Global Infrastructure

Cloud providers operate physical data centers around the world, organized into a hierarchy of regions, availability zones, and edge locations. Understanding this topology is critical for building performant, resilient, and compliant systems.

Regions

A region is a geographic area containing two or more availability zones. Each region is independent — data does not flow between regions unless you explicitly enable it. Examples: us-east-1 (N. Virginia), eu-west-1 (Ireland), ap-southeast-1 (Singapore).

When selecting a region, consider:

• Latency to users: Deploy close to your user base. A region on the opposite side of the planet adds 100-300 ms of round-trip time.
• Data residency: Some regulations (GDPR, Brazil’s LGPD, China’s Cybersecurity Law) require data to stay within national borders. Choose a region that satisfies these requirements.
• Service availability: Not all services are available in every region. Newer services often launch in a subset of regions first.
• Pricing: Resource costs vary by region. us-east-1 is typically the least expensive; regions like sa-east-1 (São Paulo) can be 30-60% more.

Availability Zones

An availability zone (AZ) is one or more discrete data centers within a region, each with independent power, cooling, and networking. AZs are connected by high-bandwidth, low-latency fiber, typically within 2 ms of each other.

Designing for availability zone redundancy is the primary mechanism for high availability in the cloud:

Region: us-east-1
├── us-east-1a (data center)
├── us-east-1b (data center)
├── us-east-1c (data center)
└── us-east-1d (data center)

Deploying application instances across three AZs means that if one AZ fails, the remaining two continue serving traffic with no interruption.

Edge Locations

Edge locations are points of presence (PoPs) that cache content closer to end users. They power content delivery networks (CDNs) like CloudFront (AWS), Azure CDN, and Cloud CDN (GCP).

Edge locations do not run your application code (that is what regions and AZs are for). They cache static assets — images, JavaScript bundles, CSS, API responses — reducing latency and offloading origin servers.

Latency Optimization

End-to-end latency is the sum of several components:

1. Propagation delay: Speed-of-light limitation (roughly 1 ms per 100 km of fiber).
2. Serialization delay: Time to push bits onto a wire (negligible at modern speeds).
3. Queuing delay: Time packets wait at routers.
4. Processing delay: Time for the server to handle the request.

To minimize latency: choose regions near your users, use a CDN for static content, and offload compute to edge locations when possible.

Data Residency and Sovereignty

Many countries require that certain data types remain within their borders. Cloud providers address this through region isolation:

• Data stored in eu-central-1 (Frankfurt) never leaves Germany unless you explicitly replicate it.
• AWS Artifact, Azure Compliance offerings, and GCP Compliance Reports provide documentation for regulatory audits.
• Some providers offer dedicated regions for government workloads (AWS GovCloud, Azure Government).

Disaster Recovery Strategies

Global infrastructure enables several disaster recovery patterns:

RPO (Recovery Point Objective) measures acceptable data loss; RTO (Recovery Time Objective) measures acceptable downtime. Your DR strategy should match the criticality of your workload.