Topic 35 of 40

In-Memory Caching

Overview

If a user visits your homepage, and the backend runs a complex SQL query that takes 200ms to calculate the 'Top 10 Trending Products', that is fine. But if 10,000 users visit your homepage simultaneously, running that identical heavy SQL query 10,000 times will melt the database. Caching solves this. You run the query exactly ONCE, and store the resulting JSON directly in lightning-fast RAM (usually using Redis or Valkey). The next 9,999 users bypass the database entirely and fetch the JSON directly from RAM in 1 millisecond.

Syntax

javascript

// Utilizing Redis (or Valkey) via the 'redis' npm package
const redis = require('redis');
const client = redis.createClient({ url: process.env.REDIS_URL });
await client.connect();

app.get('/api/trending', async (req, res) => {
    // 1. CHECK THE CACHE FIRST!
    const cachedData = await client.get('trending_products');
    
    // 2. CACHE HIT! The data was in RAM. Return it instantly.
    if (cachedData) {
        return res.json(JSON.parse(cachedData));
    }
    
    // 3. CACHE MISS! The data is not in RAM. We must hit the heavy Database.
    const products = await db.query('SELECT ... HEAVY QUERY ...');
    
    // 4. POPULATE THE CACHE! Save the result in Redis for the next user.
    // 'EX 60' means Expire in 60 seconds (Cache Invalidation)
    await client.set('trending_products', JSON.stringify(products), { EX: 60 });
    
    res.json(products);
});

Common Pitfalls

Cache Invalidation (The hardest problem in Computer Science). If you cache a user's Profile Data forever, and they update their profile picture in the database, the API will continue serving the old picture from the cache. You MUST implement strict TTLs (Time-To-Live, e.g., expire after 5 mins) or explicitly run client.del('profile_42') every time they run an UPDATE query.
Caching sensitive user data globally. If you cache a query like SELECT * FROM user_bank_accounts under a generic key like cache:accounts, User B might hit the cache and receive User A's bank details. Always append unique identifiers to cache keys for sensitive data (e.g., cache:accounts:user_42).

Interview Questions

Why use an external service like Redis for caching instead of just storing the data in a standard JavaScript Object/Map inside the Node.js process (const cache = {})?

If you scale your app to 5 Node.js servers behind a Load Balancer, local memory is completely isolated. User A hits Server 1 (populates the local cache). User A refreshes, the Load Balancer routes them to Server 2 (Cache Miss!). Redis is a centralized, external RAM store that ALL 5 servers share perfectly.

Real-World Example

Using Redis to implement a distributed, high-performance API Rate Limiter.

example

javascript

// Instead of keeping counts in local Node.js variables,
// Redis guarantees atomic incrementation across all servers!
const currentRequests = await redis.incr(`rate_limit:${req.ip}`);

if (currentRequests === 1) {
    // If it's their first request, set the key to expire in 1 minute!
    await redis.expire(`rate_limit:${req.ip}`, 60);
}

if (currentRequests > 100) {
    return res.status(429).send("Too Many Requests");
}

Check Your Knowledge

Test your understanding of In-Memory Caching with these quick questions.

What is the industry-standard term for a scenario where the application checks the Cache, fails to find the requested data, and is forced to query the primary Database?

Take Node.js Quiz

Overview

Syntax

javascript

// Utilizing Redis (or Valkey) via the 'redis' npm package
const redis = require('redis');
const client = redis.createClient({ url: process.env.REDIS_URL });
await client.connect();

app.get('/api/trending', async (req, res) => {
    // 1. CHECK THE CACHE FIRST!
    const cachedData = await client.get('trending_products');
    
    // 2. CACHE HIT! The data was in RAM. Return it instantly.
    if (cachedData) {
        return res.json(JSON.parse(cachedData));
    }
    
    // 3. CACHE MISS! The data is not in RAM. We must hit the heavy Database.
    const products = await db.query('SELECT ... HEAVY QUERY ...');
    
    // 4. POPULATE THE CACHE! Save the result in Redis for the next user.
    // 'EX 60' means Expire in 60 seconds (Cache Invalidation)
    await client.set('trending_products', JSON.stringify(products), { EX: 60 });
    
    res.json(products);
});

Common Pitfalls

Cache Invalidation (The hardest problem in Computer Science). If you cache a user's Profile Data forever, and they update their profile picture in the database, the API will continue serving the old picture from the cache. You MUST implement strict TTLs (Time-To-Live, e.g., expire after 5 mins) or explicitly run client.del('profile_42') every time they run an UPDATE query.
Caching sensitive user data globally. If you cache a query like SELECT * FROM user_bank_accounts under a generic key like cache:accounts, User B might hit the cache and receive User A's bank details. Always append unique identifiers to cache keys for sensitive data (e.g., cache:accounts:user_42).

Interview Questions

Why use an external service like Redis for caching instead of just storing the data in a standard JavaScript Object/Map inside the Node.js process (const cache = {})?

Real-World Example

Using Redis to implement a distributed, high-performance API Rate Limiter.

example

javascript

// Instead of keeping counts in local Node.js variables,
// Redis guarantees atomic incrementation across all servers!
const currentRequests = await redis.incr(`rate_limit:${req.ip}`);

if (currentRequests === 1) {
    // If it's their first request, set the key to expire in 1 minute!
    await redis.expire(`rate_limit:${req.ip}`, 60);
}

if (currentRequests > 100) {
    return res.status(429).send("Too Many Requests");
}