SIP/TCP instead of SIP/UDP

However, for real-time delivery of audio and video, TCP and other reliable transport protocols such as XTP are inappropriate. The three main reasons are:

* Reliable transmission is inappropriate for delay-sensitive data such as real-time audio and video. By the time the sender has discovered the missing packet and retransmitted it, at least one round-trip time, likely more, has elapsed. The receiver either has to wait for the retransmission, increasing delay and incurring an audible gap in playout, or discard the retransmitted packet, defeating the TCP mechanism. Standard TCP implementations force the receiver application to wait, so that packet losses would always yield increased delay. Note that a single packet lost repeatedly could drastically increase delay, which would persist at least until the end of talkspurt.
* TCP cannot support multicast.
* The TCP congestion control mechanisms decreases the congestion window when packet losses are detected ("slow start"). Audio and video, on the other hand, have "natural" rates that cannot be suddenly decreased without starving the receiver. For example, standard PCM audio requires 64 kb/s, plus any header overhead, and cannot be delivered in less than that. Video could be more easily throttled simply by slowing the acquisition of frames at the sender when the transmitter's send buffer is full, with the corresponding delay. The correct congestion response for these media is to change the audio/video encoding, video frame rate, or video image size at the transmitter, based, for example, on feedback received through RTCP receiver report packets.

An additional small disadvantage is that the TCP and XTP headers are larger than a UDP header (40 bytes for TCP and XTP 3.6, 32 bytes for XTP 4.0, compared to 8 bytes). Also, these reliable transport protocols do not contain the necessary timestamp and encoding information needed by the receiving application, so that they cannot replace RTP. (They would not need the sequence number as these protocols assure that no losses or reordering takes place.)

While LANs often have sufficient bandwidth and low enough losses not to trigger these problems, TCP does not offer any advantages in that scenario either, except for the recovery from rare packet losses. Even in a LAN with no losses, the TCP slow start mechanism would limit the initial rate of the source for the first few round-trip times.

Our ISP has blocked inbound UPD, so remote extensions do not work.