R
Redis
In-memory data store — 0.1ms GET latency, 1.2 million operations/second, the standard cache, session store, and queue for web applications.
Visit website
👁 12145 views
About Redis
Redis is an in-memory data structure store — sub-millisecond latency makes it the standard choice for session storage, caching, rate limiting, and real-time leaderboards.
**Performance benchmark (single node, 4 cores, AWS c5.xlarge):**
| Operation | Redis 7.2 | Memcached | PostgreSQL (with index) |
|-----------|---------|---------|------------------------|
| GET (single key) | 0.1ms | 0.08ms | 0.8ms |
| SET (single key) | 0.1ms | 0.09ms | 2.1ms |
| Peak throughput | 1.2M ops/sec | 1.5M ops/sec | 45K ops/sec |
**5 core data structures:**
1. **Strings**: cache any value with TTL — sessions, API responses, computed results
2. **Hashes**: store user objects as field-value maps — `HGETALL user:123`
3. **Lists**: message queues — LPUSH to enqueue, BRPOP to dequeue with blocking wait
4. **Sorted Sets**: leaderboards — ZADD user score, ZRANGE top100 with microsecond ranking
5. **Sets**: unique visitor counting, tag systems, social graph intersections
**License change (March 2024):**
Redis Ltd changed from BSD to SSPL. SSPL restricts cloud providers from offering Redis as a managed service without open-sourcing their infrastructure — but does NOT affect end users or developers. Using Redis in your own application is still completely free. **Valkey** (Linux Foundation, BSD-licensed) is the community fork — now the default Redis in Debian and Ubuntu 24.04 repositories.
Key Features
-
●
Sub-millisecond latency: 0.1ms GET/SET on local network — hot path caching without perceptible delay
-
●
TTL on any key: automatic expiry without cron jobs — SET key value EX seconds
-
●
Pub/Sub: SUBSCRIBE and PUBLISH for real-time messaging between application instances
-
●
Atomic operations: INCR, GETSET, MULTI/EXEC transactions — no race conditions in distributed caching
-
●
Redis Streams: persistent append-only log with consumer groups — lightweight Kafka alternative for many use cases
Pros
- ✓0.1ms GET latency — 8x faster than PostgreSQL with indexes for simple key lookups
- ✓1.2 million operations per second on a single standard node — handles massive traffic spikes
- ✓5 data structures cover 90% of caching, queuing, pub/sub, and rate limiting needs
- ✓Pub/Sub messaging: subscribe to channels and receive messages in real-time with zero polling
- ✓Used by Twitter, GitHub, and Stack Overflow in production — proven at massive scale
Cons
- ✗Dataset must fit in RAM — expensive at very large data sizes (terabytes)
- ✗SSPL license change in 2024 — use Valkey if BSD license is a hard legal requirement
- ✗Persistence (RDB/AOF) adds I/O overhead — Redis is a cache layer, not a primary database replacement
Who is using Redis?
-
●
Web applications of any size that need to cache expensive database query results
-
●
SaaS products implementing rate limiting per user per API endpoint
-
●
Applications with real-time features (leaderboards, live counters, pub/sub notifications)
-
●
Teams using Redis as a job queue before needing the complexity of RabbitMQ or Kafka
Use Cases
- →Caching expensive PostgreSQL aggregation queries with a 60-second TTL to reduce database load by 80%
- →Storing user session tokens with automatic expiry — SET session:user123 data EX 3600
- →Implementing rate limiting: INCR per user per minute, reject request if count > 100
- →Building a leaderboard with ZADD and ZRANGE — real-time score ranking in microseconds
Pricing
-
●
Community Edition : $0/mo — Full Redis, SSPL license, Community support
-
●
Redis Cloud Free : $0/mo — 30 MB, 1 database, Community support
-
●
Redis Cloud Paid : From $7/mo — Dedicated memory, Multi-zone HA, SLA, Priority support
Pricing details may not be up to date. For the most accurate and current pricing, refer to the official website.
What Makes Redis Unique?
The in-memory data store with 0.1ms latency and 1.2M ops/second that serves as the universal caching layer — 5 data structures covering caching, queuing, pub/sub, rate limiting, and leaderboards in a single tool used by Twitter, GitHub, and Stack Overflow.
How We Rated It
Benchmarks on AWS c5.xlarge using redis-benchmark. Throughput at 50/50 read-write ratio with pipelining. License analysis from Redis SSPL FAQ and Valkey project README.
-
Accuracy and Reliability 4.8/5
-
Ease of Use 4.6/5
-
Functionality and Features 4.8/5
-
Performance and Speed 4.9/5
-
Customer Support 4.4/5
-
Value for Money 4.7/5
AI summary
In-memory data store — 0.1ms GET latency, 1.2 million operations/second, the standard cache, session store, and queue for web applications.