Comparison of collision avoidance techniques for efficient voice transmission on Ethernet

Ethernet with standard CSMA/CD (carrier sense multiple access with collision detection) is not suitable for voice due to its high bandwidth requirements and low tolerance to network delays. Collision avoidance approaches offer attractive pay-offs in terms of deterministic delays and increased throughputs. Two such mechanisms, virtual slotting and pseudo-token-pass were previously proposed. A collision avoidance technique, frame sensing, is proposed to improve throughput performance of Ethernet for voice transmission. While virtual slotting and pseudo-token-pass form part of the client layer in the OSI model, additional hardware is required for the frame sensing technique. The performance of the three techniques is compared in terms of access delays and the number of voice channels on the network. All the methods provide a substantial improvement over CSMA/CD. Therefore, collision, avoidance mechanisms make it feasible to use Ethernet efficiently for voice transmission. The result is significant in the light of earlier pessimism regarding the suitability of CSMA/CD as voice/data protocol