DNS-Based Load Balancing

Round-robin, weighted.

Overview

DNS-based load balancing distributes traffic across endpoints by returning different addresses to different clients. The four policies below cover most real-world distribution needs and pick the right one for the workload at hand.

The approach

Three habits keep DNS load balancing operationally sound: short TTLs to enable fast changes, latency-based as the default policy, and explicit weights for canary rollouts.

Why this compounds

The first DNS load-balancing policy teaches the team how the trade-off between latency, weight, and health behaves at their scale. Subsequent policies reuse the framework.