If any application is scanning the ports of other machines, to find out whether any particular service/application is running, will it be considered harmful? Is this treated as hacking?

How else can one find out on which port the desired application is running (without the user input)?

Let's say I only know the port range in which the other application could be running, but not the exact port. In this case, my application ping each of the port in range to check whether the other application is listening on it, using already defined protocol. Is this a normal design? Or is this considered harmful for the security?