Purpose of This Lesson

Amazon Elastic Load Balancing (ELB) provides a single application endpoint for users while distributing connections across healthy targets, such as EC2 instances in an Auto Scaling group. The key assessment skill is choosing the correct load balancer type based on the required protocol layer, routing information, performance characteristics, and target type.

Key Concepts

Elastic Load Balancing workflow

A load balancer is placed in front of application targets, commonly EC2 instances distributed across Availability Zones and subnets. Users connect to the load balancer’s published endpoint rather than directly to individual instances.

The load balancer distributes traffic across targets in a target group. It continuously performs health checks to determine whether the application is still responding correctly. When a target fails its health check, ELB removes it from service and routes new traffic to healthy targets.

When ELB health checks are enabled for an Auto Scaling group, the group can receive the load balancer’s health status. Auto Scaling can terminate an unhealthy instance and launch a replacement, restoring the desired capacity.

Application Load Balancer (ALB)

An Application Load Balancer operates at OSI Layer 7, the application layer, and makes routing decisions at the HTTP/HTTPS request level. Because it can inspect HTTP and HTTPS information, it supports routing based on:

• URL path
• Host name
• Query-string parameters
• Source IP address

An ALB can route requests to different target groups. Supported target types from this lesson include EC2 instances, IP addresses, Lambda functions, and containers.

ALB is a natural fit for web applications and microservices architectures where request-level information determines the destination.

Network Load Balancer (NLB)

A Network Load Balancer operates at OSI Layer 4, the connection level. It does not inspect HTTP or HTTPS headers, so it cannot make the Layer 7 routing decisions available to an ALB.

NLB is designed for TCP- and UDP-based applications that require ultra-high performance, low latency, or high throughput. It also supports TLS offloading at scale, static or Elastic IP addresses, and static IPs as targets. NLB supports VPC endpoint services, which is relevant when exposing services through AWS PrivateLink-style endpoint service architectures covered elsewhere in the course.

Reliability and Business Continuity Context

ELB contributes to availability by presenting one endpoint while distributing traffic across multiple targets. Health checks prevent traffic from continuing to flow to a target that is no longer responding correctly.

A typical recovery sequence is:

1. A target fails its ELB health check.
2. ELB removes the target from service.
3. New connections are routed to healthy targets.
4. If ELB health checks are enabled for the Auto Scaling group, Auto Scaling receives the unhealthy status.
5. Auto Scaling terminates the failed instance and launches a replacement.

This sequence separates two responsibilities: ELB controls traffic eligibility, while Auto Scaling maintains the target fleet. ELB does not itself replace an EC2 instance; Auto Scaling performs that replacement when configured to use the relevant health information.

Exam- or Assessment-Relevant Takeaways

• Choose ALB when routing depends on HTTP/HTTPS request data such as host names, paths, query strings, or source IP addresses.
• Choose NLB for TCP or UDP applications, very low latency, high throughput, or a requirement for static or Elastic IP addresses.
• ALB is Layer 7 and request-aware; NLB is Layer 4 and connection-oriented.
• An NLB can support applications that use HTTP or HTTPS at the application level, but the NLB itself does not inspect HTTP/HTTPS headers for Layer 7 routing.
• ELB health checks determine whether targets remain in service.
• To have unhealthy load-balanced instances replaced, configure the Auto Scaling group to use ELB health checks. Auto Scaling may also detect instance problems through its own status checks.
• ALB target types include instances, IP addresses, Lambda functions, and containers.
• NLB supports TCP and UDP workloads, static or Elastic IP addresses, TLS offloading at scale, and VPC endpoint services.

Tool / Feature Decision Guide

Requirement	Preferred choice	Reason
Route web requests by URL path	ALB	It can inspect HTTP/HTTPS request information.
Route by host name	ALB	Host-based routing is a Layer 7 capability.
Route by query-string parameter	ALB	Query-string routing uses request-level information.
Support a TCP application	NLB	NLB operates at the connection layer and supports TCP.
Support a UDP application	NLB	UDP support is a stated NLB capability.
Minimize latency and support high throughput	NLB	NLB is intended for ultra-high-performance, low-latency use cases.
Provide a static or Elastic public IP address	NLB	Static and Elastic IP support is associated with NLB in this lesson.
Route to Lambda functions or containers	ALB	These are listed ALB target types.
Expose a service through a VPC endpoint service	NLB	NLB supports VPC endpoint services.

