Course contentsShow
← System Design
Lesson 123 of 1,9194. Caching StrategiesFree lesson

Distributed Cache Layer (Redis/Memcached)

Centralized caching across multiple application servers using external cache stores like Redis or Memcached.

Distributed Cache Layer (Redis/Memcached)

What you'll learn: How to use external cache stores like Redis or Memcached to share cached data across multiple application servers.

The Problem with Server-Specific Caches

Remember application-level in-memory caching? Each server stored its own cached data in RAM. This creates a problem when you scale horizontally: if you have 10 application servers, the same piece of data might be cached 10 times (wasting memory), or worse—a cache hit on Server 1 is a cache miss on Server 2, making your caching inefficient.

Centralized Caching Solution

A distributed cache layer moves the cache outside your application servers into a specialized external service. Think of it like replacing individual desk drawers (per-server caches) with a shared filing cabinet that everyone can access.

Redis and Memcached are the two most popular distributed cache stores:

  • They run as separate services on dedicated servers
  • All your application servers connect to them over the network
  • They store key-value pairs entirely in memory for ultra-fast access
  • They handle eviction policies, expiration, and memory management automatically

How It Works

When Server A caches user profile data in Redis, Server B can instantly retrieve that same cached data—no duplicate work, no wasted memory. The cache becomes a shared resource across your entire fleet of application servers.

The tradeoff? Network latency. Accessing Redis over the network (~1ms) is slower than in-process memory (~0.001ms), but still vastly faster than hitting a database (~10-100ms).

Key Takeaway: Distributed cache layers like Redis and Memcached centralize caching across multiple servers, eliminating duplication and improving cache hit rates at the cost of minimal network overhead.