
Database Scaling and Selection

#176 - Acing the System Design Interview - Zhiyong Tan
Topic 9 of 13
Published May 27, 2024 · Clip 28:41–33:13 (4:32)
Zhiyong explains database choice is about trade-offs against non-functional requirements. Start with SQL for simplicity and speed to market, then as load grows trade ACID properties for NoSQL options like Redis, Cassandra, or MongoDB depending on read/write patterns and latency needs. He discusses introducing event streaming like Kafka feeding HDFS to reduce cost, and bringing up sampling and approximation as scalability reducers.