A starting point for anyone new to the cloud. It covers what the cloud is and why it is used, how IaaS, PaaS, SaaS, and serverless differ, a map of the core services every cloud has in common, a comparison of AWS, GCP, Azure, NCP, and Cloudflare, the basics of cost and security, and a step-by-step learning roadmap.
Beginner
Updated Oct 1, 2026
~8 min read
12 sections
1 code examples
Learning Paths that include this guide
This guide is one step in the paths below. Study it together with the steps before and after.
Understanding Cloud Intro & Roadmap as a real-world workflow
A starting point for anyone new to the cloud. It covers what the cloud is and why it is used, how IaaS, PaaS, SaaS, and serverless differ, a map of the core services every cloud has in common, a comparison of AWS, GCP, Azure, NCP, and Cloudflare, the basics of cost and security, and a step-by-step learning roadmap. Rather than listing concepts, this guide is organized so you can follow the content in the order you have to make decisions on a real project.
Key perspective
Infrastructure / operations
Rather than memorizing install commands, understand the order in which traffic, runtime, observability, and incident response connect.
The big picture of cloud conceptsUnderstand the service modelsCompare cloud providersCost and security basicsPreparing for cloud certifications
Architecture diagrams
To make what you read stick, it helps to capture the flow as a picture first. The two diagrams below are reference maps you can keep coming back to while learning Cloud Intro & Roadmap.
Learning flow
Rendering diagram…
Architecture view
Rendering diagram…
What the cloud is: a data center you rent
When you first open Cloud Intro & Roadmap, the big picture matters more than individual commands. This section starts with the problems the concepts ahead were created to solve.
It used to take weeks to run a service: buying servers, installing them in a data center, and connecting the network. If traffic was higher than expected there were not enough servers, and if it was lower, expensive hardware sat idle.
The cloud is a way of renting all of that over the internet, as much as you need for as long as you need. With a few clicks (or a few lines of code) you create a server in minutes, and you delete it to stop paying when you no longer need it.
Aspect
On-premises (your own servers)
Cloud
Upfront cost
Buying servers and equipment (a large upfront investment)
None — pay for what you use
Lead time
Weeks to months
Minutes
Scaling
Need to buy more hardware
Scale up and down immediately
Management responsibility
Everything from the hardware up
The provider handles hardware and physical security
Global deployment
Secure overseas data centers yourself
Just pick a region
Service models: how much do you manage yourself
Service models: how much do you manage yourself is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
Cloud services are divided by "how much you have to manage". The management burden shrinks as you go down the list, but so does fine-grained control.
Model
You manage
The provider manages
Examples
IaaS
OS, runtime, app, data
Hardware, network, virtualization
AWS EC2, GCP Compute Engine, NCP Server
CaaS (containers)
Container images, configuration
The cluster management plane
EKS, GKE, AKS, Cloud Run
PaaS
App code, data
OS, runtime, scaling
Elastic Beanstalk, App Engine, Azure App Service
Serverless (FaaS)
Function code
Servers, scaling, and availability entirely
AWS Lambda, Cloudflare Workers
SaaS
Just use it
Everything
Gmail, Slack, Notion
Map of the common core services: different names, same concepts
Map of the common core services: different names, same concepts is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
Service names differ from provider to provider, but the core concepts are nearly identical. Learn one cloud properly and you can adapt to another quickly, at roughly the level of "swapping the names".
Category
Concept
AWS
GCP
Azure
Compute
Virtual servers
EC2
Compute Engine
Virtual Machines
Containers
Managed Kubernetes
EKS
GKE
AKS
Serverless
Function execution
Lambda
Cloud Functions / Cloud Run
Azure Functions
Storage
Object storage
S3
Cloud Storage
Blob Storage
Database
Managed relational DB
RDS / Aurora
Cloud SQL
Azure SQL / Database for PostgreSQL
Networking
Virtual private network
VPC
VPC
VNet
Security
Access management
IAM
IAM
Entra ID + RBAC
Observability
Monitoring and logs
CloudWatch
Cloud Monitoring / Logging
Azure Monitor
Comparing cloud providers
Comparing cloud providers is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
AWS is the recommended first cloud. It has the most learning material and hiring demand, and the concepts you learn there carry straight over to other clouds. This site (aidevops.us) runs on Cloudflare.
Provider
Strengths
Typically chosen when
Guide
AWS
The largest market share, the most services, vast documentation
Cloud networking basics: VPC and subnets is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
The first wall you hit in the cloud is networking. Most cases of "I created a server but cannot connect" are network configuration problems. A VPC is your own isolated network inside the cloud, and you divide it into subnets by purpose.
Component
Role
Typical placement
Public subnet
Can talk to the internet directly
Load balancers, Bastion (jump server)
Private subnet
Not reachable directly from outside
App servers, DB
Internet gateway
The path between the VPC ↔ the internet
Attached to public subnets
NAT gateway
The way out for private servers (inbound stays blocked)
For installing packages and calling external APIs
Security group
A per-instance firewall (allow rules)
The DB allows 3306/5432 only from the app servers
Your first deployment: publish a static site from the CLI
Here you look at Your first deployment: publish a static site from the CLI alongside real code. Rather than copying the example as-is, read it with an eye on the inputs, the outputs, and the parts most likely to change.
