Elastic Load Balancing
#37 · Level 2 · Scaling and resilience · Scaling and trafficConcept
Section titled “Concept”Elastic Load Balancing spreads incoming requests across several servers. The places requests are sent to are called targets — such as EC2 instances, containers, IP addresses or Lambda functions. The load balancer runs regular health checks, and stops sending requests to unhealthy targets. Its main types are the Application Load Balancer (for web traffic), the Network Load Balancer (for fast TCP/UDP traffic) and the Gateway Load Balancer. Because it runs spread across several Availability Zones, it is often used together with EC2 Auto Scaling.
Simple example
Section titled “Simple example”The line manager at a supermarket’s checkout counters. They send each customer to whichever counter is free. If a counter closes, they stop sending anyone there.
Why it’s needed
Section titled “Why it’s needed”If all the load falls on one server, it slows down or stops. Spreading the load means everyone is served quickly, and the site keeps running even if one server fails.
Where it’s used
Section titled “Where it’s used”It sits in front of a website, with several EC2 instances or containers behind it in different AZs.
Key terms
Section titled “Key terms”| English | Simple meaning |
|---|---|
| Application Load Balancer | For websites and APIs; can look at the URL to send requests to the right place |
| Network Load Balancer | Very fast, for very high traffic |
Azure equivalent
Section titled “Azure equivalent”Azure Load Balancer / Application Gateway
Related
Section titled “Related”Amazon EC2 Auto Scaling · Amazon EC2 · Availability Zone (AZ)
Learn more
Section titled “Learn more”For details and the latest information, see the official documentation: