A study of asynchronous time division multiplexing for time-sharing computer systems

In order to reduce the communications costs in time-sharing systems and multicomputer communication systems, multiplexing techniques have been introduced to increase channel utilization. A commonly used technique is Synchronous Time Division Multiplexing (STDM). In Synchronous Time Division Multiplexing, for example, consider the transmission of messages from terminals to computer, each terminal is assigned a fixed time duration. After one user's time duration has elapsed, the channel is switched to another user. With synchronous operation, buffering is limited to one character per user line, and addressing is usually not required. The STDM technique, however, has certain disadvantages. As shown in Figure 1, it is inefficient in capacity and cost to permanently assign a segment of bandwidth that is utilized only for a portion of the time. A more flexible system that efficiently uses the transmission facility on an "instantaneous time-shared" basis could be used instead. The objective would be to switch from one user to another user whenever the one user is idle, and to asynchronously time multiplex the data. With such an arrangement, each user would be granted access to the channel only when he has a message to transmit. This is known as an Asynchronous Time Division Multiplexing System (ATDM). A segment of a typical ATDM data stream is shown in Figure 2. The crucial attributes of such a multiplexing technique are: 1. An address is required for each transmitted message, and 2. Buffering is required to handle the random message arrivals.