Layer 4 Load Balancing

Load balancing performed at the OSI transport layer (Layer 4) using TCP/UDP IP addresses and port numbers to route traffic without inspecting application-layer content such as HTTP headers or URLs. L4 load balancers establish two TCP connections — one from client to LB and one from LB to backend — and forward bytes between them at very high throughput with minimal latency overhead. Because they do not parse application data, L4 balancers cannot perform URL-based routing, cookie-based sticky sessions, or TLS termination. Common implementations include AWS NLB, HAProxy TCP mode, and Linux's IPVS (used by kube-proxy in Kubernetes).