Trying to find information on the organization of fault-tolerant services. I do not want to reinvent their bikes, I assume that a similar problem has been solved, but little idea which direction to dig. I will try to briefly describe the task.
There are a couple of servers on the windows server (let them be A and b), an application that gathers data from a device U and sends them to subscribers. Servers A and b are physically separated, the device U is located near the server B.
In normal mode, the application runs on the server and collects data from U.
When a disconnection between A and b need to launch the application on the server (switching subscribers - a separate issue), and when connection is restored to complete it.