The cloud can be driven from the web console, but in real work the default is the CLI and code. Below is an example of publishing a static site to S3 with the AWS CLI. Do the same task in the console once and then repeat it with the CLI to feel the difference.
deploy-static.shBASH
# Prerequisite: set up credentials with aws configure (never use root account keys)BUCKET=my-first-site-20260924# 1. Create the bucket (in the region you use)aws s3 mb s3://$BUCKET --region us-east-1# 2. Upload the build output (syncs only changed files)aws s3 sync ./dist s3://$BUCKET --delete# 3. Configure static website hostingaws s3 website s3://$BUCKET --index-document index.html --error-document 404.html# Clean up when the lab is done — make a habit of deleting unused resources right away!# aws s3 rb s3://$BUCKET --force
Cost and security basics: what to take care of from day one
Cost and security basics: what to take care of from day one is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
The two scariest things for cloud newcomers are unexpected bills and security incidents. The cloud follows a "shared responsibility model". The provider is responsible for the data centers and hardware, and the configuration, data, and permissions you put on top are your responsibility.
Item
What to do
Why
Root account
Enable MFA, then do not use it for everyday work
A stolen root = the whole account stolen
IAM
Least-privilege accounts per person and service
Limits the blast radius of a leak
Access keys
Never in code or on GitHub; temporary credentials where possible
Leaked keys are used to spin up mining servers → a huge bill
Budget alerts
Set a monthly budget and usage alerts
Catch billing accidents early
Resource cleanup
Delete instances, NAT, and load balancers after a lab
They bill by the hour just for existing
Tags
Tag project, owner, and environment
Track cost and identify what to clean up
Before you start
Before you start is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
In the end, the cloud is "Linux servers and networks in someone else's data center". The firmer your Linux and networking fundamentals, the faster you learn the cloud.
Step-by-step roadmap: from the first server to cloud architecture
Step-by-step roadmap: from the first server to cloud architecture is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
8-week study plan: 1-2 hours a day is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
The "all resources deleted" step in the last week is an important exercise too. If you built everything with Terraform, a single terraform destroy cleans it all up — the moment you feel the value of IaC.
Week
Topic
Deliverable
Week 1
Account security + IAM + budget alerts
A safely configured cloud account
Week 2
Virtual servers + security groups
A web server you can SSH into
Week 3
Object storage + CDN
A static site deployment + custom domain
Week 4
VPC design
A 2-tier public/private subnet layout
Week 5
Managed DB + load balancer
A 3-tier architecture: 2 app servers + DB
Week 6
Container services / serverless
A container app deployment + a serverless function
Week 7
Terraform
Weeks 1-6 reproduced as code
Week 8
Monitoring + cost analysis + cleanup
Dashboard + cost report + all resources deleted
Common mistakes of cloud beginners
Common mistakes of cloud beginners is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
Most cloud incidents come not from technically hard problems but from configuration mistakes. The habit of making a checklist and going through it every time is the most powerful security tool.
Mistake
Problem
Do this instead
Doing everyday work as the root account
Total damage if the account is hijacked
MFA + IAM users/roles
Committing access keys to GitHub
Harvested automatically within minutes and abused
Environment variables or a secret manager, revoke the key immediately
Opening a security group to 0.0.0.0/0
Exposed to brute-force attacks
Allow only the ports and IPs needed
Leaving lab resources running
A huge bill at the end of the month
Delete right after the lab, budget alerts
Building only by clicking in the console
Cannot be reproduced, reviewed, or recovered
Record it with the CLI and IaC
Putting everything in a single availability zone
Everything goes down if the data center fails
A multi-AZ setup
Working without checking the region
Resources cannot be found, or latency goes up
Check the region before you start (e.g. us-east-1)
Key terms
Key terms is a point where the options diverge. Using the table to compare what each approach is for and how it differs in operation makes later decisions much easier.
Term
One-line explanation
Region
The geographic area where cloud data centers are located (e.g. N. Virginia)
Availability zone (AZ)
A group of physically separate data centers within one region
VPC
A logically isolated private network you create inside the cloud
CIDR
Notation for IP address ranges (10.0.0.0/16 = about 65,000 addresses)
IAM
Access management that defines who can do what to which resources
Security group
Virtual firewall rules applied to instances
Object storage
A service that stores files without limit as key-object pairs (S3 and others)
Serverless
An execution model billed per code execution, with no servers to manage
Shared responsibility model
The principle that the provider is responsible for the cloud itself and the user for the configuration and data in it
Free tier
A benefit that gives new sign-ups a certain period or amount of usage for free
Cloud Intro & Roadmap: frequently asked questions
Should I learn AWS, Azure, or GCP first?
Learning the cloud used by the company you want to join, or by your current organization, is best. If nothing is decided, many people start with AWS because it has the most users and learning material. Virtual servers, networking, storage, and access management are similar across clouds under different names, so learning one makes the others easy to pick up.
How do IaaS, PaaS, and SaaS differ?
With IaaS you rent infrastructure such as virtual servers and networks and manage everything from the operating system up. With PaaS you upload code and the platform manages the runtime. With SaaS you use finished software as it is. The higher you go, the less there is to manage and the less fine control you have.
How do I keep the cloud bill from being higher than expected?
Set a budget alert as soon as you create the account, and delete resources when a lab is finished. Common causes are servers left running and forgotten, detached disks and static IPs, outbound data transfer, and managed databases. Tagging resources makes it easy to trace where costs come from.