Common Traps / Misconceptions

• “HTTP automatically means ALB.” An application may use HTTP or HTTPS behind an NLB, but NLB still operates at Layer 4 and cannot inspect HTTP headers. The decisive requirement is whether request-aware routing is needed.
• “A failed health check immediately terminates the instance.” ELB removes the target from service first. Instance termination and replacement are Auto Scaling responsibilities when ELB health checks are enabled for the group.
• “The load balancer replaces failed instances.” ELB redirects traffic and tracks target health; it does not maintain the desired number of EC2 instances.
• “One public endpoint means one backend instance.” The endpoint is shared while traffic is distributed across targets in target groups.
• “NLB is only for public applications.” The lesson specifically identifies NLB support for VPC endpoint services, so NLB can also be relevant to service exposure through VPC-based endpoint architectures.
• “ALB and NLB differ only in speed.” Their fundamental distinction is Layer 7 request routing versus Layer 4 connection handling, with different protocol and feature implications.

Real-World Engineer / Analyst Notes

• Design target groups around the routing and application boundaries your service requires. ALB can select among target groups using request attributes, while NLB forwards at the connection level.
• Treat health checks as an operational contract: a target must respond correctly to the configured check for ELB to keep sending it traffic.
• Pair ELB health checks with Auto Scaling health-check configuration when failed application targets should be replaced automatically.
• When investigating an outage, distinguish between a target being removed from service, an instance failing its own status checks, and Auto Scaling launching a replacement. These are related events but not the same mechanism.
• Choose the simplest layer that satisfies the requirement. If no HTTP/HTTPS inspection is needed and the workload requires TCP/UDP handling, low latency, or static IPs, NLB may be more appropriate than ALB.

Quick Reference Summary

• ELB purpose: Provide one endpoint and distribute traffic across healthy targets.
• ALB: Layer 7, HTTP/HTTPS, request-level routing, web applications, microservices, Lambda and container targets.
• ALB routing: Path, host, query-string parameter, and source IP address.
• NLB: Layer 4, TCP/UDP, ultra-low latency, high throughput, static or Elastic IPs, TLS offloading at scale, and VPC endpoint services.
• Health checks: Failed targets are taken out of service and traffic is sent to healthy targets.
• Auto Scaling integration: ELB health status can cause Auto Scaling to terminate unhealthy instances and launch replacements.

Flashcards

Q: A web application must route /images and /api requests to different target groups. Which load balancer should be selected and why?

A: Choose an Application Load Balancer because it operates at Layer 7 and supports path-based routing using HTTP/HTTPS request information.

Q: When should an NLB be chosen instead of an ALB for a service that uses an application protocol over the network?

A: Choose NLB when the decisive requirements are TCP or UDP support, ultra-low latency, high throughput, or static/Elastic IP addresses rather than HTTP header inspection.

Q: What is the key functional difference between ALB and NLB?

A: ALB makes Layer 7, request-level decisions by inspecting HTTP/HTTPS information. NLB makes Layer 4, connection-level decisions and does not inspect HTTP/HTTPS headers.

Q: A client must reach a service through a stable public IP address. Which ELB type is indicated by this requirement?

A: NLB, because it supports static or Elastic IP addresses.

Q: What happens when an ELB health check fails for a target?

A: ELB takes the target out of service and routes new traffic to healthy targets. The target is not necessarily terminated by ELB itself.

Q: How does Auto Scaling complement ELB health checks?

A: When ELB health checks are enabled for the Auto Scaling group, Auto Scaling can learn that a load-balanced instance is unhealthy, terminate it, and launch a replacement.

Q: Which ALB routing option should be considered when the destination depends on the requested host name?

