Understanding AWS and its Global Infrastructure

Table of Contents

1. Module 1: What Is AWS?
   • 1.1 Cloud Computing
   • 1.2 AWS History
   • 1.3 How the World Uses AWS
2. Module 2: AWS Global Infrastructure
   • 2.1 Regions and Availability Zones
   • 2.2 Local Zones, Wavelength Zones and Edge Locations
   • 2.3 AWS Outposts and Dedicated Local Zones
   • 2.4 AWS Global Backbone Network
   • 2.5 How Enterprises Provision Resources
   • 2.6 An Ever-Expanding Infrastructure
3. Module 3: Connectivity and Distribution Services
   • 3.1 Amazon CloudFront
   • 3.2 AWS Global Accelerator
   • 3.3 Amazon Route 53
   • 3.4 Connectivity Services Comparison Table
4. Module 4: High Availability Architectures
   • 4.1 Multi-AZ Deployment Patterns
   • 4.2 Multi-Region Deployment Patterns
   • 4.3 Disaster Recovery
5. Quick Reference: AWS Key Numbers (2025–2026)

Module 1: What Is AWS?

1.1 Cloud Computing

What Is the Cloud?

Cloud computing is a technology designed to solve three major challenges related to distributed applications:

Problems Solved by the Cloud

Scalability: Imagine a flower sales website. During special events like Mother’s Day, traffic increases considerably. With traditional infrastructure, you’d need to install additional servers that would remain unused afterward — generating significant maintenance costs. The cloud allows dynamically adjusting resources, paying only for what is consumed.

Performance: With a single data center, users on the other side of the world experience high latency. Data travels long distances, degrading the user experience. AWS solves this problem with its 39 Regions and 750+ Points of Presence (POPs) distributed across all continents.

Fault Tolerance: A single data center creates a single point of failure. In case of natural disaster or outage, the entire application can go down. The cloud distributes resources across multiple physically separate Availability Zones (AZs) to avoid this scenario.

Cloud Computing Definition

Cloud Computing = On-demand access to computing resources
                  (compute, storage, network, databases, etc.)
                  via the Internet, with pay-as-you-go pricing.

Deployment Models

flowchart LR
    A[Traditional Infrastructure\nOn-Premises] -->|Partial migration| B[Hybrid Cloud]
    B -->|Full migration| C[Public Cloud\nAWS]
    A -->|Private controlled\ninfrastructure| D[Private Cloud]

Cloud Advantages

1.2 AWS History

AWS Development Timeline

timeline
    title AWS Evolution — From bookstore to global cloud leader
    2000 : Amazon adopts service-oriented architecture
         : Decomposition into independent services
         : Foundation of microservices
    2003 : Identification of key cloud services
         : Focus on compute, storage and databases
         : Vision for external cloud services
    2006 : Launch of first AWS services
         : Amazon S3 — object storage
         : Amazon EC2 — cloud compute
         : Beginning of the cloud revolution
    2008–2012 : Exponential growth
              : Expansion to Europe and Asia
              : Launch of Route 53, VPC, RDS
    2013–2018 : Platform maturity
              : Machine Learning, Serverless Lambda
              : Presence in 20+ Regions
    2019–2025 : Global leadership
              : Generative AI, Quantum Computing
              : 39 Regions, 123 AZs

Origin: Amazon as the Starting Point

At the turn of the millennium, Amazon — originally an online bookstore — launched new products and opened its platform to third-party sellers via Amazon Marketplace. This expansion generated an explosion of data (product images, catalogs) and posed major architectural challenges.

Key decision (2000): Instead of managing Amazon as a single large monolithic application, Amazon began defining services developed independently of each other. This principle is the foundation of the microservices architecture we know today.

Excess capacity: Data centers built by Amazon for its own needs had considerable excess capacity. At some point, Amazon realized this capacity could be offered to third parties — this is how the AWS concept was born.

First AWS Services (2006)

Amazon S3 (Simple Storage Service)

• Object storage service with virtually unlimited capacity
• Born from Amazon’s need to store millions of product images
• Developers no longer need to worry about storage capacity or data resilience
• Ideal for serving objects to globally distributed applications
• Durability: 99.999999999% (11 nines) — Standard class

Amazon EC2 (Elastic Compute Cloud)

• AWS’s first cloud compute service
• Allows provisioning virtual servers (instances) in the cloud
• Network of public and private servers, fully configurable via VPC
• Pay by the hour or second depending on instance type
• Instance families: General Purpose, Compute, Memory, Storage, GPU, Arm

