Global Accelerator

🌎 What is AWS Global Accelerator?

AWS Global Accelerator is a network-level traffic manager that improves the availability and performance of your global applications by routing user traffic through the AWS global network and accelerating it to your application’s regional endpoints.

Instead of relying on the public internet, traffic uses AWS backbone, with static IPs and automatic failover.

---

🔧 How It Works

Core Components:

Component	Description
Accelerator	The resource with 2 static IP addresses (global entry points)
Listener	Port mapping (e.g., TCP 80/443)
Endpoint Group	One per AWS Region (e.g., us-east-1, ap-south-1)
Endpoints	Can be ALB, NLB, EC2 IPs, or Elastic IPs

Think of Global Accelerator as route optimization + HA failover across regions and endpoints.

---

✅ Use Cases

Use Case	Why Use Global Accelerator
🛠️ Multi-region API	Direct users to closest healthy region with lowest latency
🌐 Global SaaS/Web app	Provide static IPs for DNS simplification and firewall allowlisting
🧪 Real-time gaming/VoIP	Lower jitter, packet loss, and latency
📦 Software download/CDN backend	Improve performance for global binary delivery
🧰 Disaster recovery/failover routing	Reroute traffic to another region automatically

---

🆚 Global Accelerator vs CloudFront

Feature	Global Accelerator	CloudFront
Level	Network Layer (TCP/UDP)	Application Layer (HTTP/HTTPS)
Protocol Support	TCP, UDP	HTTP, HTTPS only
Latency Optimization	✅ Yes (BGP + Anycast + AWS Backbone)	✅ Yes (via caching at edge)
Content Caching	❌ No	✅ Yes (full CDN)
Static IPs	✅ Yes	❌ No
Use Case	APIs, gaming, real-time apps	Static sites, streaming, media delivery

---

🛠️ Terraform Example: Global Accelerator with ALB in 2 Regions

# 1. Create the Global Accelerator
resource "aws_globalaccelerator_accelerator" "main" {
  name               = "yuva-global-accelerator"
  ip_address_type    = "IPV4"
  enabled            = true
}

# 2. Add a listener
resource "aws_globalaccelerator_listener" "http_listener" {
  accelerator_arn = aws_globalaccelerator_accelerator.main.id
  port_ranges {
    from_port = 80
    to_port   = 80
  }
  protocol = "TCP"
  client_affinity = "NONE"
}

# 3. Endpoint Group for us-east-1
resource "aws_globalaccelerator_endpoint_group" "useast1" {
  listener_arn = aws_globalaccelerator_listener.http_listener.id
  endpoint_group_region = "us-east-1"

  endpoint_configuration {
    endpoint_id = aws_lb.useast1.arn
    weight      = 128
  }
}

# 4. Endpoint Group for ap-south-1
resource "aws_globalaccelerator_endpoint_group" "apsouth1" {
  listener_arn = aws_globalaccelerator_listener.http_listener.id
  endpoint_group_region = "ap-south-1"

  endpoint_configuration {
    endpoint_id = aws_lb.apsouth1.arn
    weight      = 128
  }
}

Replace aws_lb.useast1 with your actual ALB/NLB or EC2 IPs.

---

🔐 Security & Compliance Features

Feature	Description
✅ Static IPs	Use in firewall rules or allowlists
✅ Health Checks	Remove failed endpoints automatically
✅ Regional Failover	Fast routing to healthy endpoints
✅ Traffic Distribution	Weights or failover-based
✅ AWS Shield Integration	Built-in DDoS protection

---

💡 Key Benefits

Feature	Benefit
🌍 Global Static IPs	No need to manage IPs across regions
🛣️ AWS Backbone Routing	Faster and more reliable than public internet
🔄 Automatic Failover	100% uptime without manual intervention
⚡ Latency Optimization	Routes to closest healthy endpoint

---

💰 Pricing (Simplified)

Cost Component	Notes
Accelerator fee	$0.025 per hour
Data transfer out via GA	~$0.015–0.12 per GB depending on region

🧠 Often cheaper than public internet latency, especially for performance-sensitive apps.

---

✅ TL;DR Summary

Feature	Global Accelerator
Level	Network (TCP/UDP)
Static IPs	✅ Yes
Best For	Real-time apps, global APIs, gaming
Traffic Optimization	✅ Fast path via AWS backbone
Failover	✅ Auto failover between endpoints/regions
Terraform Support	✅ Yes (aws_globalaccelerator_*)

---