AWS Global Accelerator supports two main types of accelerators: standard accelerators and custom routing accelerators, each with distinct endpoint types.
Endpoints for Standard Accelerators:
Standard accelerators route traffic to endpoints that can be:
- Network Load Balancers (NLBs)
- Application Load Balancers (ALBs)
- Amazon EC2 instances
- Elastic IP addresses
These endpoints must be located in AWS Regions and can be internet-facing or internal. Global Accelerator uses static IP addresses as fixed entry points and distributes incoming traffic to these endpoints based on factors like client location and endpoint health. You can configure endpoint weights to control the proportion of traffic routed to each endpoint, enabling use cases such as blue/green deployments and A/B testing. For dual-stack accelerators supporting both IPv4 and IPv6, only dual-stack endpoints (NLBs, ALBs, and EC2 instances that support dual-stack) can be added[1][2][4][5][6][7][8].
Endpoints for Custom Routing Accelerators:
Custom routing accelerators are designed for applications requiring custom logic to route users to specific destinations and ports. The endpoints for custom routing accelerators are Amazon VPC subnets that host one or more EC2 instances. This setup is useful for scenarios like multiplayer gaming or VoIP applications, where users need to be assigned to particular servers or sessions based on criteria such as geography or user attributes. Custom routing accelerators support only IPv4 addresses[4][9].
Additional Notes:
- On-premises resources cannot be directly configured as endpoints; however, you can use Network Load Balancers in AWS Regions to front on-premises endpoints and register those NLBs with Global Accelerator.
- Global Accelerator monitors endpoint health continuously and routes traffic only to healthy endpoints.
- Elastic IP addresses used as endpoints are static public IPs allocated to your AWS account.
- For client IP address preservation, certain requirements apply to the endpoint types used.
In summary, AWS Global Accelerator endpoints vary by accelerator type: standard accelerators use Network Load Balancers, Application Load Balancers, EC2 instances, or Elastic IP addresses, while custom routing accelerators use VPC subnets with EC2 instances as endpoints. This flexibility allows Global Accelerator to support a wide range of application architectures and traffic routing needs[1][4][6][9].
Citations:
[1] https://docs.aws.amazon.com/global-accelerator/latest/dg/about-endpoints.html
[2] https://docs.aws.amazon.com/global-accelerator/latest/dg/about-endpoint-groups.html
[3] https://docs.aws.amazon.com/general/latest/gr/global_accelerator.html
[4] https://docs.aws.amazon.com/global-accelerator/latest/dg/introduction-how-it-works.html
[5] https://docs.aws.amazon.com/global-accelerator/latest/dg/what-is-global-accelerator.html
[6] https://docs.aws.amazon.com/global-accelerator/latest/dg/about-endpoints-adding-endpoints.html
[7] https://docs.aws.amazon.com/pdfs/global-accelerator/latest/dg/global-accelerator-guide.pdf
[8] https://tutorialsdojo.com/aws-global-accelerator/
[9] https://aws.amazon.com/global-accelerator/faqs/