The main problem in networking when you are building a solution like this is how to properly load balance the traffic, and when failure happens, how to re-route the traffic to other OpenShift clusters. The main technologies we can use here are the following:
- Anycast IP addresses: In order to effectively load balance the traffic across all our data centers, we can use an anycast IP address. This will help not only load balance the traffic but also provide IP failover if an application becomes unavailable in one of the data centers.
- Application health checks: Application health checks are a must have in this solution. They will help you to identify the failure and reroute the traffic to other data centers.
- Dynamic routing protocols ...