taylor (06/02/82)
It has been my experience that RTS/CTS is not quite as simple as it sounds for flow control. (The following is my understanding, aand I welcome any necessary corrections.) RTS/CTS originated in a different environment than flow control as relates to stat muxes and data transfer. The origin was in multidropped telephone lines. Each drop must not bring up carrier unless there is data to be transmitted (only one drop may have carrier at a time.) If a terminal wants to transmit, it puts up RTS to the modem. When the modem has established carrier, (after typically a couple of hundred milliseconds,) it then provides DTS back to the terminal, which may then start transmitting. Obviously, this method works best for highly buffered, batch-type (and 3270) terminals. As a consequence, MOST asynchronous terminals do not even look at CTS before transmitting. Because RTS/CTS seems like a reasonable type of flow control, this has been adopted by some terminals (and other peripherals) for use with devices like stat muxes. The caveat here is that it works only if both devices recognize the protocol. As an example of the use, Timeplex stat muxes contain a configuration option for eia (RTS/CTS) flow control or in-band flow control (either standard x-on x-off or user defined characters). Steve Taylor NC Educational Computing Service ...!decvax!duke!mcnc!taylor (919) 549-0671