A: Host-based routing, because the ALB can inspect the HTTP/HTTPS host information and select the appropriate target group.

Q: Which target types are supported by ALB according to this lesson?

A: EC2 instances, IP addresses, Lambda functions, and containers.

Q: What is the trap when an application uses HTTPS but the requirement calls for an NLB?

A: HTTPS alone does not require ALB. NLB can carry the connection and support TLS offloading at scale, but it cannot use HTTP/HTTPS headers for Layer 7 routing.

Q: A service uses UDP and needs very low latency. Which load balancer is the better fit?

A: NLB, because it supports UDP and is intended for ultra-low-latency, high-performance workloads.

Q: Why does ELB provide a single application endpoint in a multi-instance architecture?

A: Users connect to the load balancer endpoint, and ELB determines which healthy target should receive each connection or request. Users do not need to address individual backend instances.

Q: What is the relationship between ELB health checks and EC2 instance status checks?

A: ELB checks whether the target application is responding correctly, while Auto Scaling may also learn about infrastructure problems through EC2 status checks. Either signal can contribute to identifying an unhealthy instance, but replacement is performed by Auto Scaling.

Q: Which load balancer supports VPC endpoint services in the material?

A: Network Load Balancer supports VPC endpoint services.

Practice Questions

Question 1

A company runs several web services behind one public endpoint. Requests with the host name orders.example.com must go to one target group, while billing.example.com must go to another. Which design best fits the requirement?

A. Network Load Balancer using UDP listeners
B. Application Load Balancer using host-based routing
C. Network Load Balancer using static IP targets
D. Auto Scaling without a load balancer

Correct answer: B. Host-based routing requires inspection of HTTP/HTTPS request information, which is an ALB Layer 7 capability.

Question 2

A real-time service uses UDP, requires very low latency, and must be reachable through a static public IP address. Which option is most appropriate?

A. Application Load Balancer
B. Network Load Balancer
C. Application Load Balancer with path-based routing
D. Auto Scaling health checks only

Correct answer: B. NLB supports UDP, ultra-low-latency workloads, and static or Elastic IP addresses.

Question 3

An EC2 instance behind a load balancer stops responding to the configured application health check. What should an engineer expect first?

A. ELB immediately launches a replacement instance
B. ELB removes the target from service and routes traffic to healthy targets
C. The load balancer changes from ALB to NLB
D. The instance is guaranteed to remain in service until it fails an EC2 status check

Correct answer: B. ELB responds to failed health checks by taking the target out of service; Auto Scaling is responsible for replacement when configured to use ELB health checks.

Question 4

An Auto Scaling group contains instances that are registered with an ELB. The team wants instances that fail the load balancer’s health checks to be replaced automatically. Which configuration is essential?

A. Enable ELB health checks for the Auto Scaling group
B. Use query-string routing on the load balancer
C. Assign a static IP to every ALB target
D. Disable target health checks so connections persist

Correct answer: A. Enabling ELB health checks for the Auto Scaling group allows Auto Scaling to receive the unhealthy status and replace the affected instance.

Question 5

A service uses HTTPS, but all traffic should be forwarded based only on connection-level behavior. The service also requires high throughput and does not need path- or host-based routing. Which choice is most consistent with the lesson?

A. NLB, because HTTPS can be carried without Layer 7 header-based routing
B. ALB, because every HTTPS service requires request-level routing
C. Auto Scaling alone, because load balancers cannot process HTTPS
D. ALB, because ALB is the only option that supports high throughput

Correct answer: A. NLB can support a service using HTTPS while operating at Layer 4; the decisive clues are the lack of request-level routing and the high-throughput requirement.

WordPress Metadata

Suggested Slug:
amazon-elastic-load-balancing-alb-nlb-health-checks

Meta Description:
Study Amazon Elastic Load Balancing for SOA-C03, including ALB versus NLB selection, health checks, routing options, Auto Scaling integration, and target types.

Tags:
AWS, Elastic Load Balancing, Application Load Balancer, Network Load Balancer, ALB, NLB, Auto Scaling, Health Checks, High Availability, SOA-C03