Using reservation-virtual-time-CSMA with split-channel reservation for packet voice and data

The carrier sense multiple access with collision detection (CSMA/CD) protocol, originally developed for data transmission is not very effective for the transmission of voice packets due to excessive packet delays at moderate to high loads. It has been shown by C.T. Lea, J.S. Heditch (1987), that the problem of delays can be greatly reduced by using a variation of CSMA called reservation-virtual-time-CSMA (R-VT-CSMA). This protocol takes advantage of the periodicity of voice packets during a talkspurt to reduce the collision probability and clocks out new packets onto the channel based upon their packet arrival time to ensure fairness in delay. An extension of the R-VT-CSMA protocol is proposed using split-channel reservation (W.F. Co., H.T. Mouftah, 1985), to improve channel throughput by splitting the available bandwidth into two channels, one for data transfer and the other for reservation. The advantage of this architecture is that contention resolution is performed in parallel with data transfer which results in a priori contention resolution giving an improvement in throughput.<>