1.3 How the World Uses AWS

mindmap
  root((AWS))
    Big Data
      Data Lakes on S3
      Redshift Data Warehouses
      Glue ETL Pipelines
      Athena and QuickSight Analytics
    Backup and Retention
      Application backup
      Glacier long-term archiving
      Disaster Recovery
    Object Storage
      File distribution
      CDN via CloudFront
      Public data sharing
    Internet of Things
      Industrial data collection
      IoT Core automation
      Greengrass device management
    IT Infrastructure
      IAM identity management
      WorkSpaces Virtual Desktops
      Managed Active Directory
    AI and Machine Learning
      Bedrock pre-trained models
      SageMaker training
      Rekognition vision
      Comprehend NLP
    Special Services
      Quantum Computing Braket
      GPS and mapping Location Service
      IVS media streaming
      Ground Station satellite simulation

Module 2: AWS Global Infrastructure

2.1 Regions and Availability Zones

Infrastructure Hierarchy

flowchart TB
    AWS[AWS Global Infrastructure] --> R1[Region\nus-east-1\nUS East — N. Virginia]
    AWS --> R2[Region\neu-west-1\nEU West — Ireland]
    AWS --> R3[Region\nap-southeast-1\nAsia Pacific — Singapore]
    AWS --> RN[... 36 other Regions\n39 total]

    R1 --> AZ1[Availability Zone\nus-east-1a\nData centers A]
    R1 --> AZ2[Availability Zone\nus-east-1b\nData centers B]
    R1 --> AZ3[Availability Zone\nus-east-1c\nData centers C]
    R1 --> AZ4[... up to 7 AZs\nper Region]

    AZ1 --> DC1[Data Center 1]
    AZ1 --> DC2[Data Center 2]
    AZ1 --> DC3[Data Center N]

Key Numbers (2025–2026)

What Is a Region?

An AWS Region is a physical geographic location in the world where AWS groups its data centers. Each Region is:

• Independent: resources in Region A are completely separate from those in Region B
• Composed of at least 3 Availability Zones
• Identified by a logical code (e.g., us-east-1, eu-west-3, ap-southeast-1)
• Selected based on user proximity, data sovereignty requirements, and costs

Region selection criteria:

What Is an Availability Zone (AZ)?

An Availability Zone is a cluster of one or more data centers physically separated within the same Region. Characteristics:

• Physical separation: each AZ is several kilometers from others, protecting against local natural disasters
• Independent power: redundant power supply, cooling, and networking
• Ultra-fast connectivity: private network with very low latency between AZs in the same Region (< 2 ms)
• High availability: deploying across 3 AZs guarantees availability even if an entire AZ goes down

flowchart TB
    subgraph Region["Region: us-east-1 — N. Virginia — 7 AZs"]
        subgraph AZa["AZ: us-east-1a"]
            DC_A1[Data Center A1]
            DC_A2[Data Center A2]
        end
        subgraph AZb["AZ: us-east-1b"]
            DC_B1[Data Center B1]
            DC_B2[Data Center B2]
        end
        subgraph AZc["AZ: us-east-1c"]
            DC_C1[Data Center C1]
        end
        AZa <-->|Redundant private network\n< 2 ms| AZb
        AZb <-->|Redundant private network| AZc
        AZa <-->|Redundant private network| DC_A2
    end

2.2 Local Zones, Wavelength Zones and Edge Locations

AWS Local Zones

Local Zones are AWS Region extensions located geographically very close to large metropolitan areas. They allow running applications requiring latency in the single-digit milliseconds for local users.

• 43 Local Zones deployed (2025)
• Use cases: real-time gaming, live video streaming, AR/VR, edge computing
• Available services: EC2, EBS, ELB, VPC (subset of full services)
• Each Local Zone is linked to a parent Region for services not locally available

AWS Wavelength Zones

Wavelength Zones embed AWS compute and storage services directly in 5G carrier data centers, achieving latencies below 10 milliseconds for mobile devices.

• 33 Wavelength Zones (2025)
• Partners: Verizon, Vodafone, KDDI, SK Telecom, T-Mobile
• Use cases: autonomous vehicles, industrial robotics, mobile cloud gaming, 5G AR/VR

CloudFront Edge Locations

Edge Locations are CloudFront network points of presence distributed globally for caching content as close as possible to end users.

Three layers of CloudFront infrastructure:

2.3 AWS Outposts and Dedicated Local Zones

AWS Outposts

AWS Outposts brings AWS infrastructure physically into the customer’s data center, enabling a truly hybrid cloud experience with the same AWS APIs, tools and services.

• Use cases: data that cannot leave the site (compliance), applications requiring ultra-low latency to local equipment
• Formats: full 42U racks, or Outposts servers (1U/2U format)
• Management: fully managed by AWS (installation, maintenance, updates)

AWS Dedicated Local Zones

Dedicated Local Zones are AWS infrastructure instances built for a specific customer or group of customers, to meet digital sovereignty and strict regulatory compliance requirements.

• Access restricted to a specific organization (government, defense, banking sector)
• Infrastructure physically separated from public AWS infrastructure

2.4 AWS Global Backbone Network

AWS owns one of the world’s most extensive private networks, designed to route traffic between Regions and Edge Locations with the minimum number of hops on the public Internet.

