The Physical Layer#

AWS operates the world’s largest cloud infrastructure — a global network of data centers organized into a hierarchy designed for high availability, low latency, and data sovereignty.

flowchart TD
    Global["AWS Global Infrastructure"]
    Global --> Regions["33 Regions<br>(geographic areas)"]
    Global --> Edge["450+ Edge Locations<br>(CDN endpoints)"]
    
    Regions --> AZs["105+ Availability Zones<br>(isolated data centers)"]
    Regions --> Local["Local Zones<br>(city extensions)"]
    Regions --> Wavelength["Wavelength Zones<br>(5G edge)"]

Regions#

A Region is a geographic area containing multiple isolated data centers. Each region is completely independent — a failure in one region doesn’t affect others.

Region Naming#

Choosing a Region#

flowchart TD
    Start["Choose a Region"] --> Compliance{"Data residency<br>requirements?"}
    Compliance -->|"Yes"| Restricted["Use required region<br>(e.g., EU for GDPR)"]
    Compliance -->|"No"| Latency{"Where are<br>your users?"}
    Latency --> Closest["Pick closest region"]
    Closest --> Services{"All services<br>available?"}
    Services -->|"No"| Alternative["Pick next closest<br>with all services"]
    Services -->|"Yes"| Cost{"Cost<br>sensitive?"}
    Cost -->|"Yes"| Compare["Compare pricing<br>(us-east-1 often cheapest)"]
    Cost -->|"No"| Done["✅ Use selected region"]

Key factors:

1. Compliance — GDPR requires EU data residency; some industries require specific countries
2. Latency — closer to users = faster response times
3. Service availability — new services launch in us-east-1 first
4. Cost — same service can cost 10-20% more in some regions

Availability Zones (AZs)#

Each region has 3-6 AZs. Each AZ is one or more physically separate data centers with independent power, cooling, and networking.

flowchart LR
    subgraph Region["eu-central-1 (Frankfurt)"]
        AZ1["AZ 1a<br>Data Center(s)"]
        AZ2["AZ 1b<br>Data Center(s)"]
        AZ3["AZ 1c<br>Data Center(s)"]
    end
    
    AZ1 <-->|"< 2ms latency<br>redundant fiber"| AZ2
    AZ2 <-->|"< 2ms latency<br>redundant fiber"| AZ3
    AZ1 <-->|"< 2ms latency<br>redundant fiber"| AZ3

AZ Properties#

• Physically isolated — separate buildings, power grids, flood zones
• Connected — high-bandwidth, low-latency private fiber between AZs
• Independent failure — a power outage in AZ-a doesn’t affect AZ-b
• Mapped differently per account — your eu-central-1a might be a different physical AZ than another account’s eu-central-1a

Multi-AZ Architecture#

Deploying across multiple AZs is the foundation of high availability on AWS:

flowchart TD
    Users["Users"] --> ALB["Application Load Balancer<br>(spans all AZs)"]
    
    subgraph AZ1["AZ 1a"]
        EC2a["EC2 Instance"]
        RDSp["RDS Primary"]
    end
    
    subgraph AZ2["AZ 1b"]
        EC2b["EC2 Instance"]
        RDSs["RDS Standby<br>(synchronous replica)"]
    end
    
    ALB --> EC2a
    ALB --> EC2b
    RDSp -.->|"Sync replication"| RDSs

If AZ-a fails: ALB routes all traffic to AZ-b, RDS fails over to standby.

Edge Locations#

Edge locations are AWS’s CDN (CloudFront) and DNS (Route 53) endpoints distributed globally — 450+ locations in 90+ cities.

AWS Backbone Network#

AWS operates its own global fiber network connecting all regions and edge locations. Traffic between regions travels over AWS’s private network, not the public internet.

Benefits:

• Lower latency — optimized routing
• Higher throughput — dedicated capacity
• Better security — traffic never touches public internet
• Consistent performance — no ISP congestion

Service Scope#

Services operate at different scopes:

flowchart TD
    subgraph Global["Global Services"]
        IAM["IAM"]
        R53["Route 53"]
        CF["CloudFront"]
    end
    
    subgraph Regional["Regional Services (eu-central-1)"]
        S3["S3"]
        Lambda["Lambda"]
        VPC["VPC"]
    end
    
    subgraph AZ["AZ-Scoped (eu-central-1a)"]
        EC2["EC2"]
        EBS["EBS Volume"]
        Subnet["Subnet"]
    end

Understanding scope matters for:

• Disaster recovery — regional services need cross-region replication
• High availability — AZ-scoped resources need multi-AZ deployment
• Cost — cross-region data transfer costs money

Shared Responsibility Model#

flowchart TD
    subgraph Customer["Customer Responsibility (Security IN the Cloud)"]
        CData["Customer Data"]
        CApp["Application & IAM"]
        COS["OS, Network, Firewall Config"]
        CEnc["Client-side Encryption"]
    end
    
    subgraph AWS["AWS Responsibility (Security OF the Cloud)"]
        AHW["Hardware / Global Infrastructure"]
        ANet["Networking"]
        AVirt["Virtualization (Hypervisor)"]
        APhys["Physical Security"]
    end

Key Takeaways#

1. Regions are independent — choose based on compliance, latency, services, and cost
2. AZs provide high availability — always deploy across 2+ AZs for production
3. Edge locations reduce latency — use CloudFront for static content
4. Know service scope — global vs regional vs AZ determines your HA strategy
5. Shared responsibility — AWS secures the infrastructure; you secure what you put on it
6. AWS backbone — inter-region traffic stays on AWS’s private network