A network failover apparatus and method for use in a client-server system. The method includes establishing at least a first and further path between a client and a server. The first path connects the server to the client through a first network and a first interface of the client and the further path connects the server to the client through a further network that is separate from the first network and a further interface of the client. The method also includes reaching the server through the first interface, detecting that the server is no longer reachable through the first interface, and identifying the first interface as failed. The method also includes reaching the server through the further interface after the first interface is identified as failed, testing the first interface to determine whether the server is reachable while the server is reachable through the further interface, and reestablishing a connection to the server through the first interface.