# What Are DevOps and Cloud? — The DevOps Loop, IaaS vs PaaS vs SaaS and the Big 3 Clouds

Source: https://skills101.in/learn/cloud-devops/devops-cloud-basics — Skills101 free tutorial (Cloud & DevOps). Updated 2026-11-09.

Writing code is only half of software. Someone has to **build it, test it, ship it to servers, keep it running at 3 a.m., and scale it when 10× more students show up on result day**. That's DevOps and cloud — and it's one of the fastest-growing job areas for freshers in India.

## 1. The problem DevOps solves

Old way: developers wrote code for months, then "threw it over the wall" to an operations team who had to deploy it. Releases were rare, scary and often broke things. Developers said *"it works on my machine"*; operations said *"your code broke production"*.

**DevOps** is a culture and set of practices where the **same team** builds, ships and runs the software, with heavy **automation**:

- small, frequent releases instead of big rare ones
- every change automatically built and tested (**CI** — continuous integration)
- deployment automated and repeatable (**CD** — continuous delivery/deployment)
- infrastructure written as code, versioned in Git
- monitoring, so problems are found by alerts, not angry users

## 2. The DevOps loop

![The DevOps infinity loop: plan, code, build, test on the development side, then release, deploy, operate, monitor on the operations side, with feedback flowing back into planning](https://skills101.in/learn-img/cloud-devops/devops-loop.svg)

| Stage | What happens | Typical tools |
|---|---|---|
| Plan | tickets, priorities | Jira, GitHub Issues |
| Code | write and review code | Git, GitHub, VS Code |
| Build | compile, package, build images | npm, Maven, Docker |
| Test | unit, integration, security tests | Jest, pytest, JUnit, Selenium |
| Release | version and approve | GitHub Releases, tags |
| Deploy | push to servers | GitHub Actions, Jenkins, Argo CD |
| Operate | run, scale, patch | Kubernetes, Linux, cloud consoles |
| Monitor | metrics, logs, alerts | Prometheus, Grafana, CloudWatch |

You'll learn each stage in this track.

## 3. What is "the cloud"?

The cloud is **renting computing over the internet instead of buying servers**. You pay for what you use, get machines in minutes, and can scale up or down.

| Before cloud | With cloud |
|---|---|
| buy servers weeks in advance | start a server in 1–2 minutes |
| pay upfront for peak capacity | pay per hour/second for what you use |
| your team fixes hardware | the provider handles hardware, power, cooling |
| one data centre | data centres worldwide |

## 4. Service models: who manages what?

![Cloud service models as a stack from hardware up to the app: on-premises you manage everything; with IaaS the provider runs hardware and virtualisation; with PaaS also the OS and runtime; with SaaS everything](https://skills101.in/learn-img/cloud-devops/service-models.svg)

| Model | You manage | Provider manages | Example | Pizza analogy |
|---|---|---|---|---|
| **On-premises** | everything | nothing | your own server room | make pizza at home |
| **IaaS** (Infrastructure) | OS, runtime, app, data | hardware, network, virtualisation | AWS EC2, Azure VMs | buy a frozen base, bake at home |
| **PaaS** (Platform) | your app and data | everything below | Heroku, Render, Google App Engine, Elastic Beanstalk | order delivery, eat at home |
| **SaaS** (Software) | just your usage/data | everything | Gmail, Zoom, Google Docs | eat at the restaurant |
| **Serverless / FaaS** | individual functions | servers, scaling | AWS Lambda, Azure Functions | pay per slice |

**Rule of thumb:** the higher up the stack, the less control but the less work. Startups often begin on PaaS and move to IaaS/Kubernetes when they need more control or lower cost at scale.

## 5. Regions and availability zones

- **Region**: a geographic area with several data centres — e.g. AWS **Mumbai (ap-south-1)** and **Hyderabad (ap-south-2)**, Azure Central India, Google Cloud **Mumbai** and **Delhi**.
- **Availability Zone (AZ)**: one or more separate data centres inside a region, with independent power and networking.

Run your app in **2+ AZs** so one data-centre failure doesn't take you down. Choose a region close to your users (lower latency) and one that meets data rules (some Indian sectors require data stored in India).

## 6. The Big 3 clouds side by side

| Need | AWS | Microsoft Azure | Google Cloud |
|---|---|---|---|
| Virtual machine | EC2 | Virtual Machines | Compute Engine |
| File/object storage | S3 | Blob Storage | Cloud Storage |
| Managed SQL database | RDS / Aurora | Azure SQL / Database for PostgreSQL | Cloud SQL |
| Serverless functions | Lambda | Functions | Cloud Run functions |
| Kubernetes | EKS | AKS | GKE |
| Identity & permissions | IAM | Entra ID + RBAC | IAM |
| Monitoring | CloudWatch | Azure Monitor | Cloud Monitoring |
| CDN | CloudFront | Front Door / CDN | Cloud CDN |

Concepts transfer between clouds — learn one deeply (AWS has the largest market share and the most job posts), and you can pick up the others quickly.

## 7. Case study: result day at a coaching platform

A test-prep site normally serves 2,000 students at a time. On result day, 60,000 arrive within an hour.

- **Without cloud/DevOps:** the single server maxes out, the site crashes, someone SSHes in and restarts it repeatedly.
- **With cloud/DevOps:** the app runs as containers behind a load balancer across two AZs; **autoscaling** adds servers when CPU goes above 60%; static pages are cached on a **CDN**; alerts warn the team before users notice; after the rush, extra servers shut down so the bill drops back.

You'll be able to explain (and build) every piece of that by the end of this track.

## 8. Careers

| Role | Focus |
|---|---|
| DevOps engineer | CI/CD, automation, infrastructure as code |
| Cloud engineer | designing and running cloud infrastructure |
| Site Reliability Engineer (SRE) | reliability, monitoring, incident response, performance |
| Platform engineer | internal tools so developers ship easily |

**Fresher path:** Linux + Git + one language (Python/Go/bash) → Docker → CI/CD → one cloud (AWS) → Kubernetes → Terraform. A cloud certification such as **AWS Certified Cloud Practitioner** or **Solutions Architect – Associate** helps you clear HR filters; a deployed project helps you clear interviews.

## Common mistakes

- Thinking DevOps is just "a tool" (Jenkins, Docker). It's a way of working; tools support it.
- Running everything in one availability zone.
- Leaving test servers running and getting a surprise bill — always set a **billing alert**.

## Quick recap

- **DevOps** = one team builds, ships and runs software, with automation (CI/CD, IaC, monitoring).
- **Cloud** = rent compute on demand. Models: **IaaS → PaaS → SaaS** (more managed going up), plus **serverless**.
- **Region** = geographic area; **AZ** = isolated data centre(s) inside it. Use 2+ AZs.
- AWS, Azure and GCP offer the same building blocks under different names.


Practice quiz and free certificate: https://skills101.in/learn/cloud-devops/devops-cloud-basics