GSLB

acronym

cloudnetworking

Global Server Load Balancing: DNS-level traffic steering across sites and regions - answering each resolver with the healthiest, closest, or least-loaded destination.

Failover at the speed of TTLs; the discipline behind every 'active-active across regions' claim.

Global load balancing answers a question local load balancing cannot: not which server, but which site. When an application runs in more than one data centre or region, something has to decide which one a given user reaches, and because that decision has to happen before any connection is made, it is usually made in DNS.

That DNS dependency is the source of most surprises. The answer is cached by resolvers and by the client, so a change does not take effect everywhere the moment you make it, and a TTL that seemed sensible during design becomes a failover delay during an outage. Health monitoring decides whether a site is eligible at all, and topology rules decide who is preferred, with the most specific matching rule winning.

The failure modes are worth knowing in advance. A site that is up but degraded may still pass a naive health check. A resolver far from the user makes the geographic decision look wrong when it is working exactly as configured. And an aggressive TTL that gives you fast failover also multiplies query load, which is the trade you are actually choosing.

All glossary entries