Cloud

title: AWS, Azure & GCP: The Big Three description: A newcomer's comparison of Amazon Web Services, Microsoft Azure, and Google Cloud — core service categories side by side, regions and availability zones, free tiers, the pay-as-you-go billing model with cost-control basics, and honest guidance on which to learn first.

AWS, Azure & GCP: The Big Three¶

Three providers dominate the public cloud market: Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP). They offer the same fundamental building blocks under different names, and the concepts transfer cleanly between them — so learning one teaches you most of the others.

Applies to

Cloud concepts that span all major public clouds. Examples reference the real services of AWS, Azure, and GCP. For what "the cloud" means at all, start with What Is the Cloud.

The same ideas, three vocabularies¶

Every major cloud rents the same primitives: machines to run code, places to store data, networks to connect them, and an identity system to control who can do what. The hard part for a newcomer is that each provider brands these differently. Once you map the vocabulary, a skill learned on one platform mostly carries over.

Core service categories¶

Category	AWS	Microsoft Azure	Google Cloud
Compute (virtual machines)	EC2 (Elastic Compute Cloud)	Azure Virtual Machines	Compute Engine
Serverless functions	Lambda	Azure Functions	Cloud Functions
Object storage	S3 (Simple Storage Service)	Blob Storage	Cloud Storage
Block storage (VM disks)	EBS (Elastic Block Store)	Managed Disks	Persistent Disk
Managed relational database	RDS (Relational Database Service)	Azure SQL Database	Cloud SQL
Virtual network	VPC (Virtual Private Cloud)	VNet (Virtual Network)	VPC
Identity & access	IAM (Identity and Access Management)	Microsoft Entra ID (formerly Azure AD)	Cloud IAM

The names are branding; the concepts are not

A virtual machine is a virtual machine whether it is called an EC2 instance or a Compute Engine VM. The deeper ideas — virtualization, networking, storage durability, identity — are shared across every provider. See Virtualization and Hypervisors and Cloud Storage and Networking.

Regions and availability zones¶

All three clouds organize their hardware the same way, which is one of the clearest examples of transferable concepts.

Term	What it is	Why it matters
Region	A geographic area (e.g. `us-east-1`, West Europe, `asia-south1`) containing multiple data centers	Choose one close to your users for low latency, and mind data-residency laws
Availability Zone (AZ)	One or more isolated data centers within a region, with independent power, cooling, and networking	Spreading resources across AZs survives the failure of a single data center
Edge / PoP	Smaller sites for content delivery and DNS, closer to end users	Used by CDNs (CloudFront, Azure CDN, Cloud CDN) and DNS, not for running general workloads

The terminology is nearly identical across providers:

Concept	AWS	Azure	GCP
Geographic area	Region	Region	Region
Isolated data center group	Availability Zone	Availability Zone	Zone

Design for failure across zones

A region can have an outage; an availability zone can fail on its own. Production systems place redundant copies across multiple AZs in one region, and critical systems also replicate to a second region for disaster recovery.

The pay-as-you-go billing model¶

Public cloud is utility billing: you pay for what you consume, usually by the second or hour for compute and by the gigabyte-month for storage, with no upfront hardware purchase. This is the core economic difference from owning servers.

Common ways to pay for compute:

Pricing model	How it works	Best for
On-demand / pay-as-you-go	Full price, billed per second or hour, no commitment	Spiky or unpredictable workloads, learning, testing
Reserved / committed use	Commit to 1–3 years for a large discount (AWS Savings Plans/Reserved Instances, Azure Reservations, GCP Committed Use Discounts)	Steady, predictable baseline workloads
Spot / Spot VMs / Preemptible	Deep discount on spare capacity that the provider can reclaim at any time	Fault-tolerant, interruptible batch jobs

The bill grows quietly

The number one beginner mistake is leaving resources running — a forgotten VM, an unattached disk, an idle managed database — and being surprised by the bill. Cloud resources cost money until you explicitly delete them.

