Strings, hashes, sorted sets, streams, Lua scripts, Redlock.

Redis — Data Structures, Expiry, and Distributed Patterns

Redis is an in-memory data structure store. It excels at caching, session storage, rate limiting, pub/sub, and leaderboards. Key data structures: Strings (SET/GET with TTL), Hashes (HSET/HGETALL for objects), Lists (LPUSH/RPOP for queues), Sorted Sets (ZADD/ZRANGE for leaderboards), Streams (XADD/XREAD for event logs). All commands are atomic. Lua scripts run atomically. Redlock is the distributed mutual-exclusion algorithm across multiple Redis nodes.

Interactive Pipeline Flow

UnderstandImplementOptimise

🔌 Connect — connection pool

Use a connection pool (StackExchange.Redis in .NET, ioredis in Node.js). Creating a new connection per request is expensive — reuse the multiplexer.

Architecture Diagrams

Sliding window rate limiter using Sorted Sets.

Loading diagram...

Key Design Decisions

🚫Never use KEYS * in production

KEYS * scans all keys — O(N) and blocks the single-threaded Redis during scan. Use SCAN 0 MATCH pattern COUNT 100 with cursor iteration instead.

📡Use Streams for durable messaging

PUBLISH/SUBSCRIBE loses messages if no subscriber is connected. Use Streams (XADD/XREADGROUP) for durable, acknowledged, consumer-group-based messaging.

⏱️Always set TTLs on cache keys

Cache keys without TTL grow forever until eviction policy kicks in. Set EX at write time. Monitor memory with INFO memory and keyspace.

Discussion