Mastering Redis: Core and Advanced Data Structures, Architecture, and Operations

Redis Data Structures

Basic structures:

String

Hash

List

Set

Sorted Set (Zset)

Advanced structures:

HyperLogLog – cardinality estimation

Geo – geolocation indexing

Pub/Sub – publish/subscribe messaging

Redis Architecture Overview

Implemented as a single‑process, single‑threaded key‑value store

Fully in‑memory, providing extremely low latency

Relies on I/O multiplexing (epoll/kqueue) to handle many client connections efficiently

Simple data structures lead to straightforward operations and high performance

Advanced Operations

Finding a large number of keys with a known prefix (e.g., 100 000 out of 100 million):

KEYS + regex – works but blocks the server and should be avoided in production

SCAN – incremental, non‑blocking cursor‑based scan suitable for large datasets

Setting the same expiration time for many keys:

Do not concentrate expirations at a single timestamp; add a random offset to each key to spread the load and prevent latency spikes.

Persistence mechanisms:

AOF (Append‑Only File) – incremental log of write commands; durability depends on the sync configuration (e.g., every write, every second, or never). Files are larger than RDB snapshots.

RDB (Redis Database) – point‑in‑time snapshot created by a forked child process; minimal impact on the main thread but may lose recent data if a crash occurs before the next snapshot.