SkarpChapter 2 of 26
AWS Global Infrastructure, Availability Zones, and Core Building Blocks
Before you design anything, you need to know where it runs: dive into Regions, Availability Zones, and the core services that almost every exam scenario builds upon.
Big Picture: Why Global Infrastructure Matters
Why Global Infrastructure?
Before designing on AWS, you need a clear map of where workloads run and how AWS is physically organized. The exam quietly tests this in many scenario questions.
Core Pieces
AWS is built from Regions, Availability Zones (AZs), and edge locations. Regions are geographic areas, AZs are isolated data center clusters, and edge locations are global cache/edge sites.
Well-Architected Connection
Regions and AZs are central to the Reliability, Performance efficiency, and Cost optimization pillars. They affect fault tolerance, latency, data residency, and pricing.
Mental Model
Imagine a Region as a city, each AZ as a neighborhood with its own power and roads, and edge locations as small local post offices and caches sprinkled worldwide.
What You Will Do
You will decode Regions/AZs/edges, learn EC2 and S3 basics, and practice mapping scenarios to the right building blocks by always asking: Where is this running, and what if it fails?
Regions: Separate Geographic Areas with Boundaries
What Is a Region?
An AWS Region is a physically separate geographic area containing multiple Availability Zones, like `us-east-1` or `eu-central-1`. Regions are the top-level building blocks.
Isolation and Resilience
Regions are isolated from each other for fault tolerance and security. A failure in one Region should not directly impact workloads in another Region.
Compliance and Residency
Region choice often follows laws and policies, such as GDPR for EU data. If a scenario says data must stay in the EU, think EU Regions like `eu-central-1` or `eu-west-1`.
Latency and Cost
Users usually get the best performance from the closest Region. Pricing can also differ by Region, so cost-sensitive designs may compare Regions explicitly.
Global vs Regional Services
Some services are global (IAM, Route 53, CloudFront). Others are regional (EC2, RDS, S3 buckets). Regional resources do not automatically span Regions.
Availability Zones: The Core of High Availability
What Is an AZ?
An Availability Zone (AZ) is one or more discrete data centers in a Region, with independent power, cooling, and networking, plus high-speed links to other AZs.
Failure Isolation
AZs are designed so that a failure in one AZ should not take down others in the same Region. This is the foundation of fault-tolerant AWS architectures.
Single vs Multi-AZ
Single-AZ designs are cheaper but fragile. Multi-AZ designs are the default for high availability and directly support the Reliability pillar of Well-Architected.
Service Patterns
Common patterns: EC2 instances spread across AZs behind a load balancer, RDS Multi-AZ for failover, and S3 storing data across multiple AZs automatically.
Subnets and AZs
Subnets are AZ-scoped. To use multiple AZs, you create one subnet per AZ in your VPC. Exam questions often hint at this with “subnet in each AZ”.
Edge Locations, CloudFront, and Global Reach
What Are Edge Locations?
Edge locations are globally distributed sites where CloudFront and other edge services run. They sit closer to users than Regions, reducing latency.
CloudFront Basics
Amazon CloudFront is AWS’s CDN. It caches content at edge locations and fetches it from origins like S3, EC2, on-prem HTTP servers, or load balancers.
When to Use CloudFront
Use CloudFront when you have global users, need low-latency static or dynamic content, want to offload origin traffic, or need signed URLs/cookies for access control.
Global Service Behavior
CloudFront is a global service. You create a distribution once; AWS automatically serves users from the nearest edge location. You do not pick AZs for edges.
Exam Clues
Phrases like “improve performance for users worldwide” or “cache static content closer to users” are strong hints that CloudFront and edge locations are the right answer.
Design Walkthrough: Multi-AZ vs Multi-Region
The Scenario
Online learning platform, mostly EU users, some in North America. Uses EC2, RDS, S3. Need to decide between multi-AZ and multi-Region designs.
Option 1: Multi-AZ Single Region
Place EC2 in two AZs behind an ALB, use RDS Multi-AZ, and S3 in `eu-central-1`. This survives AZ outages and is simpler and cheaper than multi-Region.
Multi-AZ Pros and Cons
Pros: high availability inside the Region, simpler ops, good EU latency. Cons: weaker for NA users, no protection from Region-wide failures.
Option 2: Active-Passive Multi-Region
Run a primary stack in `eu-central-1` and a standby in `us-east-1`, with async DB replication and S3 cross-Region replication for critical data.
Multi-Region Tradeoffs
Pros: can survive Region failure, strong DR. Cons: more complex routing (Route 53), higher cost and operational complexity. Exam: match design to the stated requirement.
Amazon EC2: Virtual Servers as a Core Building Block
What Is EC2?
