Field Guide · concept

A load balancer distributes incoming network traffic across multiple servers so that no single machine is overwhelmed, improving capacity, responsiveness, and availability.1

clients Load balancer spread · health-check server server server ✕ server removed a failed health check is pulled from the pool — the rest keep serving
The balancer fronts a pool of identical backends and spreads requests among them (round-robin, least-connections, or client hashing). Health checks pull a dead server out automatically, so the pool survives losing any one member — the basis of both scaling out and staying available.

Overview

A load balancer sits in front of a pool of identical backends and spreads requests among them using rules such as round-robin, least-connections, or hashing on the client. It also runs health checks, automatically removing a server that stops responding so users are not routed to a dead machine. Balancers operate at the transport layer (L4, forwarding TCP/UDP) or the application layer (L7, inspecting HTTP to route by path or host). They may be dedicated hardware appliances or software running on commodity servers.

Where it fits

Load balancing is fundamental to scalability — adding servers behind a balancer increases throughput — and to high availability, since the pool survives the loss of any one member. It overlaps with a reverse proxy, which also fronts backends, and with a content delivery network for global traffic. Orchestrators like Kubernetes provide built-in balancing across pods.

Sources

  1. Load balancing (computing) — Wikipedia, on load balancing techniques. ↩

See also