Backbone characteristics:

Benefits of intra-AWS communication: When AWS services communicate with each other (EC2 to S3, for example), traffic stays on the private AWS network — not on the public Internet. This ensures performance, security and reliability.

2.5 How Enterprises Provision Resources

Structure: Users → Accounts → Organizations

flowchart TB
    Org[AWS Organization\n1 consolidated bill] --> MA[Management Account\nCentralized billing]
    Org --> A1[AWS Account\nProduction]
    Org --> A2[AWS Account\nDevelopment]
    Org --> A3[AWS Account\nData / Analytics]
    Org --> A4[AWS Account\nSecurity / Audit]

    A1 --> U1[User: Dev Team]
    A1 --> U2[User: Ops Team]
    A1 --> R1[Resources: EC2, RDS\nS3, Lambda...]
    R1 --> REG1[Region us-east-1]
    R1 --> REG2[Region eu-west-1]

Core Concepts

2.6 An Ever-Expanding Infrastructure

Scale economy principle: As AWS acquires equipment and improves its global infrastructure, costs decrease. AWS passes on these savings through regular price cuts — historically dozens of price reductions since 2006.

Module 3: Connectivity and Distribution Services

3.1 Amazon CloudFront

Amazon CloudFront is AWS’s CDN (Content Delivery Network). It distributes content (static, dynamic, streaming) to end users from the nearest Edge Locations.

CloudFront Key Features

CloudFront Functions vs Lambda@Edge:

3.2 AWS Global Accelerator

AWS Global Accelerator improves application availability and performance by routing user traffic through the AWS backbone (private network) rather than the public Internet.

flowchart LR
    User[User\nTokyo] -->|Public Internet\nunpredictable paths| GA[AWS Global Accelerator\n2 static anycast IPs]
    GA -->|AWS backbone\nlow latency, stable| EP1[Endpoint\nUS East — ALB]
    GA -->|Health check\nautomatic failover| EP2[Endpoint\nEU West — ALB]

Key difference from CloudFront:

3.3 Amazon Route 53

Amazon Route 53 is AWS’s DNS (Domain Name System) service, with advanced routing and health checking features.

Route 53 routing policies:

3.4 Connectivity Services Comparison Table

Module 4: High Availability Architectures

4.1 Multi-AZ Deployment Patterns

The Multi-AZ pattern distributes resources across multiple Availability Zones in the same Region.

flowchart TB
    LB[Application Load Balancer\nMulti-AZ] --> AZ_A[AZ us-east-1a\nEC2 instances\nRDS Primary]
    LB --> AZ_B[AZ us-east-1b\nEC2 instances\nRDS Standby]
    LB --> AZ_C[AZ us-east-1c\nEC2 instances]

    AZ_A <-->|Synchronous replication| AZ_B

RTO/RPO for Multi-AZ:

• Recovery Time Objective (RTO): time to restore service (< 60 seconds for RDS Multi-AZ)
• Recovery Point Objective (RPO): maximum data loss (near zero for synchronous replication)

4.2 Multi-Region Deployment Patterns

For highest availability and compliance with geographic data requirements:

4.3 Disaster Recovery

The 4 DR strategies (in order of cost and RTO):

flowchart LR
    A[Backup & Restore\nRTO: hours\nLowest cost] --> B[Pilot Light\nRTO: 10-30 min\nLow cost]
    B --> C[Warm Standby\nRTO: minutes\nMedium cost]
    C --> D[Multi-Site Active/Active\nRTO: seconds\nHighest cost]

Quick Reference: AWS Key Numbers (2025–2026)

Global Infrastructure Topology

flowchart TB
    subgraph Global["AWS Global Infrastructure"]
        subgraph Tier1["Tier 1: Regions — 39 launched"]
            NA["North America\n9 Regions"]
            EU["Europe\n8+ Regions"]
            APAC["Asia-Pacific\n12+ Regions"]
            SA["South America\n1 Region"]
            ME["Middle East / Africa\n5+ Regions"]
        end

        subgraph Tier2["Tier 2: Edge and Local"]
            LZ["Local Zones\n43 zones\nnear metro areas"]
            WZ["Wavelength Zones\n33 zones\nin 5G networks"]
            OP["AWS Outposts\nIn customer data centers"]
        end

        subgraph Tier3["Tier 3: Content Distribution"]
            REC["Regional Edge Caches\n15 points\nin AWS Regions"]
            POPs["CloudFront POPs\n750+ points\nin global cities"]
            EPOP["Embedded POPs\n1,140+ points\nin ISP networks"]
        end

        Tier1 <-->|AWS Backbone\n20 million km fiber| Tier2
        Tier1 <-->|AWS Backbone| Tier3
        Tier2 <-->|AWS private network| Tier3
    end

Search Terms

aws · global · infrastructure · core · services · amazon · web · cloud · zones · local · availability · cloudfront · deployment · computing · connectivity · dedicated · edge · locations · numbers · outposts · patterns · wavelength