There are a few implementations of this approach such as Apache or nginx web servers, however in their implementations they choose the least loaded server and this algorithm could not be precise if there are more then one server configuration in cluster. In this case it could cause overloading of some servers while the other ones will have some spare resources.