Cloud Load Balancing
High-performance, global Anycast load balancing scaling to 1M+ queries/sec with zero pre-warming.
High-performance, global Anycast load balancing scaling to 1M+ queries/sec with zero pre-warming.
What is Cloud Load Balancing?
Google Cloud Load Balancing is a software-defined, distributed proxy service that routes incoming user traffic to the nearest healthy backend compute instance worldwide. Using Anycast IP routing, client requests enter Google's fiber network at the closest Point of Presence (PoP) edge datacenter, minimizing latency and providing automated multi-region global failover with zero pre-warming required.
Decision Guide: When to Choose This Service
Use Cloud Load Balancing to distribute HTTP(S), TCP, and UDP traffic across backend instance groups, GKE pods, Cloud Run services, and multi-region server fleets with a single global Anycast IP address.
Do NOT provision a dedicated load balancer for simple Cloud Run services if native Cloud Run HTTPS routing with traffic splitting is sufficient.
Core Architectural Pillars
A single IPv4 and IPv6 address is advertised globally from all Google edge locations, automatically routing users to the geographically closest healthy region.
Global External App LB (Layer 7 HTTP/S), Regional External App LB, Internal App LB (Layer 7 inside VPC), and Network LB (Layer 4 TCP/UDP passthrough).
Backend Services route traffic to Instance Groups, GKE Network Endpoint Groups (NEGs), or Cloud Run / Cloud Functions (Serverless NEGs).
Attach DDoS protection, rate limiting, IP whitelists, and edge cache acceleration with a single click.