The Real-time Transport Protocol (RTP) is an Internet protocol standard used to conduct real-time multimedia unicast and multicast communications. It consists of two components: the transport protocol and the Real-time Transport Control Protocol (RTCP). The former provides the Internet Protocol (IP) specifications to transmit multimedia streams across networks in real-time. The latter provides basic session management and Quality of Service (QoS) capabilities such as looking out for data packet loss and compensating for transmission delays. Commonly used in Voice over Internet Protocol (VoIP) telecommunications, Real-time Transport Protocol was originally developed by the Internet Engineering Task Force's Audio-Video Working Group to provide a means of conducting real-time videoconferencing between multiple participants in geographically dispersed locations.
Audio and video data streams are transmitted separately in RTP. Separate RTP and RTCP packets are transmitted for each using two different communications ports and/or multicast addresses. Participants are thus able to choose to receive only one medium. Synchronized playback of both audio and video is achieved by making use of timing information in the RTCP packets for both audio and video sessions.
The Real-time Transport Protocol header describes how the codec bit streams are assembled into packets. It also contains the instructions that enable receiving network devices to reconstruct the data packets. Other components of RTP include the following: frame identification, which marks the start and end of each frame; intramedia synchronization, which uses timestamps to detect and compensate for delay jitter; and payload identification, which describes the media encoding method so that adjustments can be made for variations in bandwidth.
Also part of the Real-time Transport Protocol are a sequence number to detect lost packets and a source identification. Components of RTCP include identification that includes participants' names, email addresses, telephone numbers, and intermedia synchronization, which enable the transmission of separate audio and video streams. Session control enables participants to indicate they are leaving a session while quality of service (QoS) feedback keeps track of the number of lost packets; round-trip transmission time, and jitter, enables the source to adjust data rates as required.
Though it does provide basic monitoring capabilities to assure QoS, RTP does not guarantee real-time delivery of multimedia communications; nor does RTP assure other QoS parameters such as packets being received in the correct order. It relies on Internet protocols in the Network and Transport layers of the Open Systems Interconnection (OSI) Model to do so. RTP generally runs on top of the User Datagram Protocol (UDP), although other transport protocols, including Session initiation Protocol (SIP) and H.323, can be used as well.