A simulation-based comparison of voice transmission on CSMA/CD networks and on token buses

Digitized speech can be transmitted over a variety of digital media. An interesting choice is the use of a Local-Area Network (LAN), for which digitized speech is packetized at the transmitter and depacketized at the receiver. Many local-area networks exhibit good throughput but poor delay characteristics; variable or excessive transmission delay can become noticeable and objectionable to the users of such a voice system. A number of simulations were performed to assess the delay characteristics of a Carrier Sense Multiple Access/Collision Detection (CSMA/CD) LAN and of a similar token bus LAN. A comparison of the results shows that the token bus performs somewhat better. The CSMA/CD LAN's performance was characterized by carrying voice well until a point of collapse is reached; the token bus's performance degraded more continuously. In either case, throughput close to the theoretical capacity of the LAN was found achievable with appropriate techniques.