Overview of Redis

Redis (Remote Dictionary Server) is an open-source, in-memory data structure store used as a database, cache, and message broker. In high-concurrency environments, traditional relational databases often suffer from disk I/O bottlenecks and scaling limitations. Redis addresses these by keeping data in RAM and utilizing a highly efficient single-threaded event loop architecture.

Core Data Types and Use Cases

Redis supports five primary data structures. Choosing the right one is critical for optimizing performance and memory usage.

1. Strings: The Foundation

Strings are the most basic Redis type. They can store text, serialized objects, or even binary data up to 512MB.

Key Applications

• Atomic Counters: Use INCR or DECR to manage global IDs or view counts across distributed systems without race conditions.
• Temporary Tokens: Use SETEX to store verification codes or session tokens that automatically expire.

# Basic string operations
SET session:user:99 "active"
GET session:user:99
EXPIRE session:user:99 3600  # Expires in 1 hour

# Atomic increment for a page view counter
INCR page:home:views
INCRBY page:home:views 5

2. Hashes: Object Mapping

Hash are ideal for representing objects because they allow you to access individual fields without fetching the entire object.

Application: Shopping Carts

By using a user ID as the key and product IDs as fields, you can manage cart quantities efficiently.

# Adding items to a cart
HSET cart:user:501 item:202 2
HSET cart:user:501 item:305 1

# Incrementing quantity
HINCRBY cart:user:501 item:202 1

# Retrieving all items in the cart
HGETALL cart:user:501

3. Lists: Sequential Data

Redis Lists are implemented via doubly linked lists, making insertions at the ends very fast (O(1)).

Application: Activity Feeds

Lists are perfect for "latest news" or "recent comments" where the order of arrival matters.

# Push new notification to the head
LPUSH notify:user:77 "Your order has shipped"
LPUSH notify:user:77 "Payment received"

# Fetch the 10 most recent notifications
LRANGE notify:user:77 0 9

# Blocking pop for simple message queues
BRPOP task:queue 30

4. Sets: Unique Collections

Sets are used when you need to store unique items and perform set-theory operations like finding common elements.

Application: Social Graphs and UV Tracking

# Adding followers
SADD following:user:A user:B user:C user:D
SADD following:user:X user:B user:D user:Z

# Find mutual friends (Intersection)
SINTER following:user:A following:user:X

# Track unique visitors per day
SADD uv:2023-10-27 "ip_address_1"
SCARD uv:2023-10-27

5. Sorted Sets (ZSets): Ranking and Priority

Sorted Sets maintain uniqueness like Sets but assign a "score" to each element, which Redis uses to keep the set ordered.

Application: Real-time Leaderboards

# Update player scores
ZADD leaderboard:gaming 5000 "PlayerAlpha"
ZADD leaderboard:gaming 7500 "PlayerBeta"
ZADD leaderboard:gaming 6200 "PlayerGamma"

# Get top 3 players
ZREVRANGE leaderboard:gaming 0 2 WITHSCORES

# Increment a score
ZINCRBY leaderboard:gaming 100 "PlayerAlpha"

Advanced Implemantation: Rate Limiting

A common scenario is limiting API calls per minute for a specific user. This can be achieved by combining strings with expiration logic.

# Strategy: Allow 5 requests per minute
# Key: rate_limit:user_id
# Logic: If key doesn't exist, SET with EXPIRE. If exists, INCR and check value.

# Example implementation logic:
# 1. Check if key exists
# 2. If not: SET rate_limit:user:101 1 EX 60
# 3. If yes: INCR rate_limit:user:101
# 4. If result > 5: Reject request

Key Management and Database Operations

Redis provides administrative commands to manage the lifecycle of keys and organizational structure.

• TTL (Time To Live): Use TTL key to check how many seconds remain before a key expires.
• Key Patterns: KEYS user:* finds all keys starting with "user:", though SCAN is preferred in production to avoid blocking.
• Multiple Databases: Redis instances provide 16 logical databases (0-15). Use SELECT 1 to switch.
• Persistence: While in-memory, Redis can persist data via RDB (snapshots) or AOF (append-only files) to prevent data loss after a restart.