What is AWS Global Accelerator?
AWS Global Accelerator is a networking service that improves the availability and performance of applications by routing user traffic through the AWS global network.
It gives your application static Anycast IP addresses that act as fixed entry points.
- For IPv4 accelerators, AWS provides two static IPv4 addresses
- For dual-stack accelerators, AWS provides two IPv4 addresses and two IPv6 addresses
- User traffic enters the AWS network at a nearby edge location
- AWS then routes the traffic to a healthy application endpoint in one or more AWS Regions
Note
Global Accelerator does not cache content at the edge. It accelerates network routing to your application endpoints.
Unicast IP vs Anycast IP
Unicast IP
With a unicast IP, one server or endpoint owns one IP address.
- Different servers usually have different IP addresses
- The client connects to a specific server address
Anycast IP
With an Anycast IP, multiple edge locations can advertise the same IP address.
- The client uses the same IP address
- The network routes the client to a nearby or optimal edge location
- From there, AWS carries the traffic across its global network
Important
In Global Accelerator, the static IP addresses are Anycast from AWS edge locations. This means users connect to the same IPs globally, but traffic enters AWS close to the user.
How traffic flows
A typical request flow looks like this:
- The client connects to one of the Global Accelerator static IP addresses
- Traffic reaches the nearest suitable AWS edge location
- Traffic travels over the AWS global network
- Global Accelerator sends the request to a healthy endpoint
For TCP traffic, Global Accelerator can terminate the client connection at the edge and establish a separate connection to the application endpoint.
Supported endpoint types
For standard accelerators, endpoints can include:
- Application Load Balancers
- Network Load Balancers
- EC2 instances
- Elastic IP addresses
Application Load Balancers and Network Load Balancers can be either internet-facing or internal.
Tip
Global Accelerator is useful when you want a stable global entry point while still being able to change or add regional backend endpoints.
Health checks and failover
Global Accelerator continuously checks the health of endpoints in a standard accelerator.
- Traffic is routed only to healthy active endpoints
- If an endpoint becomes unhealthy, new traffic can be routed to another healthy endpoint
- This is useful for multi-Region failover and disaster recovery
You can also control traffic using:
- Endpoint weights: adjust how much traffic each endpoint receives
- Traffic dials: adjust how much traffic goes to an endpoint group or Region
Important
If an exam question mentions fast regional failover, healthy endpoints, or multi-Region routing with static IPs, Global Accelerator is often a strong fit.
Security
Global Accelerator can improve the network security posture of an application.
- Clients only need to allow-list a small set of static IP addresses
- For IPv4 accelerators, this usually means 2 fixed public IPs
- The static entry points are protected by AWS Shield
- Origins can be masked behind Global Accelerator instead of being exposed directly to the public internet
Note
Global Accelerator helps with DDoS resilience and fixed entry points, but it is not a replacement for security groups, authentication, encryption, or AWS WAF where needed.
AWS Global Accelerator vs CloudFront
Both Global Accelerator and CloudFront use AWS edge locations and the AWS global network, but they solve different problems.
| Feature | CloudFront | Global Accelerator |
|---|---|---|
| Main purpose | Content delivery network | Network accelerator |
| Best for | Static and dynamic web content | TCP or UDP applications |
| Caching | ✅ Can cache content at the edge | ❌ Does not cache content |
| Edge behavior | Serves cacheable content from edge locations | Routes traffic through AWS edge locations |
| Entry point | CloudFront distribution domain name | Static Anycast IP addresses |
| Common protocols | HTTP and HTTPS web traffic | TCP and UDP |
| Good use cases | Websites, APIs, images, videos, dynamic web apps | Gaming, IoT, VoIP, static IP allow-listing, multi-Region failover |
| Regional failover | Possible with origin failover | Built around routing to healthy regional endpoints |
Important
Use CloudFront when the question focuses on CDN, caching, web content, images, videos, or serving content from the edge.
Use Global Accelerator when the question focuses on static IPs, TCP/UDP traffic, low-latency routing, or fast regional failover.
Common use cases
Static IP allow-listing
Some clients or enterprise networks require fixed IP addresses to be allow-listed.
Global Accelerator gives the application stable IPs even if the backend endpoints change.
Multi-Region failover
Global Accelerator can route users to healthy endpoints across Regions.
This is useful for high availability and disaster recovery.
Low-latency TCP or UDP applications
Global Accelerator is useful for latency-sensitive applications such as:
- Gaming
- IoT
- Voice over IP
- Real-time applications
Blue/green or traffic shifting
Endpoint weights and traffic dials can be used to gradually shift traffic between endpoints or Regions.
What to remember
- Global Accelerator gives applications static Anycast IP addresses
- Traffic enters the AWS global network close to the user
- It routes traffic to healthy regional endpoints
- It supports TCP and UDP
- It does not cache content
- CloudFront is usually better for CDN and web caching use cases
- Global Accelerator is usually better for static IP, non-HTTP, and fast failover use cases