for analytical queries) Real-time analytics, big data
TimescaleDB TBs+ (built on PostgreSQL) Time-series data, IoT, metrics
Elasticsearch PBs+ (depends on sharding/indexing) Full-text search, logging, analytics
Oracle DB TBs–PBs (Exadata optimized for large systems) Enterprise-grade, mission-critical
MS SQL Server TBs (524PB per database theoretically) Enterprise, BI, ERP systems
Neo4j (Graph DB) TBs (limit depends on JVM and disk) Graph analytics, social networks
DynamoDB (AWS) Unlimited (scales automatically) Serverless NoSQL, event-driven apps
BigQuery / Snowflake Virtually Unlimited Cloud data warehouses

💡 Notes

  • MySQL/PostgreSQL: Best for structured relational data, very mature, widely used.
  • MongoDB: Great for flexible schemas and massive scale with horizontal sharding.
  • Redis: Lightning fast, but all data must fit in memory unless using Redis-on-Flash or persistence.
  • SQLite: File-based, great for apps or small-scale data; not recommended for very large systems.
  • Cassandra: Ideal for massive write-heavy workloads and distributed environments.
  • ClickHouse / BigQuery / Snowflake: Tailored for analytics at petabyte scale, not for frequent updates.

🚀 Scaling Up

  • Vertical Scaling: Add more CPU/RAM/Storage to a single server (limited).
  • Horizontal Scaling: Add more nodes (Cassandra, MongoDB, ClickHouse, etc. support this better).