Elastic Load Balancing

About

Elastic Load Balancing (ELB) automa­tically distri­butes incoming applic­ation traffic across multiple targets, such as EC2’s, contai­ners, IP addresses, & Lambda functions.

Elastic Load Balancing is a best practice to assign incoming traffic to a single target, such as an EC2 Instance, and then distribute the rest of the traffic across the target’s resources. An elastic load balancer distributes traffic across an arbitrary number of targets.

• Documentation
• User Guide

Types of Load Balancers

Compare ALB vs NLB vs CLB

Scaling

Vertical Scaling

• Increasing the size of the instances (ie- increase in RAM and vCPUs )
• Ex: from t2.micro to t3.2xlarge (doesn’t have to be the same instance family)
• In vertical scaling, you scale up/down
• Vertical scaling usually happens in databases, to handle high workloads as your application grows

*Horizontal Scaling

• Increasing the no. of instances
• In horizontal scaling, you scale out/in
• Ex: ASG scaling out EC2 instances to match increased workload for your web application

Cross Zone Load Balancing

ALB

• Enabled by default. Cannot disable it
• Not charges for data transfer between AZs (inter AZ data transfer)

CLB

• Enabled by default. Can disable it
• Not charged for data transfer between AZs

NLB

• Disabled by default. Can enable it
• Charged for data transfer between AZs

Digest

• ELB(Elastic Load Balancing) distributes application or network traffic across multiple targets, such as EC2 instances, containers(ECS), and IP addresses, in multiple AZs.
• Cross Zone Load Balancing – when enabled, each load balancer node distributes traffic across the registered targets in all enabled AZs.
• 3 Types of Load balancers - Application, Network, Classic
• Deleting ELB won’t delete the instances registered to it.
• Termination protection will be disabled by default; enable it to prevent accidental delete.
• 504 error means the gateway has timed out and the application is not responding within the idle timeout period
• Look for the X-Forwarded-For header, if you need the end user IPv4 address
• ASG (Auto Scaling Group) ensures you’ve the correct number of EC2 instances available. Specify minimum, maximum and desired number of instances.
• Lifecycle hook - perform custom actions when instances launch or terminate
• Cool down period - ensure not to launch additional instances before previous scaling activities complete
• Launch configuration - Instance configuration template the ASG uses to launch EC2 instances

Price

Current price

Use Cases

Type: Scale your network design

• AWS discourages the use of Classic Load Balancer in favor of its newer load balancers

• Application Load Balancer is typically used for web applications.

• Network Load Balancer would be used for anything that ALBs don’t cover. A typical use case would be a near real-time data streaming service (video, stock quotes, etc.) Another typical case is that you would need to use an NLB if your application uses non-HTTP protocols

Practice

Create Classic Load Balancer

Questions

Q1

Which load balancer would you use for services which use HTTP or HTTPS traffic?

Q2

What are possible target groups for ALB (Application Load Balancer)?

Q3

Your would like to optimize the performance of their web application by routing inbound traffic to api.mysite.com to Compute Optimized EC2 instances and inbound traffic to mobile.mysite.com to Memory Optimized EC2 instances.

Which solution below would be best to implement for this?

1. Enable X-Forwarded For on the web servers and use a Classic Load Balancer
2. Configure proxy servers to forward the traffic to the correct instances
3. Use an Application Load Balancer with path-based routing rules to forward the traffic to the correct instances
4. Use an Application Load Balancer with host-based routing rules to forward the traffic to the correct instances

Q4

What is the primary reason why you should be using an elastic load balancer for a website with high activity?

1. ELBs help you scale servers easily without manual intervention
2. ELBs can distribute traffic equally to your backend targets to handle the incoming traffic load
3. ELBs help tighten security through the use of security groups
4. ELBs boost your website’s overall performance

Q5

After a year of development, the company’s 100-node high-performance computing (HPC) application is now ready to be deployed to an Amazon Elastic Kubernetes Service (Amazon EKS) cluster with Horizontal Pod Autoscaler (HPA). The application must be capable of receiving millions of UDP traffic per second from the public Internet while maintaining low latency.

Which of the following is the most operationally efficient solution that should be implemented to meet the above requirements?

1. Launch a Gateway Load Balancer and integrate an Elastic Fabric Adapter (EFA) to each Kubernetes pod deployed in the Amazon EKS cluster
2. Integrate the AWS Load Balancer Controller add-on to the EKS cluster. Launch a Network Load Balancer to load balance network traffic to individual Kubernetes pods.
3. Install the AWS Load Balancer Controller add-on to the EKS cluster and launch an Application Load Balancer to distribute the incoming traffic to the Kubernetes pods.
4. Set up the kube-proxy Amazon EKS add-on to the cluster and configure the Source Network Address Translation (SNAT) of the Kubernetes pods by setting the AWS_VPC_K8S_CNI_EXTERNALSNAT configuration to true.