Understanding Load Balancing: Principles, Types, and Application Scenarios

Why Load Balancing Is Needed

When a system receives massive traffic, horizontal scaling by adding servers and using a load balancer distributes requests, improving processing capacity and avoiding single‑point failures.

Load Balancing Principles

Systems can scale vertically (more CPU, memory) or horizontally (more machines). Horizontal scaling requires a load‑balancing layer that forwards traffic to a cluster of servers.

Typical architecture includes an application cluster and a load‑balancing device that selects a backend server based on a balancing algorithm.

Functions of Load Balancing

Alleviate concurrency pressure and increase throughput.

Provide fault‑tolerance and high availability.

Enable elastic scaling by adding or removing servers.

Offer security features such as filtering and black/white lists.

Load Balancing Types

Based on OSI layers:

Layer‑2 (MAC) – virtual MAC address.

Layer‑3 (IP) – virtual IP address.

Layer‑4 (TCP) – IP + port forwarding.

Layer‑7 (HTTP) – virtual URL/hostname routing.

Common L4 and L7 Load Balancers

L4 solutions: F5 (hardware), LVS (heavy‑weight), Nginx (lightweight with cache), HAProxy (flexible).

L7 solutions: HAProxy, Nginx, Apache, MySQL Proxy.

Generally LVS is used for L4, Nginx for L7, while HAProxy can handle both.

Application Scenarios

High‑traffic services – distribute load across multiple servers.

Horizontal expansion – add or remove servers as demand changes.

Eliminate single‑point failures – automatically bypass faulty nodes.

Cross‑zone disaster recovery – switch between availability zones quickly.