The Real Reason AWS Elastic Load Balancing (ELB, ALB, NLB) Was Designed the Way It Was

In the world of AWS, Elastic Load Balancing is a foundational service, quietly directing traffic to keep applications running smoothly. But the fact that it’s not one service, but a family of three, isn’t an accident. It’s a story of evolution.

The Original Problem: One Balancer to Rule Them All

Back in 2009, when AWS launched the first Elastic Load Balancer, the cloud was a simpler place. The goal was straightforward: give developers a way to automatically distribute traffic across multiple EC2 instances without the headache of managing physical

hardware. This initial service, now known as the Classic Load Balancer (CLB), was a jack-of-all-trades. It operated at both the connection level (Layer 4) and the application level (Layer 7), providing a simple, elastic black box that just worked. For the monolithic applications of the era, this was revolutionary. It offered basic health checks and fault tolerance, ensuring traffic only went to healthy servers.

When 'Simple' Suddenly Wasn't Enough

As the cloud matured, so did application architecture. The monolithic model started giving way to microservices—small, independent services that work together. This new paradigm broke the Classic Load Balancer's "one-size-fits-all" model. Developers needed to route traffic based on more than just an open port; they needed to direct requests based on the URL path, hostname, or other application-specific data. For instance, traffic for `/api/users` needed to go to the user service, while `/api/orders` went to the order service. The Classic Load Balancer couldn't do this efficiently, lacking support for path-based or host-based routing. It became a bottleneck for modern application delivery.

A Fork in the Road: The Application-Aware Brain (ALB)

To solve this, AWS introduced the Application Load Balancer (ALB) in 2016. Unlike its predecessor, the ALB was purpose-built for Layer 7, the application layer. This specialization was its superpower. The ALB could inspect HTTP/HTTPS requests and make intelligent routing decisions based on their content. Suddenly, developers could create complex rules to send traffic to different backend services, called target groups, based on URL paths or hostnames. This was a perfect fit for microservices and containerized applications running on services like Amazon ECS, which could now dynamically register tasks on different ports of the same instance. The ALB was designed to be the smart traffic cop for modern, content-rich applications.

Back to Basics for Raw Speed: The Network Load Balancer (NLB)

While the ALB solved the application routing problem, another need emerged: extreme performance. Some workloads, like high-frequency trading platforms, real-time gaming servers, or IoT applications, don't need intelligent HTTP routing. They need raw, unadulterated speed and the ability to handle millions of requests per second with ultra-low latency. For this, AWS created the Network Load Balancer (NLB) in 2017. The NLB operates purely at Layer 4 (the transport layer), forwarding TCP and UDP traffic with minimal overhead. It doesn't inspect application content; it just moves packets as fast as possible. The NLB was also designed to provide a static IP address for each Availability Zone, a critical feature for services that need a fixed, predictable entry point. It was built to be a high-performance network pipe, handling volatile traffic patterns without breaking a sweat.