Amazon Web Services (AWS) is the largest public cloud. This note covers the foundations: global infrastructure, the responsibility split, pricing, and design principles. See also The Big Three and Cloud Computing Services.
Services at a glance
| Service | Purpose |
|---|---|
| AWS Organizations | Multi-account management, consolidated billing, SCPs |
| AWS Control Tower | Set up a governed multi-account landing zone |
| AWS Cost Explorer / Budgets | Analyse and cap spend |
| AWS Pricing Calculator | Estimate costs |
| AWS Trusted Advisor | Checks for cost, performance, security, fault tolerance, limits |
| AWS Well-Architected Tool | Review workloads against the framework |
| AWS Health Dashboard | Service health and events affecting you |
| AWS CloudTrail | API call audit log |
| AWS Config | Resource configuration history and compliance |
| AWS Artifact | Compliance reports and agreements |
| AWS Support | Basic, Developer, Business, Enterprise On-Ramp, Enterprise |
Hands-on companion: sharifndlovu123/aws-playground, a lab of about 60 AWS services explored through the CLI, IaC and SDKs.
Why Cloud (Benefits)
- Trade upfront (CapEx) for variable (OpEx) expense
- Economies of scale lead to lower per-unit prices
- Stop guessing capacity: scale up and down on demand
- Speed and agility: resources in minutes
- Stop spending on running datacentres
- Go global in minutes
Global Infrastructure
| Term | Meaning |
|---|---|
| Region | A geographic area with multiple isolated Availability Zones (e.g. af-south-1 Cape Town, eu-west-1 Ireland) |
| Availability Zone (AZ) | One or more discrete datacentres with redundant power, networking and connectivity; each region has 3+ |
| Edge locations | Points of presence for CloudFront and Route 53 that cache and route near users |
| Regional edge caches | Larger caches between origin and edge |
| Local Zones | Extend a region closer to a city for single-digit ms latency |
| Wavelength | Compute at 5G network edges |
| Outposts | AWS hardware in your datacentre |
Choosing a region: compliance / data residency, latency to users, service availability, price.
Services are regional (EC2, VPC, S3 buckets live in a region), global (IAM, Route 53, CloudFront) or AZ-scoped (a subnet, an EC2 instance).
Shared Responsibility Model
- AWS: security of the cloud: hardware, facilities, global network, managed service software
- Customer: security in the cloud: data, IAM, OS patching (EC2), security groups, encryption choices, client-side config
| Example | Responsible |
|---|---|
| Patching the hypervisor | AWS |
| Patching the guest OS on EC2 | You |
| Patching the OS under RDS | AWS |
| S3 bucket policy and public access | You |
| Physical datacentre access | AWS |
| IAM users and MFA | You |
Responsibility shifts toward AWS as you move EC2 โ RDS โ Lambda โ SaaS.
Pricing Principles
- Pay as you go: on-demand per second / hour / GB
- Save when you commit: Savings Plans, Reserved Instances (1 or 3 year)
- Pay less as you use more: tiered pricing, volume discounts
- Free tier: always-free, 12-month free, and short trials
- Data transfer: inbound free, outbound to the internet charged, transfer between AZs/regions charged
EC2 purchase options
| Option | Best for |
|---|---|
| On-Demand | Short, unpredictable workloads |
| Savings Plans | Steady usage, flexible across instance families |
| Reserved Instances | Steady, specific instance config |
| Spot | Fault-tolerant, flexible jobs, up to ~90% off, can be reclaimed with 2 min warning |
| Dedicated Hosts / Instances | Licence rules, compliance |
| Capacity Reservations | Guarantee capacity in an AZ |
Cost tools
- Cost Explorer: visualise and forecast
- Budgets: alerts on actual or forecast spend
- Cost and Usage Report (CUR): most detailed data
- Cost allocation tags, Consolidated Billing (Organizations) for volume discounts
AWS Well-Architected Framework: Six Pillars
- Operational Excellence: run and monitor, improve continuously
- Security: protect data and systems, least privilege
- Reliability: recover from failure, scale to meet demand
- Performance Efficiency: use resources efficiently
- Cost Optimization: avoid unneeded cost
- Sustainability: minimise environmental impact
AWS Cloud Adoption Framework (CAF)
Perspectives: Business, People, Governance, Platform, Security, Operations. Guides organisations through cloud transformation.
Migration: the 7 Rs
Retire, Retain, Rehost (lift-and-shift), Relocate, Repurchase, Replatform, Refactor. Tools: Application Migration Service, Database Migration Service (DMS), Migration Hub, Snow Family (Snowcone, Snowball, Snowmobile) for offline transfer.
Ways to Interact with AWS
- Management Console: web GUI
- AWS CLI:
awscommands - SDKs: Python (Boto3), JavaScript, Java, .NET, Go...
- CloudShell: browser shell with CLI pre-installed
- Infrastructure as Code: CloudFormation, CDK
$ aws configure
$ aws sts get-caller-identity
$ aws ec2 describe-regions --output table
Related: 2. AWS IAM and Security, 3. AWS Compute, 4. AWS Storage and Databases, 5. AWS Networking, 6. AWS Developer Services and Serverless, 7. AWS Solutions Architecture