Skip to main content
In this lesson, we explore how Amazon EC2 instances, Elastic Load Balancers (ELBs), and target groups work together to deliver scalable, fault-tolerant applications on AWS.

​
Why Use a Load Balancer?

Running a single EC2 instance for mywebsite.com can become a bottleneck as traffic grows. There are two primary scaling strategies: Vertical Scaling
Scale up by increasing instance size.
Vertical scaling has limits and can be cost-inefficient when traffic is variable.
The image is a diagram illustrating the need for a load balancer, showing a connection from AWS Cloud to a website and users.
Horizontal Scaling
Run multiple identical instances behind a single endpoint. DNS still points to one IP, so without a load balancer new servers receive no traffic.
The image is a diagram illustrating the need for a load balancer in an AWS cloud setup, showing multiple servers connected to a website and users.
Introducing an Elastic Load Balancer solves this:
The image is a diagram illustrating the use of a load balancer in an AWS cloud setup, showing multiple servers distributing traffic to a website.
As demand grows, new EC2 instances register automatically with the ELB, making scaling transparent to end users.

​
How Load Balancing Works in AWS

AWS ELB is highly available and fault tolerant across multiple Availability Zones (AZs). It distributes incoming traffic to target groups—collections of instances, IP addresses, or Lambda functions. Each load balancer uses one or more listeners to check for connections on a port (e.g., port 80 for HTTP). Listener rules determine how to forward requests:
The image illustrates how a load balancer works within an AWS cloud environment, showing traffic distribution across public subnets and target groups to a website.
  • Default HTTPS port: 443
  • Custom listener ports are supported (e.g., 8080)
  • Rules are evaluated by priority
  • Load balancers can be public-facing or internal

​
Protocols and Ports

Load Balancer TypeDefault PortSupported Protocols
HTTP/HTTPS80 / 443HTTP, HTTPS, HTTP/2, gRPC
TCP/UDPAnyTCP, UDP, TLS
Their main goal is to provide scalability and high availability.

​
Target Groups

A target group defines the routing and health-check configuration for one or more registered targets. You specify:
  • Protocol and port
  • Health check path and thresholds
  • Target type: instance, IP, or Lambda
Supported protocols include HTTP, HTTPS, TCP, UDP, HTTP/1, HTTP/2, and gRPC.

​
Health Checks

Load balancer nodes perform periodic health checks on each target. Unhealthy targets are removed from rotation until they recover.
Configure health check intervals and thresholds to balance rapid failover with avoiding false positives.
The image illustrates a load balancer health check process within an AWS cloud environment, showing client requests being directed through public subnets.

​
Public vs. Private Load Balancers

TypeInternet-FacingWithin VPC
Public Load BalancerYesNo
Private Load BalancerNoYes
Both improve fault tolerance and allow seamless scaling.
The image illustrates a network architecture with public and private load balancers in an AWS cloud environment, showing connections between private and public subnets, a website, and users.

​
Cross-Zone Load Balancing

With cross-zone load balancing enabled, each ELB node evenly distributes traffic across all registered targets in every AZ. This prevents hotspots and idle instances.
The image illustrates the concept of cross-zone load balancing in AWS, showing multiple public subnets with instances connected to a load balancer, which distributes traffic across zones.
Disable cross-zone load balancing only for very specific network requirements. Most deployments should keep it enabled.

​
Types of AWS Load Balancers

Load Balancer TypeOSI LayerUse Case
Application Load Balancer (ALB)Layer 7Advanced HTTP/HTTPS routing and features
Network Load Balancer (NLB)Layer 4High performance TCP/UDP load balancing
The image describes two types of load balancers supported by AWS: Application Load Balancer (ALB) and Network Load Balancer (NLB), highlighting their features and the OSI layers they operate on.

​
Network Load Balancer (NLB)

  • Targets: instance IDs, IP addresses, or ALBs
  • One network interface per AZ for a static IP
  • Optionally assign Elastic IPs to each subnet
  • Routes TCP/UDP/TLS to target groups by port
  • Can route to resources outside the VPC (VPN or Direct Connect)
The image illustrates how a Network Load Balancer (NLB) works, showing the flow of data between a corporate data center and AWS Cloud through public subnets, with connections to various ports (8080, 80, 443).

​
Application Load Balancer (ALB)

ALB rules consist of conditions and a single action (forward, redirect, or fixed-response). Supported conditions:
  • Host header
  • Path
  • HTTP method
  • Source IP
  • HTTP header
  • Query string
The image illustrates the features of an AWS Application Load Balancer (ALB), showing various rules like Host Header, Path, and HTTP Request Method, with options for forwarding, redirecting, and fixed responses.

​
Host Header Routing

Forward requests based on the Host header.
Example:
  • If Host: blog.mywebsite.com, route to blog target group.
  • Default: route all other traffic to default.
The image is a diagram illustrating an AWS Application Load Balancer (ALB) configuration with a host header rule for "blog.mywebsite.com" and a default target group.

​
Path-Based Routing

Match URL paths (e.g., /blog) and forward to the corresponding target group.

​
HTTP Method Routing

Match HTTP methods (e.g., POST) and forward to an API target group.

​
Source IP Routing

Allow requests from specific IP addresses to a designated group.

​
HTTP Header Routing

Match custom headers (e.g., x-environment: staging) to route to a staging environment:
The image is a diagram illustrating an AWS Application Load Balancer (ALB) HTTP header configuration, showing how requests are routed based on header rules to different target groups.

​
Query String Routing

Inspect query parameters (e.g., ?category=books) and forward accordingly:
The image illustrates an AWS Application Load Balancer (ALB) configuration with query string rules, showing how requests with the query string "category=books" are directed to a specific target group, while others follow a default rule.

​
Rule Priority Example

curl -H "x-client: premium" http://mywebsite.com/api
curl -H "x-client: medium"  http://mywebsite.com/api
curl                        http://mywebsite.com/api
  1. Priority 1: Path /api + header x-client: premium → large
  2. Priority 2: Path /api + header x-client: medium → medium
  3. Default: All other /api → default-api

​
Elastic Load Balancer Integrations

ELBs integrate with many AWS services to deliver end-to-end, scalable architectures:
The image is a diagram showing the integration of Elastic Load Balancing with various AWS services, including Amazon EC2, Amazon ECS, AWS Lambda, AWS WAF, Amazon Route 53, and Autoscaling.
  • EC2: Distribute incoming traffic across instances
  • ECS: Balance containerized workloads
  • Lambda: Route requests to serverless functions
  • WAF: Apply security rules to incoming traffic
  • Route 53: Use DNS to map domains to load balancers
  • Auto Scaling: Automatically adjust capacity and register/deregister targets

Watch Video

Practice Lab