• .Net Core
• 1
• June-09-2025
• by Ihsan Ullah

What is Caching?

Caching stores a copy of data in a faster storage layer (like memory) so future requests can be served quicker without hitting the database or doing expensive computation again.

---

🧠 Common Caching Strategies for APIs

1. In-Memory Caching (e.g. Redis, Memcached)

• Best for: High-performance APIs with frequent reads.
• How it works: Data is stored in RAM (like a key-value store).
• Tools: Redis, Memcached

Example:

// Pseudo-code

if (cache.has("user:123")) {

   return cache.get("user:123");

}

const user = db.getUser(123);

cache.set("user:123", user);

return user;

2. HTTP Caching (Client-side + CDN)

• Best for: Public APIs and static content (like images, product listings).
• How it works: Uses HTTP headers like Cache-Control, ETag, Expires to instruct browsers/CDNs to cache.
• Tools: CDNs (Cloudflare, Fastly), browsers

Example headers:

Cache-Control: public, max-age=3600

ETag: "abc123"

3. Database Query Result Caching

• Best for: Repeated complex queries with infrequent changes.
• How it works: Cache the result of the DB query in memory (like Redis) and return the cached result if it exists.

---

4. Application-level Caching

• Store computed results (like filtered lists, analytics) in memory or files.
• Useful when the logic is expensive but doesn’t change often.

---

5. Persistent Caching

• Saves cached data to disk or a database so it survives app restarts.
• Good for infrequent data changes but high read demand.

---

6. Stale-While-Revalidate

• Serve stale data immediately, while fetching and updating the cache in the background.
• Keeps responses fast while data stays reasonably fresh.

Libraries: swr in React, custom middleware in Node/Express.

---

7. Per-User or Per-Session Caching

• Useful for personalized data.
• Cache is scoped to a user or token: e.g., cache.get("user:42:dashboard")

---

8. Reverse Proxy Caching (e.g. Varnish, Nginx)

• Acts as a cache layer between your app and the outside world.
• Caches whole HTTP responses.

---

🔐 Cache Invalidation (Hard Part)

Caching is fast, but invalidating (updating or removing) stale data is tricky:

• Time-based (TTL): Set expiration time.
• Event-based: Invalidate cache on database change or API call.
• Manual purge: Admins or developers trigger cache clear.

---

✅ Best Practices

• Use cache only where data is relatively stable.
• Set appropriate TTL (time to live).
• Watch out for race conditions or cache poisoning.
• Use tools like Redis with expiry, HTTP caching, and CDNs together.

---

🧪 Example Stack

For a MERN app:

• Use Redis for backend caching
• Use node-cache or lru-cache for lightweight in-memory cache
• Add Cloudflare/CDN caching on static/public endpoints
• Use ETags for smart HTTP caching