Edge and 5G Computing - Achieving Ultra-Low Latency with AWS Wavelength and Local Zones
An explanation of ultra-low latency computing with AWS Wavelength and Local Zones. It introduces processing at the 5G network edge, placement close to urban areas, use cases, and how to choose between these and standard Regions.
The Demand for Ultra-Low Latency Computing
In standard AWS Regions, latency can grow depending on the distance to end users, and for some applications this latency is unacceptable. In game streaming (cloud gaming), low latency determines whether play feels comfortable. In AR/VR applications, latency causes motion sickness. In inference for self-driving cars, even slight latency affects safety. For these ultra-low latency requirements, AWS offers two edge computing options: Wavelength and Local Zones. Wavelength places AWS compute inside the networks of 5G carriers, and traffic from 5G devices reaches AWS directly without going through the internet. Local Zones place AWS infrastructure in major metropolitan areas and run processing close to end users in those cities.
How Wavelength Works
A Wavelength Zone is AWS infrastructure installed inside a carrier's 5G network. Traffic from 5G devices is processed within the carrier's network and does not go through the internet, so latency stays low. Wavelength Zones are installed inside the networks of partner carriers, and their locations differ by country and region. In a Wavelength Zone, you can use EC2 instances, EBS volumes, and VPC subnets. You deploy by creating a carrier gateway in the VPC of the parent Region and adding a subnet in the Wavelength Zone. Instances in a Wavelength Zone can access S3, DynamoDB, RDS, and other services in the Region, and you can build an architecture that combines processing at the edge with data storage in the Region.
Local Zones and When to Use Each
Local Zones are AWS infrastructure placed in major metropolitan areas where population and industry are concentrated (Los Angeles, New York, Chicago, Dallas, and others), and they deliver applications to users in those areas with low latency. While Wavelength is for 5G devices, Local Zones deliver low latency to users in that city regardless of whether they connect over wired or wireless networks. You can use EC2, EBS, ECS, EKS, RDS, ElastiCache, and more. As a way to choose, you use Wavelength when you need ultra-low latency from 5G devices, Local Zones when you need low latency for users in a specific city, and standard Regions otherwise. The locations where Local Zones and Wavelength Zones are available differ by country and region, so you check the official list before using them.
Wavelength and Local Zones Pricing
EC2 instance pricing in Wavelength Zones is on par with standard Regions, but the available instance types are limited. Data transfer through the carrier gateway costs about $0.05 per GB (as of August 2026). Resources in Local Zones have pricing that differs from the parent Region. In both cases the available services are more limited than in a Region. For this reason, a design that places only the components with strict latency requirements in Wavelength/Local Zones and handles everything else in the Region makes it easier to keep costs down.
Deciding What to Place at the Edge
One of the important points in designing edge computing is deciding what to place at the edge and what to keep in standard Regions. Simply putting everything at the edge is not the answer. Compute resources at the edge are limited, and there are also constraints on the services you can place there. You estimate the latency you can tolerate and place at the edge only the highly time-sensitive processing whose experience suffers once that range is exceeded. Large-scale data accumulation, time-consuming analysis, and management functions stay in the Region. The key to a design that makes effective use of limited edge resources is to break down processing with the latency budget in mind and carefully select the parts that truly need the edge.
Typical Use Cases and Selection
There are several typical examples of uses that require ultra-low latency. They include AR and VR, where immersion matters; games, where reaction speed decides who wins; processing that analyzes video on the spot; and real-time control of vehicles and industrial equipment, among others. In these cases, there is no room to send data on a round trip to a distant data center. When selecting a method, you choose based on your requirements. If you need ultra-low latency from mobile connections, you use the approach of placing compute inside the carrier network. If you want to broadly provide performance close to that of a Region in a specific metro area, you use the approach of geographically nearby sites. It is important to clarify what you want to solve and choose the form of edge that fits.
Summary - Guidelines for Using Edge Computing
AWS Wavelength and Local Zones are edge computing options for applications that need ultra-low latency. Wavelength handles low-latency processing inside 5G networks, and Local Zones handle it in major metropolitan areas. Consider them for use cases where latency affects user experience or business value, such as game streaming, AR/VR, real-time video analysis, and financial trading. If standard Regions are sufficient for your latency requirements, we recommend using Regions from the perspective of cost and available services.
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.
- AWS Wavelength official page
- AWS Local Zones official page
- Amazon EC2 official page
- AWS Documentation (official)
If this page and the official documentation disagree, treat the official documentation as authoritative.