About This Architecture

Multi-region NTP Stratum 2 architecture leveraging AWS with on-premises Stratum 1 sources and anycast BGP failover across us-east-1 and eu-west-1. Chrony NTP servers in Auto Scaling Groups behind Network Load Balancers in each region advertise a shared anycast VIP via BGP, enabling clients to reach the nearest healthy endpoint without DNS changes. Transit Gateway and Direct Connect integrate on-premises atomic clock sources and legacy devices, while CloudWatch alarms and Route 53 health checks monitor drift and regional failover. This pattern solves precision timekeeping at scale for distributed workloads—EC2, ECS, EKS, Lambda, and RDS—without single points of failure. Fork and customize this diagram on Diagrams.so to adapt subnets, instance types, or add additional regions.

People also ask

How do I build a highly available multi-region NTP service on AWS that syncs with on-premises atomic clocks and fails over automatically?

This diagram shows a production NTP architecture spanning AWS regions us-east-1 and eu-west-1, with Chrony Stratum 2 servers in Auto Scaling Groups behind Network Load Balancers. Both regions advertise an anycast VIP via BGP, allowing clients to reach the nearest healthy endpoint. On-premises Stratum 1 sources feed via Transit Gateway and Direct Connect, while CloudWatch and Route 53 health checks