A Comprehensive Guide to Elastic Load Balancing on AWS

default image

Elastic Load Balancing (ELB) is an invaluable service on AWS for scaling and managing traffic to applications. In my experience as a cloud architect, implementing a well-architected load balancing solution should be a priority for virtually any application hosted on AWS.

In this detailed guide, I‘ll arm you with in-depth knowledge of how ELB works, key capabilities of each load balancer option, use cases, tips and best practices. My goal is to provide the information you need to take full advantage of load balancing on AWS.

Balancing Loads: A Core AWS Capability

Load balancing is one of the most fundamental concepts in networking and AWS provides a variety of robust services for traffic distribution. According to AWS, Elastic Load Balancing handles 265 billion requests per month at peak!

At its core, a load balancer acts as a reverse proxy, sitting in front of your application instances and distributing requests across them. The primary benefits this provides are:

  • High availability – By spreading requests, you minimize disruption if a server goes down.

  • Scalability – ELB makes it easy to scale up or down as your traffic changes.

  • Performance – Load is shared evenly so no one server gets overloaded.

  • Flexibility – ELB supports a wide array of application architectures and traffic types.

AWS offers several purpose-built load balancers to pick the right tool for your needs. Next we‘ll do a deeper comparison.

Comparing Types of Load Balancers

AWS provides three main types of load balancers, each optimized for specific use cases.

Application Load Balancer

The Application Load Balancer (ALB) routes traffic based on the content of the HTTP request itself, at the layer 7 level. This allows for very fine-grained control – you can route different URLs to different target groups.

Some key capabilities:

  • Path-based Routing – Requests can be routed based on the URL path. Useful for routing microservices.
  • Host-based Routing – Routes can be based on hostname in the HTTP request header.
  • Container Support – ALB integrates with services like ECS and Kubernetes for container-based apps.
  • WebSockets – Provides native support for WebSocket connections.
  • TLS Termination -HTTPS connections are terminated at the load balancer, not the application server.

Here‘s an example architecture using an ALB:

Multi-tier web application using an application load balancer

The ALB is well suited for traditional web apps, APIs, and microservices architectures.

Network Load Balancer

The Network Load Balancer (NLB) operates at layer 4, the TCP connection level. This means it can handle much higher throughput – over 1 million requests per second!

Key capabilities:

  • Ultra Low Latency – Around 100ms vs 400ms for ALB. Critical for high performance apps.
  • TLS Pass-Through – NLB does not terminate SSL, allowing for end-to-end encryption.
  • Static IP Support – Associating static IPs is possible for whitelisting.
  • TCP/UDP – Supports load balancing TCP and UDP traffic, not just HTTP.

NLB is ideal for these use cases:

  • High performance TCP workloads
  • Real-time multiplayer gaming
  • Finance apps with UDP protocols
  • IoT applications

Here‘s an example architecture for a high throughput TCP workload:

Using a network load balancer for TCP traffic

Classic Load Balancer

The Classic Load Balancer is the legacy option, supporting basic HTTP/HTTPS and TCP load balancing. It integrates with EC2 Classic.

The Classic Load Balancer should only be used for legacy applications, as the other two balancer types now supersede it in functionality.

Now that we‘ve done a comparison, let‘s look at some key considerations when selecting a load balancer on AWS.

Choosing the Right Load Balancer

Deciding which load balancer to use depends on your specific application and traffic requirements:

Traffic Type – ALB for HTTP/HTTPS apps, NLB for TCP or UDP protocols.

Performance Needs – NLB handles much higher throughput and lower latency.

Advanced Routing – ALB enables finer-grained control for microservices.

TLS Termination – NLB supports pass-through of end-to-end encrypted traffic.

Monitoring – CloudWatch provides detailed metrics for each ELB type.

This table summarizes the key differences:

Feature ALB NLB Classic LB
Performance Up to 25,000 req/sec Over 1 million req/sec Moderate
Latency ~400ms ~100ms ~300ms
Routing methods Host, path, headers, query params TCP/UDP connection Host, path
TLS support Offloading (termination) Pass-through Offloading
Target type EC2, ECS, Lambda, IP EC2, IP EC2

Understanding these options will help you select the right load balancing approach.

Load Balancer Algorithms

Another important consideration is the load balancer algorithm that determines which target instance receives a request. The options on AWS are:

  • Round Robin – Requests alternating evenly across targets.
  • Least Outstanding Requests – Sends requests to target with least current load.
  • Least Response Time – Routes to target with lowest observed latency.
  • Source IP – Requests from the same source go to the same target.

Round Robin is the default and works well in most cases. But you can select the algorithm that best fits your app needs and traffic patterns.

In recent years, the Application Load Balancer has seen massive growth in popularity on AWS. Here are some interesting stats:

  • 75% – Portion of new load balancer instances that are ALBs according to AWS
  • 13x – Application load balancer usage growth from 2016 to 2018
  • 60% CAGR – Projected growth rate of ALB use through 2022

This shift shows the migration from legacy options toward application-based traffic routing. For most web and mobile workloads, an ALB is likely the right fit.

Load Balancer Gotchas

While Elastic Load Balancers provide great functionality, there are some common "gotchas" to be aware of:

  • Sticky Sessions – Can be an issue for stateful apps, requiring special handling.
  • Caching – Load balancers don‘t cache content, so implementing a CDN may help.
  • Logging – Make sure to enable access logs for traffic insights.
  • SSL Certificates – Require dedicated certificates for your domain on each load balancer instance.

Understanding these limitations upfront prevents surprises down the line!

Best Practices for Load Balancing on AWS

Based on my hands-on experience, here are some key best practices to follow as you architect your ELB solution:

  • Consider your traffic pattern – sudden spikes or steady loads will impact choice of balancer.
  • Use multiple target groups for more granular traffic routing.
  • Enable cross-zone load balancing for higher availability across data centers.
  • Monitor logs and metrics in CloudWatch for visibility.
  • Set up SSL certificates through ACM for better security.
  • Restrict access to backend instances through VPC constructs.
  • Utilize multiple availability zones to protect from failures.
  • Consider pairing with a content delivery network like CloudFront to cache content.

Following AWS recommended practices around high availability, security, and monitoring will ensure your load balancing yields maximum benefits.

The Future of Load Balancing on AWS

It‘s an exciting time in the field of load balancing, with new capabilities rapidly emerging:

  • Smart algorithms that dynamically adapt to real-time traffic changes.
  • Integrations with service mesh architectures to simplify microservices deployment.
  • Predictive load balancing using ML to forecast traffic and scale ahead of demand.
  • Automatic reconfiguration to instantly react and adjustconfigs to match usage patterns.

We can expect rapid innovation from AWS in this space in coming years!


Load balancing is a critical capability on AWS that enables highly available and resilient application architectures. There is an expansive set of tools available through Elastic Load Balancing that likely has a solution tailored to your specific use case.

Understanding the key differentiators between each balancer option allows you to make the optimal choice. Implementation best practices around availability, scaling, and monitoring is key to realizing the maximum benefit.

I hope this guide has armed you with a comprehensive overview of load balancing on AWS and how it can benefit your applications! Let me know if you have any other questions.

Written by