HTTP Keep-Alive

HTTP Keep-Alive (persistent connections) allows multiple HTTP requests/responses over a single TCP connection, avoiding the overhead of establishing a new connection for each request.

Without Keep-Alive (HTTP/1.0)

Request 1: TCP handshake → Request → Response → Close
Request 2: TCP handshake → Request → Response → Close
Request 3: TCP handshake → Request → Response → Close

Total: 3 TCP handshakes + 3 requests

With Keep-Alive (HTTP/1.1+)

Connection Established (1 TCP handshake)
  |
  +-- Request 1 → Response 1
  +-- Request 2 → Response 2
  +-- Request 3 → Response 3
  |
Connection Closed

Total: 1 TCP handshake + 3 requests

Keep-Alive Headers

# Client request
Connection: keep-alive

# Server response
Connection: keep-alive
Keep-Alive: timeout=5, max=100

# Close connection explicitly
Connection: close

Keep-Alive Parameters

Parameter	Description
timeout	Seconds to keep idle connection alive
max	Maximum requests per connection

Performance Impact

Without Keep-Alive:
- 3 requests × (TCP handshake: 30ms + TLS: 50ms) = 240ms overhead

With Keep-Alive:
- 1 handshake (80ms) for 3 requests = 80ms overhead
- 67% reduction in connection overhead

Keep-Alive in Practice

// Node.js
const http = require('http');

const server = http.createServer((req, res) => {
  res.end('Hello');
});

// Enable keep-alive (default in HTTP/1.1)
server.keepAliveTimeout = 5000; // 5 seconds

Why Keep-Alive Matters

1. Reduces latency — No repeated TCP/TLS handshakes
2. Improves throughput — Multiple requests pipelined
3. Reduces server load — Fewer connections to manage
4. Better UX — Faster page loads

Methods

• GET: Read (safe, idempotent)
• POST: Create
• PUT: Replace (idempotent)
• PATCH: Partial update
• DELETE: Remove (idempotent)

Headers

• Content-Type: Body format
• Cache-Control: Caching rules
• Authorization: Authentication
• Accept: Desired response format

Status Codes

• 2xx: Success
• 3xx: Redirection
• 4xx: Client error
• 5xx: Server error

Key Points

• Understanding HTTP Keep-Alive is essential for production systems
• Always consider scalability and maintainability
• Test thoroughly before deploying to production
• Monitor performance and set up alerting

Common Patterns

1. Validation: Always validate input at the boundary
2. Error Handling: Use structured error responses
3. Logging: Log key events for debugging
4. Testing: Unit, integration, and load tests
5. Documentation: Keep docs updated with code changes