Open this lesson in your favourite AI. It'll walk you through the why, explain the demo, and quiz you on the try-it list.
Every conversation about cloud architecture eventually hinges on which layer you are buying. If you cannot tell IaaS from PaaS from SaaS you will end up paying for managed services you did not need, or running raw VMs when a managed runtime would have shipped you in a week. Knowing where each provider draws the line is the first filter you apply when picking a service.
List a representative service from each layer on AWS and Azure to see the boundary in concrete terms.
Use these three in order. Each builds on the one before.
In one paragraph, explain IaaS vs PaaS vs SaaS like I'm new to it, using a concrete example service for each.
Walk me through how a request to a PaaS service like AWS Elastic Beanstalk actually flows step by step from the load balancer down to my code, and identify which layers I am responsible for vs the provider.
Given a startup that needs to ship a CRUD API in two weeks but expects 100x traffic growth in 18 months, how would you decide between running raw EC2 (IaaS), App Runner (PaaS), and a fully managed BaaS like Supabase, and why?
# IaaS — raw compute, you own the OS
aws ec2 describe-instance-types --instance-types t3.micro \
--query 'InstanceTypes[0].{vcpu:VCpuInfo.DefaultVCpus,memMiB:MemoryInfo.SizeInMiB}'
# PaaS — managed runtime, you push code
aws elasticbeanstalk list-available-solution-stacks \
--query 'SolutionStacks[?contains(@,`Node.js 20`)]'
# SaaS-like — fully managed product surface
aws workmail list-organizations --region us-east-1 \
--query 'OrganizationSummaries[*].[Alias,State]' --output table