DEV Community

Cover image for Caching — The Secret Weapon Behind Fast, Scalable Systems — Architecture Series: Part 4
MUHAMMAD USMAN AWAN
MUHAMMAD USMAN AWAN

Posted on

Caching — The Secret Weapon Behind Fast, Scalable Systems — Architecture Series: Part 4

#webdev #programming #learning #architecture

🚀 Caching — The Secret Weapon Behind Fast, Scalable Systems

🧐 What is Caching?

Caching means storing frequently accessed or expensive-to-compute data in a faster medium (in-memory, browser, CDN, Redis) so future requests are served instantly.

How it works:

  1. Check cache → ✔️ If data is found → Cache Hit → return instantly ❌ If not found → Cache Miss → fetch from DB/API → store → return

Caching = trading RAM for speed.

🤔 Where Data Is Cached? (Cache Layers)

1. Application Cache (In your code)

  • Memoization
  • Function-level caching
  • Example: caching expensive calculations

2. Server-Side / Distributed Cache

  • Redis, Memcached
  • For sharing cache across multiple backend instances

3. Database Cache

  • Query caching
  • Materialized views
  • Example: PostgreSQL caching recent SELECT queries

4. Browser Cache

  • Uses HTTP headers
  • localStorage, sessionStorage, IndexedDB

5. CDN Cache (Global Edge Servers)

  • Cloudflare, CloudFront
  • Caches images, CSS, JS, static HTML at edge

🤨 Why Do We Use Caching? (The Real Benefits)

⚡️ Performance Boost

Database query: ~100–300ms
Redis fetch: ~0.5–2ms
CDN fetch: ~20–50ms
Browser cache: ~0–1ms

🧠 Scalability

Reduce database load by 80–90%.

💰 Cost Savings

Lower DB reads, less compute, fewer servers.

🚨 Reliability & Fallback

If DB is slow or temporarily unavailable → cache can still serve data.

✨ Better User Experience

Faster pages → lower bounce rates → higher conversion.

😲 Real-World Examples

Twitter (X)

Trending topics are recalculated every few seconds, stored in Redis → millions of users see instantly.

Instagram

Profile + feed metadata cached → reduces reads on sharded DB.

YouTube

Video metadata cached on edge → instant loads worldwide.

😎 Caching Patterns (VERY Important for Interviews + Real Systems)

1. Cache-Aside (Lazy Loading)

Most common in backend systems (Redis + Node.js).

Flow:
Check cache → if miss → fetch DB → store → return.

2. Read-Through Cache

Application always queries cache.
Cache itself fetches from DB on miss.

3. Write-Through Cache

Write to cache → cache writes to DB.
(Data always fresh, but slower writes)

4. Write-Back (Write-Behind)

Write to cache → return immediately → cache writes to DB later.
(Fast writes, but riskier)

😉 Cache Invalidation (One of the hardest problems)

This decides whether cache stays fresh or stale.

Invalidation Techniques:

  1. TTL (Time-to-live) Auto-expire after X seconds.
  2. Manual Invalidate redis.del('user:123')
  3. Event-based On DB update → publish event → invalidate distributed caches.
  4. Versioning Add version numbers to keys: posts:v2:latest

🔥 Common Issues & Solutions

1. Stale Data

Fix using TTL or events.

2. Cache Stampede / Thundering Herd

Millions request same “cold key” = DB burst.

Fix with:

  • Locking (Redis SETNX)
  • Stale-while-revalidate
  • Request coalescing

3. Memory Pressure

Fix via eviction policies:

  • LRU (least recently used)
  • LFU
  • FIFO

4. Cold Start

Warm cache using popular keys on boot.

🧐 How to Use Redis Caching in Backend (Node.js / Express Example)

This is the simplest and most widely used pattern:

import express from "express";
import Redis from "ioredis";
import User from "./models/User.js";

const app = express();
const redis = new Redis();

app.get("/user/:id", async (req, res) => {
  const id = req.params.id;
  const key = `user:${id}`;

  // Check cache
  const cached = await redis.get(key);
  if (cached) {
    return res.json(JSON.parse(cached));
  }

  // Fetch DB
  const user = await User.findById(id);
  if (!user) return res.status(404).json({ message: "User not found" });

  // Save to cache
  await redis.set(key, JSON.stringify(user), "EX", 300); // 5 min TTL

  return res.json(user);
});

app.put("/user/:id", async (req, res) => {
  const id = req.params.id;

  // Update DB logic here...

  // Invalidate cache
  await redis.del(`user:${id}`);

  res.json({ message: "Updated + Cache invalidated" });
});

app.listen(3000);
Enter fullscreen mode Exit fullscreen mode

🧐 Frontend Caching (React / Next.js)

Use React Query / SWR for app-level caching.

Example with React Query:

const { data } = useQuery({
  queryKey: ["user", id],
  queryFn: () => axios.get(`/api/user/${id}`).then(res => res.data),
  staleTime: 5 * 60 * 1000, // 5 min fresh
});
Enter fullscreen mode Exit fullscreen mode

Browser-level caching via headers: 

Cache-Control: public, max-age=3600
ETag
Last-Modified
Enter fullscreen mode Exit fullscreen mode

Storage API options:

  • localStorage (persistent)
  • sessionStorage (session only)
  • IndexedDB (large data)
  • Service Workers (PWA offline caching)
  • Next.js built-in image caching

🤨 CDN Caching

Used for:

  • Static assets (JS, CSS, images)
  • API responses (if allowed)
  • Edge caching for SSR (Next.js + Cloudflare)

Cache-Control example for static assets: 

Cache-Control: public, max-age=31536000, immutable
Enter fullscreen mode Exit fullscreen mode

🔥 When Should You Use Caching?

Use caching when:

✔️ Same data requested repeatedly
✔️ DB/API becomes bottleneck
✔️ Read-heavy systems
✔️ Expensive computations
✔️ Static/semi-static data
✔️ Large traffic spikes

Avoid caching:

❌ Highly volatile data (stock prices, payments)
❌ Data requiring strong consistency
❌ User-sensitive content on shared caches

If you're following along with this Architecture Series and happened to miss the earlier parts, no worries — you can dive into them anytime. Each part builds your foundation step by step. Here are the previous topics we covered:

1️⃣ Pagination — Architecture Series: Part 1
2️⃣ Indexing — Architecture Series: Part 2
3️⃣ Virtualization — Architecture Series: Part 3

Top comments (0)