A system, method, and apparatus are directed towards managing a Voice over IP (VOIP) message over a network, where the VOIP message may employ the Real-time Transport Protocol (RTP) and possibly Session Initiation Protocol (SIP) over the User Datagram Protocol (UDP). The invention enables a VOIP client device, such as an IP phone, and the like, to communicate a message to a local proxy residing on a local computing device. The communications between the VOIP device and the local proxy may employ SIP/RTP over UDP. Upon receipt of the communications, the local proxy converts the transport protocol to another transport protocol, such as Transmission Control Protocol (TCP). The local proxy may also perform a port translation on the message. The converted communications may then be sent to a remote server, where it may be employed in its present SIP/RTP over TCP format, or be converted back to using UDP.