Reply by: Backend_Developer_5yr
Also consider your team's expertise. If everyone knows SQL and PostgreSQL, dont switch to MongoDB just because its trendy. PostgreSQL can handle millions of IoT records if properly indexed. With JSONB columns you get flexibility of NoSQL within PostgreSQL. Start with what you know, optimize later if you hit performance issues. Premature optimization is root of all evil as they say.
Reply by: IoTplatform_CTO
We use combination - time-series database (InfluxDB) for sensor readings, PostgreSQL for metadata, user accounts, device registry. This hybrid approach works well. Sensor data is append-only time-series which InfluxDB handles brilliantly. Relational data like users, devices needs SQL. Dont force everything into one database type, use right tool for each data type.
Reply by: DatabaseArchitect_Senior
For IoT time-series data, consider time-series databases like InfluxDB or TimescaleDB (which is built on PostgreSQL). These are optimized for time-series workloads. Regular MongoDB or PostgreSQL will work but not optimal. NoSQL is not always better - depends on your data structure and query patterns. If you need complex queries with joins, SQL is better. If you need flexible schema and horizontal scaling, NoSQL is better.
User: NoSQL_vs_SQL_confused
Subject: When to use NoSQL database vs traditional SQL?
Planning database architecture for new IoT application that will collect sensor data. Should I use NoSQL like MongoDB or stick with PostgreSQL? Everyone says NoSQL is better for IoT but not clear why. What are deciding factors?
