Networking and Content Delivery
Elastic Load Balancing (ELB)
Elastic Load Balancing distributes traffic across healthy targets using Application, Network, Gateway, or Classic Load Balancers. Custom GenAI services commonly use ALB for HTTP routing or NLB for high-performance transport behavior.
Key points
- Health checks remove unhealthy targets from routing
- ALB supports content-based HTTP routing and WebSockets
- NLB preserves transport-level performance and static endpoint characteristics
When to use it
- Balance requests across ECS inference gateway tasks
- Route streaming chat and administrative paths to separate target groups
Exam tips
- Choose the load-balancer type by protocol and routing needs, not brand familiarity
- Set deregistration delay and application shutdown to avoid cutting off long streamed responses