Edge Computing & Hybrid
Explore AWS edge and hybrid computing architectures to understand how to extend AWS Region capabilities closer to users and devices. Learn to select and design deployments with Outposts, Local Zones, and Wavelength based on latency, data residency, and local processing constraints, ensuring optimal application performance and compliance.
Modern AWS architectures increasingly span beyond traditional Regions to meet deterministic latency, data sovereignty, and local processing requirements. Architects must understand how AWS extends compute to on-premises data centers, 5G carrier networks, and metropolitan areas while maintaining centralized governance. This lesson builds a progression from Region-centric designs to distributed, edge-aware architectures, equipping you to select the correct deployment model when scenarios demand local execution rather than content-delivery acceleration.
Why edge and hybrid architectures matter
AWS Regions provide comprehensive service catalogs and global reach, yet certain workloads cannot tolerate the variable latency of traversing public internet paths to distant Regions. Three primary constraints drive edge adoption:
Deterministic low latency demands compute placement within predictable network boundaries, where even single-digit-millisecond variations disrupt real-time processing, such as industrial control loops or live media production.
Data sovereignty and compliance mandates require that data never leaves a specific jurisdiction or physical facility, making Region-only architectures insufficient when regulations prohibit cross-border data movement.
Locality requirements arise when data volumes are too large to backhaul efficiently or when processing must occur at the point of data generation for operational reasons.
A critical architectural distinction separates content delivery acceleration from compute locality. Compute locality refers to the physical placement of execution environments where application logic runs, rather than mechanisms that only accelerate the delivery of precomputed content. Amazon CloudFront caches content at edge locations, and AWS Global Accelerator optimizes routing through the AWS backbone, but neither executes application workloads locally. When requirements emphasize local processing, data residency, on-premises execution, or deterministic low-latency compute, the correct architectural choices involve services such as AWS Outposts, AWS Local Zones, or AWS Wavelength rather than CDN or traffic acceleration services.
The following diagram illustrates how each deployment model extends from a parent Region with distinct connectivity patterns and latency characteristics.
This architectural model treats edge locations as extensions of Regional capabilities rather than independent islands, which directly informs how governance, networking, and operational patterns must be designed.
AWS Outposts for hybrid on-premises operations
AWS Outposts delivers fully managed AWS infrastructure into customer data centers or co-location facilities, running identical hardware and software stacks found in AWS Regions. This consistency enables the ...