Amazon EC2 provides resizable virtual servers in the cloud. It is a core exam service, used in many architectures for compute workloads.
Instance Types and Lifecycle
Instance types are grouped by optimization (compute, memory, storage, network). Lifecycle: `pending` → `running` → `stopping`/`stopped` → `terminated`.
Pricing Models
On-Demand for flexibility, Reserved Instances or Savings Plans for long-term savings, and Spot Instances for cheap but interruptible capacity.
Placement and Networking
Each EC2 instance runs in a subnet tied to one AZ. High availability comes from multiple instances across AZs, usually behind a load balancer.
Storage and Exam Traps
EC2 uses EBS (AZ-scoped block storage) or instance store (ephemeral). A single instance is never multi-AZ; HA requires multiple instances across AZs.
Amazon S3: Durable Object Storage Across AZs
What Is S3?
Amazon S3 is object storage. It stores data as objects in buckets and is designed for massive scale, high durability, and simple HTTP-based access.
Durability and Availability
S3 Standard is designed for 99.999999999% durability by storing data across multiple AZs in a Region, with high availability SLAs.
Buckets and Objects
You create buckets in a specific Region. Each object has a key (its path/name) and optional metadata. Capacity is effectively unlimited.
Storage Classes and Access
Storage classes balance cost and access patterns. Access is controlled via IAM, bucket policies, ACLs, and S3 Block Public Access settings.
Region Behavior and Traps
Buckets are Regional and multi-AZ by default. S3 is not block storage. One Zone-IA is single-AZ and cheaper but less resilient.
Thought Exercise: Mapping Workloads to Regions, AZs, and Edges
Work through these scenarios and decide which global infrastructure features you would use. There are no single “correct” answers here, but your reasoning should align with exam logic.
- Internal HR application for a single country
- Users: Employees in one country.
- Data: Must remain in-country by law.
- Latency: Moderate requirements.
- Question: Would you use multiple Regions? How many AZs? Would CloudFront help?
Think: Data residency and compliance likely drive you to a single Region in that country or nearest legal option, with multi-AZ for availability. CloudFront is probably not necessary unless you serve large media files.
- Global news website with heavy static content
- Users: Worldwide.
- Data: Public articles and images.
- Latency: Users expect fast page loads everywhere.
- Question: What combination of Regions, AZs, and edge locations would you choose?
Think: A core origin (S3 or EC2) in one or a few Regions, CloudFront to cache at edge locations, and multi-AZ in each origin Region for resilience.
- Mission-critical trading platform
- Users: Regional (for example, only in APAC), but with strict uptime and RTO/RPO.
- Data: Financial transactions.
- Latency: Very low latency required.
- Question: Is multi-AZ enough, or do you need multi-Region? How would you justify the extra cost?
Think: Multi-AZ is the baseline. If the scenario emphasizes surviving Region-wide failures or has extremely strict DR requirements, multi-Region may be justified despite higher cost.
Write down your answers in your own words. After that, re-read each scenario and underline which requirement (compliance, latency, availability, or cost) drove your decision. This mirrors how exam questions are structured.
Check Understanding: Regions, AZs, and Edge Locations
Answer this question to test your grasp of AWS global infrastructure.
A company runs an API on EC2 instances in a single Region. Users are located around the world and report high latency. The company does NOT want to manage additional application stacks in other Regions. What is the MOST effective solution to improve global performance?
- Create a second copy of the API stack in another Region and use Route 53 latency-based routing.
- Place the existing EC2 instances in multiple Availability Zones within the current Region.
- Use Amazon CloudFront with the existing API endpoint as the origin.
- Move the EC2 instances to a larger instance type in the current Region.
Show Answer
Answer: C) Use Amazon CloudFront with the existing API endpoint as the origin.
Using Amazon CloudFront with the existing API endpoint as the origin leverages **edge locations** to reduce latency for global users without deploying additional Regional stacks. Multi-AZ (option B) improves availability, not global latency. Adding another Region (option A) increases complexity and contradicts the requirement. Larger instances (option D) may help throughput but not network distance for global users.
Check Understanding: EC2 and S3 in Multi-AZ Designs
Test how well you can apply EC2 and S3 behavior to availability scenarios.
An application stores user uploads on an S3 bucket in `us-east-1` and processes them with EC2 instances in two Availability Zones within the same Region. Which statement is MOST accurate?
- If one Availability Zone fails, the S3 bucket will become unavailable because it is tied to that AZ.
- The S3 bucket is designed to remain available because S3 Standard stores data across multiple AZs in the Region.
- The EC2 instances are automatically moved to another Region if the current Region fails.
- To make the S3 data multi-AZ, you must configure S3 Cross-Region Replication.
