Course contentsShow
← System Design
Lesson 143 of 1,9194. Caching StrategiesFree lesson

Multi-Tier Caching Pattern

Layering caches from local in-memory to distributed remote caches for different data access speeds and scopes.

Multi-Tier Caching Pattern

What you'll learn: How to layer multiple caches—from fast local memory to slower distributed caches—to balance speed, scope, and cost.

The Core Idea

Multi-tier caching organizes your caches into layers, each with different characteristics:

  • Tier 1 (Local/In-Memory): Fastest but smallest. Lives inside your application process (like a HashMap or LRU cache in memory). Access time: microseconds.
  • Tier 2 (Distributed/Remote): Slower but shared across services. A separate cache cluster like Redis or Memcached. Access time: milliseconds.
  • Tier 3 (Database/Origin): Slowest but authoritative source. Your database or external API.

When a request arrives, you check Tier 1 first. If it's a miss, check Tier 2. If that's also a miss, fetch from Tier 3 (the database), then populate both Tier 2 and Tier 1 on the way back.

Why Layer Caches?

Think of it like storing groceries: you keep frequently-used items (salt, pepper) on the countertop (Tier 1), less-frequent items in the pantry (Tier 2), and rarely-used bulk goods in the garage (Tier 3). You always check the countertop first because it's closest.

Benefits:

  • Speed: Hot data served from local memory (no network calls)
  • Reduced load: Distributed cache shields the database from repeated queries
  • Scope: Local cache is instance-specific; distributed cache is shared across all instances

Trade-offs: More complexity, potential for stale data across tiers, and cache coherence challenges.

Key Takeaway: Multi-tier caching creates a hierarchy of speed and scope—use fast local caches for hot data and slower distributed caches for shared data, checking each tier in order before hitting the database.