Cost-control basics¶

A few habits keep spending under control on any provider:

Set a budget and alerts. Use AWS Budgets, Azure Cost Management + Budgets, or GCP Budgets & Alerts to email you when spend crosses a threshold.
Right-size. Match instance size to actual CPU/memory usage; most workloads are over-provisioned. All three offer recommendations (AWS Compute Optimizer, Azure Advisor, GCP Recommender).
Turn things off. Stop dev/test VMs outside working hours; delete unused disks, snapshots, and idle load balancers.
Tag everything. Apply labels/tags (e.g. project, owner, env) so cost reports can attribute spend.
Use the calculators. The AWS Pricing Calculator, Azure Pricing Calculator, and Google Cloud Pricing Calculator estimate cost before you deploy.

Free tiers — where to practice¶

All three let you learn at little or no cost, but the offers differ. Always confirm current details on the provider's pricing page, since terms change.

Provider	Free offer (typical)
AWS	A mix of 12-months-free (e.g. a small EC2 instance, RDS hours), always-free services (e.g. Lambda monthly free requests, DynamoDB), and short-term trials
Azure	A starting credit for the first 30 days, 12 months free on popular services, plus a set of always-free services
GCP	A starting credit (90 days) plus an Always Free tier (e.g. a small Compute Engine instance, Cloud Storage allowance) within usage limits

Free tiers are not free if you exceed the limits

Going over a free-tier limit, or using a service that is not in the free tier, bills at normal rates. Set a budget alert before you start experimenting, and prefer free-tier-eligible sizes.

Which should you learn first?¶

Honest answer: pick one and go deep — the concepts transfer, and employers value depth in any of the three over a shallow tour of all of them.

If…	Consider
You want the largest job market and the broadest service catalog	AWS — the market-share leader and the most common certification path
Your organization already runs Microsoft (Windows Server, Microsoft 365, Active Directory)	Azure — tight integration with Entra ID and Windows ecosystems
You work with data, analytics, Kubernetes, or AI/ML pipelines	GCP — strong data tooling, and Google originated Kubernetes

Concepts first, provider second

Learn the ideas — virtual machines, object vs block storage, VPCs and subnets, IAM, regions and AZs — and you can move between clouds. The first cloud you learn is the hardest; the second is mostly vocabulary. When you are ready to validate the skill, see Certifications for the entry-level paths (AWS Certified Cloud Practitioner, Azure Fundamentals AZ-900, Google Cloud Digital Leader).

Verify your work¶

Confirm you have internalized the core mappings — no cloud account required.

Without looking, name each provider's object storage service. (Answer: S3, Blob Storage, Cloud Storage.)
Explain the difference between a region and an availability zone, and why production spans multiple AZs.
Describe two concrete steps to avoid a surprise bill (e.g. set a budget alert; delete idle resources).
State which provider you would pick first and why, in one sentence.

If you have an account, also confirm a budget alert is configured before launching anything:

AWS    : Billing → Budgets → Create budget (set an amount + email alert)
Azure  : Cost Management → Budgets → Add
GCP    : Billing → Budgets & alerts → Create budget

Summary¶

AWS, Azure, and GCP offer the same primitives under different names — compute, serverless, object/block storage, managed databases, virtual networks, and identity.
The core mappings: EC2/Azure VMs/Compute Engine; Lambda/Functions/Cloud Functions; S3/Blob/Cloud Storage; EBS/Managed Disks/Persistent Disk; RDS/Azure SQL/Cloud SQL; VPC/VNet/VPC; IAM/Entra ID/Cloud IAM.
All three use regions (geographic areas) containing availability zones (isolated data centers); spread workloads across AZs for resilience.
Billing is pay-as-you-go, with reserved/committed and spot/preemptible options for discounts; control cost with budgets, right-sizing, tagging, and turning resources off.
Each has a free tier for learning, but exceeding limits bills at normal rates — set a budget alert first.
Pick one cloud and learn it deeply; the concepts transfer to the others.