Introduction to Cloud Computing

Almost every modern application runs on “the cloud” — but the term is vague. This page gives a clear definition, explains the three cloud service models, and points to beginner-friendly services that can be used immediately.

Target reader: Beginners who have heard about cloud computing but want a clear explanation. Estimated time: 15 min read Prerequisites: Servers and Clients (recommended)

What the Cloud Is

Cloud computing means accessing computing resources — servers, storage, databases, networking — over the Internet, on demand, from a provider that manages the physical hardware.

The alternative is on-premises (on-prem): buying and managing physical servers in a company’s own building or data center.

Aspect	On-Premises	Cloud
Hardware	Purchased, owned, maintained by the company	Owned and maintained by the cloud provider
Upfront cost	High (buy servers before using them)	None (pay for what is used)
Scaling	Requires ordering, shipping, installing new hardware	Provisioning new capacity takes seconds via a dashboard or API
Geographic reach	Limited to owned locations	Providers have data centers on every continent
Responsibility	Company manages hardware failures, power, cooling	Provider handles the physical layer

The shift to cloud computing has made it possible for a developer to launch a globally distributed application in an afternoon, without purchasing a single physical server.

Benefits of the Cloud

No upfront cost: Instead of spending money on hardware before a project proves itself, resources are rented by the hour or by usage. Early-stage projects can run cheaply and scale up only when needed.

Elasticity: Traffic spikes (a product launch, a news event) can be handled by provisioning more capacity on demand. When the spike passes, the extra capacity is released and billing stops.

Global distribution: Providers have data centers in regions around the world. Deploying an application close to users reduces latency. A startup in Tokyo can serve users in New York from a nearby data center without owning hardware there.

Managed services: Rather than installing and maintaining PostgreSQL on a virtual machine, a developer can use a managed database service. The provider handles updates, backups, and failover. This frees engineering time for product work.

IaaS, PaaS, and SaaS

Cloud services are grouped into three models that differ in how much the provider manages versus how much the customer manages.

A pizza analogy makes the distinction concrete:

                    Make it yourself   |  Buy the dough  |  Order delivery
                    (On-Premises)      |  (IaaS)         |  (PaaS / SaaS)
                                       |                 |
You manage:         Everything         |  App + Data     |  App only / Nothing
Provider manages:   Nothing            |  Hardware + OS  |  Hardware + OS + Runtime

IaaS — Infrastructure as a Service

The provider supplies virtual machines, networking, and storage. The customer installs the operating system, runtime, and application.

What the customer manages: Operating system, runtime, middleware, application code, data
What the provider manages: Physical hardware, virtualization, networking infrastructure
Examples: AWS EC2, Google Compute Engine, Azure Virtual Machines
Analogy: Renting a kitchen with equipment, but buying and cooking the ingredients

IaaS offers the most control and flexibility, but requires the most operational expertise.

PaaS — Platform as a Service

The provider manages the infrastructure and the runtime environment. The customer deploys application code.

What the customer manages: Application code and data
What the provider manages: Everything below the application layer (OS, runtime, scaling)
Examples: Heroku, Google App Engine, AWS Elastic Beanstalk, Vercel, Netlify
Analogy: Buying pizza dough — the base is provided, but the toppings are up to the customer

PaaS is where most web developers start. A git push is often all that is needed to deploy.

SaaS — Software as a Service

The provider delivers a fully working application over the Internet. The customer uses it directly, with no deployment work.

What the customer manages: Their own data within the application
What the provider manages: Everything, including the application itself
Examples: Gmail, Slack, Notion, Figma, GitHub
Analogy: Ordering pizza delivery — the whole pizza arrives ready to eat

SaaS requires no technical setup. As a developer, SaaS is what most of the daily tools are.

Side-by-Side Comparison

             Customer Responsibility
             High ─────────────────────────── Low
             |                                  |
         On-Prem    IaaS        PaaS       SaaS
             |        |           |           |
         Full        App +     App code    Just
         control     OS +      only        use it
                     Runtime

Major Cloud Providers

Provider	Strengths	Notable services
AWS (Amazon Web Services)	Largest market share; widest service catalog	EC2, S3, Lambda, RDS, CloudFront
GCP (Google Cloud Platform)	Strong in data/ML; Kubernetes originated here	Compute Engine, BigQuery, Vertex AI, GKE
Azure (Microsoft)	Deep enterprise and Microsoft ecosystem integration	Azure VMs, Azure SQL, Azure AI services
Vercel	Deployment of frontend apps and serverless functions	Next.js hosting, Edge Functions, Analytics
Netlify	Static site and JAMstack deployment	Netlify Functions, Forms, Identity
Supabase	Open-source Firebase alternative; PostgreSQL-backed	Database, Auth, Storage, Realtime
Cloudflare	CDN, DNS, edge computing, DDoS protection	Workers, Pages, R2, KV

AWS, GCP, and Azure are the three hyperscale providers. Vercel, Supabase, and Netlify are developer-focused platforms that make deployment and backend services accessible with minimal configuration.

Beginner-Friendly Services to Start With

For learning and small projects, these services have free tiers and require no credit card or infrastructure knowledge:

Service	What to use it for
Vercel	Deploying a frontend app (React, Next.js, Astro)
Netlify	Static site deployment with form handling
Supabase	PostgreSQL database + authentication + file storage
Railway	Deploy a backend service or database with one click
Cloudflare Pages	Fast static site hosting with global edge distribution

A common beginner stack: build a frontend with Astro or Next.js, deploy it to Vercel, and use Supabase for the database. All three have free tiers sufficient for learning and small production projects.

Summary

Cloud computing means renting computing resources over the Internet instead of owning hardware.
Benefits include no upfront cost, elastic scaling, global reach, and managed services.
IaaS provides raw infrastructure (VMs); the customer manages everything above hardware.
PaaS provides the runtime platform; the customer manages only application code.
SaaS is a fully managed application; the customer only uses it.
For beginners, Vercel, Supabase, and Netlify are the most accessible starting points.

FAQ

Q: Do I need a cloud account to learn web development?

A: Not immediately. Local development (running everything on a laptop) is sufficient for learning and prototyping. Cloud services become relevant when deploying an application for others to access, or when needing a database that persists between sessions.

Q: Which cloud provider should I learn first?

A: For frontend and full-stack development, Vercel and Supabase are the most beginner-friendly. For learning cloud infrastructure concepts more broadly, AWS has the widest documentation and tutorials, and free-tier offerings are sufficient for experimentation.

Q: What is serverless?

A: Serverless is a deployment model where the provider runs application code in response to events, automatically scales, and charges only for actual execution time. There are still servers — the developer just does not manage them. Vercel Functions, AWS Lambda, and Cloudflare Workers are serverless platforms. Serverless is a subcategory of PaaS.

Q: Is the cloud always cheaper than on-premises?

A: Not at scale. For large organizations with predictable, sustained workloads, owning hardware can be cheaper than renting. The cloud’s cost advantage is strongest for variable workloads, early-stage projects, and teams that want to avoid infrastructure operations. This is why large companies often use a mix of both.

Next steps: Engineering section — terminal, Git, and development environments

Link to this page (Japanese): クラウドコンピューティング入門