What is Amazon Global Accelerator?

Amazon Global Accelerator is a networking service offered by Amazon Web Services (AWS) that helps improve the availability and performance of internet applications for global users. It acts as a traffic forwarding service that directs user requests to the optimal endpoint for their location, reducing latency, improving user experience, and providing greater resilience against failures.

Components of Amazon Global Accelerator

Here are some key features of Amazon Global Accelerator:

• Static IP addresses: Global Accelerator provides static IP addresses that act as a fixed entry point for application endpoints. These IP addresses are anycast, meaning they are geographically distributed across the AWS edge network. This ensures that users are directed to the closest endpoint regardless of location.

• Accelerators: They act as traffic directors for internet applications, intelligently routing user requests to the optimal endpoint. Amazon Global Accelerator offers two types of accelerators:

  • Standard accelerator: These are designed to improve the application’s availability. They direct user traffic over the high-performance AWS global network and route users to the nearest healthy endpoint located in an AWS Region. This means users connect to the application servers geographically closest to them, reducing latency (delay) and improving responsiveness. If an endpoint in a specific region experiences an outage, Global Accelerator automatically routes traffic from a faulty endpoint to a healthy endpoint in another region. This redundancy helps maintain application availability during regional disruptions.

  • Custom routing accelerator: These provide more granular control over how user traffic is directed. It allows us to define custom rules for traffic routing. We can map users to specific endpoints based on various factors like:

    • User location: Similar to a standard accelerator, we can prioritize endpoints based on user location for better performance.

    • User type: We can route different types of users (e.g., free vs. premium) to different endpoints with varying capabilities.

    • Application version: We can direct users to specific endpoints based on the application version they’re using, e.g., route testers to a specific endpoint that is hosting the newer version of the application that includes the new features we have to test.

• DNS name: The DNS name is a hostname provided by Amazon Global Accelerator that you can use to route traffic to your application. When you create a Global Accelerator, you’re assigned a DNS name that you can use to direct traffic to the accelerator. This DNS name resolves to one or more Anycast IP addresses, allowing clients to connect to your application via the closest edge location.

• Network zone: A network zone represents the collection of edge locations and points of presence (PoPs) where Amazon Global Accelerator operates. These network zones are geographically distributed and interconnected to form the AWS global network infrastructure. When you create an accelerator, you can choose whether to enable traffic to be routed through the entire AWS global network (including edge locations and PoPs in all regions) or through specific regions.

• Listener: A listener is a configuration object within Amazon Global Accelerator that defines how traffic should be received and processed by the accelerator. Each listener is associated with one or more ports, protocols (such as TCP or UDP), and endpoint groups. Listeners specify the port range and protocol for incoming client connections and determine how traffic is distributed to endpoint groups.

• Endpoint group: An endpoint group is a collection of application endpoints (such as EC2 instances, Network Load Balancers, or Elastic IP addresses) that receive traffic from Amazon Global Accelerator. Endpoint groups are associated with listeners and define the backend infrastructure that handles client requests. You can configure endpoint groups to distribute traffic evenly across endpoints or to prioritize specific endpoints based on routing algorithms.

• Endpoint: An endpoint represents an individual destination for client traffic within an endpoint group. Endpoints can be instances of AWS resources such as EC2 instances, Network Load Balancers, or Elastic IP addresses. When traffic is routed to an endpoint group, it is distributed among the endpoints according to the configured routing algorithm.

Use case

Let’s consider a use case in which a company operates a globally distributed web application that serves customers from multiple geographic regions. The company wants to improve its application’s performance and availability by leveraging Amazon Global Accelerator.

We can use Amazon Global Accelerator to improve the performance and availability of the application in the following ways:

• A user accesses the web application by entering the DNS name provided by the Amazon Global Accelerator into their browser.

• The DNS name resolves to one or more Anycast IP addresses advertised from multiple edge locations worldwide.

• The user’s request is routed to the closest edge location based on network conditions.

• The accelerator forwards the request to the appropriate endpoint group associated with the listener configured for the request’s port and protocol.

• The endpoint group distributes the request to one of the backend servers hosting the application content in the respective region.

• The backend server processes the request and sends the response back to the user via the Global Accelerator.

In conclusion, Amazon Global Accelerator acts as an intermediary between the users and application endpoints, ensuring optimal traffic routing for a better user experience and overall application performance.