AWS First Steps

AWS Wavelength

An infrastructure service that places compute and storage at the edge of 5G carrier networks to enable ultra-low latency applications

About 5 min readWritten and edited by Awashiro

Overview

AWS Wavelength places AWS compute resources directly inside telecommunications carriers' 5G networks. Because mobile device traffic is processed within the carrier network without traversing the public internet, it achieves single-digit millisecond latency. Wavelength Zones extend a standard AWS Region's VPC, allowing you to deploy existing services like EC2, ECS, and EKS as-is. In Japan, Wavelength Zones are available in Tokyo and Osaka through a partnership with KDDI, supporting ultra-low latency use cases such as AR/VR, real-time gaming, and industrial IoT.

Carrier Integration and Network Architecture

A Wavelength Zone is AWS infrastructure installed within a telecommunications carrier's data center. Unlike standard Availability Zones, it is positioned on a network path directly reachable from the carrier's 5G base stations. Packets from mobile devices reach the Wavelength Zone within the carrier's 5G core network, eliminating hops to the public internet. This architecture removes the latency that used to pile up across many internet hops, and AWS sets single-digit millisecond latency from the mobile device to the application as its target. A Wavelength Zone connects to the parent Region's VPC through a Carrier Gateway. The Carrier Gateway serves as the traffic entry and exit point with the carrier network and is managed as a separate resource from the Internet Gateway. Carrier IP addresses are special Elastic IPs routable only within the carrier network and are not directly reachable from the internet. The recommended design delegates backend processing and data persistence to parent Region services (DynamoDB, S3, RDS, etc.) and places only latency-sensitive frontend processing in the Wavelength Zone.

Application Deployment and Subnet Design

Deploying to a Wavelength Zone involves adding a carrier subnet to an existing VPC. First, opt in to the Wavelength Zone, then create a subnet within the VPC specifying the Wavelength Zone. Configure the subnet's route table with a default route to the Carrier Gateway to establish the communication path to the carrier network. Assigning a Carrier IP to EC2 instances enables direct access from 5G devices. ECS and EKS worker nodes can also be placed in Wavelength subnets, serving container-based microservices at low latency. Note that the instance types and EBS volume types available in Wavelength Zones are far more limited than in the parent Region and vary by zone; as of August 2026, only a subset of instances such as the t3, r5, and g4dn families and gp2 volumes are supported. In zones where GPU instances (g4dn) are available, edge inference workloads are also possible. For load balancers, ALB is available in some zones while NLB cannot be placed, so design your traffic distribution around the support status of the zones you use.

5G Use Cases and Pricing Model

Wavelength's primary use cases are in domains where latency directly impacts experience quality. In AR/VR applications, the delay from head movement to the updated image (motion-to-photon) governs experience quality, and the industry widely cites around 20 ms as a comfort guideline. Only a small slice of that budget can be spent on the network round trip, which is where Wavelength's edge processing highly effective. In cloud gaming, it minimizes the delay from input to frame rendering, maintaining a real-time feel. In industrial IoT, control signals for factory robotic arms and automated guided vehicles are processed at low latency, balancing safety and productivity. For autonomous driving assistance systems, vehicle-to-everything (V2X) communication data is aggregated and analyzed in the Wavelength Zone to deliver real-time traffic conditions. The following description of pricing is as of August 2026. The pricing model includes hourly EC2 instance charges plus data transfer fees from the Wavelength Zone. Resources in a Wavelength Zone are priced separately from the parent Region, so the exact amounts, including data transfer between the zone and the parent Region, need to be checked against the official Wavelength pricing page. Data transfer to the carrier network incurs carrier data transfer charges, and the rate differs by Wavelength Zone. Communication between instances within the same Wavelength Zone is free. The key to cost optimization is minimizing the components placed at the edge and delegating heavy processing to the parent Region.

References (Official AWS Resources)

The primary sources for this page are the official AWS website and documentation. Check the official pages below for the latest specifications and pricing.

If this page and the official documentation disagree, treat the official documentation as authoritative.

ShareXB!

Related Terms

AWS Local ZonesAWS OutpostsAWS Global AcceleratorAWS Direct Connect

Related Services

Amazon EC2Amazon VPCAWS OutpostsAWS Local ZonesAWS Global Accelerator

Related Articles

AWS Edge Strategy - The Hybrid Future Shaped by Outposts, Local Zones, and WavelengthWe explain the hybrid and edge strategy AWS deploys through three edge services - Outposts, Local Zones, and Wavelength - comparing them with Azure Stack and GCP Distributed Cloud to highlight differences in breadth of options and design philosophy.AWS Regions and Edge Locations - The Physical Advantage of 33 Regions and 600+ PoPsWe lay out how AWS regions, Availability Zones, and edge locations are structured, and walk through DR design using the Tokyo and Osaka regions, meeting data sovereignty requirements, and running processing at the edge from a practical standpoint.AWS Availability Zone Design - How Physical Separation and Fault Isolation Create a Reliability AdvantageExamine the design philosophy behind AWS AZs as physically independent data center clusters, compare them with Azure and GCP availability zones, and look at them against the published design criteria and real-world incident examples.AWS First-Mover Advantage and Economies of Scale - What 18 Years of Accumulation Since 2006 Really MeansExamine the first-mover advantage AWS has built since creating the public cloud market in 2006, comparing service breadth and depth, API backward compatibility, and ecosystem depth against Azure and GCP.AWS Global Network Backbone - How Private Submarine Cables and 600+ PoPs Deliver Superior ConnectivityExplore the advantages of AWS's global network built on private submarine cables, dedicated fiber, and 600+ edge locations, compared with GCP's Premium Tier and Azure's network design.

Similar Terms and Articles

AWS WavelengthPlace compute resources at the edge of 5G networks for ultra-low-latency applications5G Edge Computing with AWS Wavelength - Designing Ultra-Low Latency ApplicationsLearn about running applications at the 5G edge with Wavelength, Wavelength Zone design, and practical use cases.Edge and 5G Computing - Achieving Ultra-Low Latency with AWS Wavelength and Local ZonesLearn about ultra-low latency computing with AWS Wavelength and Local Zones. Covers processing at the 5G network edge, proximity placement in metro areas, use cases, and when to use standard regions instead.AWS Edge Strategy - The Hybrid Future Shaped by Outposts, Local Zones, and WavelengthWe explain the hybrid and edge strategy AWS deploys through three edge services - Outposts, Local Zones, and Wavelength - comparing them with Azure Stack and GCP Distributed Cloud to highlight differences in breadth of options and design philosophy.