Auto scaling

• All the cloud support elasticity i.e. increasing as well as decreasing.
• Scaling is of two forms
  • Vertical scaling:
    • increasing or decreasing the size of servers
    • scale up and down
  • Horizontal Scaling
    • increasing or decreasing the number of servers
    • Scale in and Scale out
• To Scale the application running on a vm we should have an Image with application pre-installed.
• The next factor is define Scaling category
  • Deterministic: We know the peak hours and scale accordingly
  • Un Deterministic: We donot know the peak hours and have to rely on some metrics to scale.
  • Fixed: Always the fixed number of servers
• Cloud Providers have started offering predictive scaling this is a ML model which learns your scaling needs
• High Availability: Ability of system to have minimal or zero downtime despite traffic peaks
• Fault Tolerance: Ability of System to continue being operational despite errors
• Application types depending on customer geo locations
  • users are global
  • users are regional or local.
• Impact of this on Application deployments
  • Multi-Regional
  • Multi-Zonal
• Auto scaling is supported in
  • AWS using Auto scaling groups (ASG)
  • Azure using Virtual Machine Scale Sets (VMSS)

Steps To Acheive

• How to create Images in AWS and Azure
• How to configure Auto Scaling

Images

• AWS uses a term called as AMI (Amazon Machine Image)
• Azure uses a term called as VM Image
  • Generalize image: User can set credentials
  • Specialized Images: Users can’t set credentials

AMI

• To create an AMI, We need to create an ec2 instance, install necessary softwares and then create image
• Create Image takes snapshot of a root volume (disk) and creates an AMI with a unique id in the same region where ec2 is running
• If we need this image in other regions, we can copy the image to multiple regions.
• Note: image copied to other region will get a new AMI id.

Create an AMI (Manual)

• AMI 1: Lets create an AMI for troweld website with name foundry
• AMI 2: Lets create an AMI for carpernter website with name carpernter
• Note: We will be installing an extra software called as stress to create artifical stress

Foundry

• Lets create an ec2 instance with ubuntu 24.04

#!/bin/bash
sudo apt update 
sudo apt install nginx unzip stress -y
cd /tmp
wget https://www.free-css.com/assets/files/free-css-templates/download/page294/troweld.zip
unzip troweld.zip
mv troweld-html/ foundry/
sudo mv foundry/ /var/www/html/

• Now lets create an AMI
  
  
  
  
  
• Now lets launch an ec2 instance with the AMI created above
  
  
• AMIs are of the following type by ownership
  • Amazon offered official AMIs: These are images of popular OS offered by Amazon for customers, some of the Amazon offered images might include hourly costs
  • MarketPlace Images: These are created by third party (Companies) with specific images which might include hourly costs
    
  • Private Images: These are created by users for their internal purposes. This will have storage costs. These images can be shared with specific AWS accounts
  • Community Images: These are created by users and shared to other users for free of cost.

Carpenter

• Now lets create an AMI by following the steps as done in previous case (refer classroom recording if necessary)

#!/bin/bash
sudo apt update 
sudo apt install nginx unzip stress -y
cd /tmp
wget https://www.free-css.com/assets/files/free-css-templates/download/page294/woody.zip
unzip woody.zip
mv carpenter-website-template/ carpenter/
sudo mv carpenter/ /var/www/html/

Auto Scaling

• For Creating an Autoscaling group, we need to create a Launch Template where we specify the redundant selection paramters such as AMI, security group, instance type etc
• Also while creating asg we need to select more than one AZ
  

Creating a Launch Template

• Watch clasroom recording.
  

Fixed Scaling with ASG

• For selection options watch classroom recording