How to organize a fault-tolerant application?

Question

High availability

I'm trying to find information on the organization of fault-tolerant services. I don’t want to reinvent my own bicycles, I assume that similar tasks have already been solved, but I have little idea in which direction to dig. I will try to briefly describe the problem.
There are a couple of servers on a windows server (let there be A and B), an application that collects data from some device U and sends it to subscribers. Servers A and B are physically distant, device U is located next to server B.
In normal mode, the application runs on server A and collects data from U.
If the connection between A and B is broken, you need to run the application on server B (switching subscribers is a separate issue), and when the connection is restored, terminate it.

Answer 1

It is not clear for which application you want to organize fault tolerance, so look towards Falovier Cluster or NLB Cluster .