Show Answer
Answer: B) The S3 bucket is designed to remain available because S3 Standard stores data across multiple AZs in the Region.
S3 Standard is designed to store data redundantly across multiple AZs within a Region, so an AZ failure should not make the bucket unavailable. EC2 instances are not automatically moved across Regions. Cross-Region Replication is for multi-Region, not multi-AZ; S3 is already multi-AZ by default.
Key Term Review: Global Infrastructure and Core Services
Use these flashcards to reinforce terminology and core ideas.
- AWS Region
- A physically separate geographic area that contains multiple Availability Zones. Regions are isolated from each other for fault tolerance, security, compliance, and latency considerations.
- Availability Zone (AZ)
- One or more discrete data centers in a Region with independent power, cooling, and networking, connected to other AZs in the Region with high-speed, low-latency links. The basic unit for high availability designs.
- Edge location
- A site used by services like Amazon CloudFront to cache content and terminate connections closer to users, reducing latency. Part of AWS’s global edge network.
- Amazon CloudFront
- AWS’s content delivery network (CDN) that uses edge locations to cache and deliver content from origins such as S3, EC2, on-prem servers, or load balancers, improving global performance.
- Amazon EC2
- Amazon Elastic Compute Cloud (EC2) provides resizable virtual servers in the cloud, with different instance types, pricing models, and placement options across subnets and Availability Zones.
- Amazon S3
- Amazon Simple Storage Service (S3) is object storage that stores data as objects in buckets, with high durability and availability by replicating data across multiple Availability Zones in a Region.
- Multi-AZ architecture
- A design that deploys resources such as EC2 instances or RDS databases across multiple Availability Zones in a single Region to improve availability and fault tolerance.
- Multi-Region architecture
- A design that deploys workloads across two or more AWS Regions to improve disaster recovery, geographic redundancy, and sometimes latency for multi-continent users.
- S3 Cross-Region Replication (CRR)
- A feature that automatically replicates objects from an S3 bucket in one Region to a bucket in another Region, used for multi-Region redundancy, compliance, and latency optimization.
- EC2 pricing models
- The main EC2 pricing models are On-Demand (no commitment), Reserved Instances/Savings Plans (long-term commitment for discounts), and Spot Instances (discounted spare capacity that can be interrupted).
Tying It Together: Mapping Building Blocks to Exam Scenarios
Regions in Scenarios
Regions are chosen for compliance, latency, service availability, and cost. Look for phrases about data residency, continents, or disaster recovery across Regions.
AZs in Scenarios
AZs are the core tool for high availability within a Region. Clues include “survive data center failure”, “no single point of failure”, and “Multi-AZ”.
Edges and CloudFront in Scenarios
Edge locations and CloudFront appear when you see “global users”, “cache static content”, or “reduce origin load” while keeping a single origin stack.
EC2 and S3 Roles
EC2 is flexible compute with OS control. S3 is durable object storage for logs, data lakes, static sites, and backups, often feeding CloudFront.
Your Mental Checklist
Always ask: Where does it run? What if an AZ or Region fails? Where are the users? Which is primary: availability, latency, compliance, or cost?
Key Terms
- Amazon S3
- Amazon Simple Storage Service (S3) is object storage that stores data as objects in buckets, with high durability and availability by replicating data across multiple Availability Zones in a Region.
- S3 bucket
- A container for objects stored in Amazon S3, created in a specific AWS Region and identified by a globally unique name.
- AWS Region
- A physically separate geographic area that contains multiple Availability Zones. Regions are isolated from each other for fault tolerance, security, compliance, and latency considerations.
- Amazon EC2
- Amazon Elastic Compute Cloud (EC2) provides resizable virtual servers in the cloud, with multiple instance types, pricing models, and placement options.
- Edge location
- A globally distributed site used by services like Amazon CloudFront to cache content and terminate connections closer to users, reducing latency.
- Instance type
- A specification for an EC2 instance that defines its CPU, memory, storage, and networking capacity, optimized for different workload types.
- Amazon CloudFront
- A content delivery network (CDN) service that uses edge locations to cache and deliver content from origins such as S3, EC2, on-premises servers, or load balancers.
- Multi-AZ architecture
- An architecture that deploys resources across multiple Availability Zones in a Region to improve availability and fault tolerance.
- Availability Zone (AZ)
- One or more discrete data centers in a Region with independent power, cooling, and networking, connected to other AZs in the Region with high-speed, low-latency links.
- Multi-Region architecture
- An architecture that deploys workloads across two or more AWS Regions to improve disaster recovery and geographic redundancy.
- S3 Cross-Region Replication (CRR)
- A feature that automatically replicates objects from an S3 bucket in one Region to a bucket in another Region for redundancy, compliance, or latency optimization.