AWS Knowledge Hub

English

Tiếng Việt

English

Tiếng Việt

Appearance

AWS Global Infrastructure ​

Summary ​

  • AWS’ global infrastructure is built from Regions, Availability Zones, Edge Locations, Local Zones, and Wavelength Zones to deliver low‑latency, highly available, and fault‑tolerant applications.
  • Choosing the right Region and designing across multiple AZs directly impacts reliability, performance, compliance, and cost.
  • For most architectures, Multi‑AZ within a Region plus CloudFront at the edge is the baseline pattern for high availability and good user experience.

Infrastructure overview (mindmap) ​

mermaid
mindmap
  root((AWS Global Infrastructure))
    Regions
      Isolated
      Data residency
      Regional pricing
    Availability Zones
      Multiple per Region
      Fault isolation
      Low-latency links
    Edge Locations
      CloudFront PoPs
      Lambda@Edge
    Local Zones
      Closer to end users
      Single-digit ms latency
    Wavelength Zones
      Embedded in 5G networks
      Mobile edge computing

Best Practices ​

  • Pick Regions deliberately based on latency to users, compliance/data residency, required services, and pricing—not just by habit.
  • Design Multi‑AZ deployments for production workloads (EC2, RDS, ECS, ALB, etc.) to survive an AZ failure without downtime.
  • Use CloudFront and Edge Locations to move static content and caching closer to users and reduce latency and origin load.
  • For ultra low‑latency workloads (gaming, media, 5G, AR/VR), evaluate Local Zones or Wavelength Zones where available.
  • Plan for disaster recovery: Multi‑AZ for high availability inside a Region, and Multi‑Region for stricter RTO/RPO or regulatory needs.

Exam Notes ​

  • Be able to clearly distinguish Region vs Availability Zone vs Edge Location, and know typical design guidance (Multi‑AZ, Region selection, CloudFront for global).
  • Many exam questions encode constraints like data residency, latency to users, or global users—which often point to a specific Region strategy or CloudFront.
  • Remember that AWS publishes SLAs (for example, EC2 Multi‑AZ 99.99%, S3 11 9’s durability), but in exams you rarely need exact numbers—just which design improves availability.

AWS documentation ​