System Design

Designing systems that handle growing load — horizontal vs vertical scaling, load distribution, stateless services.

Store frequently-accessed data in fast memory. Reduces DB load, cuts latency. Cache invalidation is the hard part.

SQL vs NoSQL, replication, sharding, indexing. Pick the right DB for the workload — not one DB for everything.

In a distributed system, you can only have 2 of 3: Consistency, Availability, Partition Tolerance. Partition tolerance is mandatory.

Decouple producers and consumers. Absorb traffic spikes, enable async processing, guarantee delivery. Kafka, RabbitMQ, SQS.

REST, GraphQL, gRPC. Rate limiting, versioning, authentication. Design APIs that are predictable, consistent, and evolvable.

Full-text search, ranking, indexing. Elasticsearch, inverted indexes, TF-IDF. How Google-scale search actually works.

Content delivery, DNS, TCP/IP, HTTP/2. How data moves from server to user in milliseconds.