Introduction to Redis: What It Is and Why It’s Fast

Source: Dev.to

What Is Redis?

Redis (Remote DIctionary Server) is an open‑source, in‑memory data‑structure store and one of the most popular NoSQL databases. Created by Salvatore Sanfilippo in 2009, it is often called a “data‑structure server” because it supports far more than simple key‑value storage.

• In‑memory database – all data resides in RAM, giving extremely fast read/write operations.
• Persistence options – despite being memory‑based, Redis can persist data to disk.
• Modern servers with hundreds of gigabytes of RAM make Redis practical for many real‑world workloads.

Core Data Types

Redis supports a rich set of optimized data structures:

• Strings – text, binary data, or numbers
• Lists – ordered collections of strings
• Sets – unordered collections of unique strings
• Hashes – field‑value maps (similar to objects)
• Sorted Sets – ordered sets with scores
• Bitmaps – space‑efficient bit‑level operations
• HyperLogLogs – probabilistic cardinality estimation
• Streams – append‑only logs for messaging and event sourcing

Persistence

Redis can store data durably using two mechanisms (independently or together):

Replication & High Availability

Redis supports master‑replica replication, allowing data to be copied to multiple replicas. This provides:

• Read scalability
• High availability
• Data redundancy

Transactions

Redis transactions execute a group of commands atomically, ensuring all commands are processed as a single unit.

Publish/Subscribe

Redis includes a Pub/Sub messaging system that enables real‑time message broadcasting to multiple subscribers.

Server‑Side Lua Scripting

Lua scripts can be run on the server, allowing multiple operations to be executed atomically and efficiently in a single script.

Performance Characteristics

Redis can handle 100,000+ operations per second on a single core. Key reasons for this speed include:

• No disk I/O bottlenecks – data lives in RAM
• Direct memory access via pointers, avoiding costly serialization
• Single‑threaded, non‑blocking I/O model
  • No context‑switching overhead
  • No race conditions or locking complexity
  • Predictable, stable latency

To utilize multiple CPU cores, run multiple Redis instances on the same machine.

Optimized Internal Data Structures

• Hash tables with incremental rehashing
• ZipLists for memory‑efficient small collections
• IntSets for integer‑only sets
• Skip Lists for sorted sets (O(log N) performance)

Implementation Details

• Written in ANSI C → manual memory management, minimal abstraction, direct system‑level optimizations
• Uses high‑performance system calls (epoll, kqueue, select) depending on the platform
• RESP (Redis Serialization Protocol) – simple and fast to parse
• Pipelining – send multiple commands without waiting for responses

Asynchronous Background Tasks

• Background RDB saves
• AOF rewriting
• Forked child processes

These keep the main event loop responsive.

Common Use Cases

Installation

Ubuntu / Debian

sudo apt-get update
sudo apt-get install redis-server

macOS (Homebrew)

brew install redis

Docker

docker run --name redis -d -p 6379:6379 redis

Basic Commands (Redis CLI)

SET user:1001 "John Doe"
GET user:1001

SET session:abc123 "data" EX 10   # expires after 10 seconds
INCR page_views
LPUSH tasks "send_email"
SADD tags "redis" "database" "cache"

Advantages

• Much faster reads and writes compared to disk‑based databases
• No JOINs or complex queries – simple key‑based access
• Schema‑less and flexible data structures
• Built‑in persistence, Pub/Sub, transactions, and Lua scripting

Limitations

• Memory cost – RAM is more expensive than disk storage
• Dataset size – data must fit into available memory
• Single‑threaded model – CPU‑heavy operations can block other requests
• Persistence trade‑offs – RDB vs. AOF require careful configuration

Summary

Redis achieves its exceptional speed through a well‑designed architecture:

1. In‑memory storage eliminates disk I/O.
2. Single‑threaded event loop avoids concurrency overhead.
3. Highly optimized C‑based data structures ensure efficiency.

While Redis is not a replacement for traditional relational databases, it excels as a cache, session store, real‑time analytics engine, message broker, and fast data processor. Its simplicity, performance, and flexibility have made it a core component of modern system architectures used by companies such as Twitter, GitHub, Stack Overflow, and Snapchat.

Whether you’re building a small application or scaling a high‑traffic platform, Redis provides sub‑millisecond response times and unmatched performance for fast‑changing data.