Caching Strategies

TL;DR

Caching = Store frequently accessed data in fast storage (RAM) to avoid slow database queries. Cache is the #1 way to improve performance - can make system 10-100x faster.

Core Concepts

Cache Hit vs Miss

• Cache hit: Data in cache (fast, ~1ms)
• Cache miss: Query database (~100ms), then store in cache
• Performance gain: 100x faster (1ms vs 100ms)

Caching Strategies

Cache-Aside (Lazy Loading) - Most Common

def get_user(user_id):
    # 1. Try cache first
    cached_user = cache.get(f"user:{user_id}")
    if cached_user:
        return cached_user  # Cache hit
    
    # 2. Cache miss - query database
    user = db.query("SELECT * FROM users WHERE id = ?", user_id)
    
    # 3. Store in cache (TTL = 5 minutes)
    cache.set(f"user:{user_id}", user, ttl=300)
    return user

Write-Through Cache

Write to cache and database simultaneously.

• ✅ Cache always consistent with DB
• ❌ Slower writes (two operations)

Write-Behind (Write-Back)

Write to cache immediately, async write to database.

• ✅ Fast writes
• ❌ Risk of data loss if cache crashes

Cache Eviction Policies

Policy	Evict	Use Case
LRU	Least recently used	General purpose (most common)
LFU	Least frequently used	Hot data stays (trending videos)
TTL	Expired items	Guaranteed freshness

Cache Stampede Solutions

# Locking solution
def get_cached(key):
    data = cache.get(key)
    if data:
        return data
    
    if cache.set_nx(f"lock:{key}", 1, ttl=10):  # Acquire lock
        data = db.query(key)
        cache.set(key, data, ttl=300)
        cache.delete(f"lock:{key}")
        return data
    else:
        time.sleep(0.1)
        return get_cached(key)  # Retry

TTL Guidelines

Data Type	TTL
Immutable	No TTL (forever)
Static content	1 year
User profiles	1 hour
Product prices	5 minutes
Real-time data	10 seconds or don't cache

Quick Reference

Strategies: Cache-aside (lazy), Write-through (sync), Write-behind (async), TTL (expiry)

Eviction: LRU (default), LFU (hot data), TTL (freshness)

Layers: Browser → CDN → App (in-memory) → Redis → Database

---