Data Transmission Characteristics

Number of Bits Transferred at One Time

Bit Parallel Data Transfer (more than one bit transferred from sender to receiver at one time). Bit parallel data transfer is typically used inside the computer and for peripherals in the "machine room." This type of data transfer involves a large number of data "wires" and transfers all bits in a data unit in one clock-cycle, so it delivers a high potential for fast transfer of information.

Because bit parallel data transfer typically uses thick cables, these cables are relatively expensive to acquire and install. Additionally, such an installation is sensitive to electronic interference and crosstalk. (Crosstalk occurs when signals on one wire affect the signals on another wire to an extent where the output cannot be accurately determined by the receiver—essentially, self-interference within the cable.) Most manufacturers place stringent limitations on the length of parallel cables.

In addition to the data wires, there are usually additional wires that provide for transferring a clock and other control signals. There must also be a "signal ground" wire to provide a route for completion of an electrical circuit between the source and destination.

Bit Serial Data Transfer (one bit transferred from sender to receiver at a time). Bit serial data transfer is most typically used for "remote" devices. This type of data transfer uses relatively thin cables and has slower speeds, which enable it to recover well from outside interference. Depending on the way that data are encoded as a signal, this type of transfer can be used over extended distances.

Serial cables also require a signal ground wire and other control wires. Transmitter and receiver and sender and receiver are, in reality, just another set of terms that mean the same as source and destination, respectively. In a minimal implementation, however, some suppliers implement a two-wire interface for a two-way (called "full duplex") circuit using the real (or "building") ground as a return path, which works as long as the two devices are quite close to each other.

Parallel-Serial Conversions. The device that makes it possible for computers (which are internally parallel devices) to communicate serially is typified by the Universal Synchronous Asynchronous Receiver Transmitter (USART). In the early days of computer-to-computer communications, USART was a large, complex electronic circuit, but it has since been reduced to a single hardware chip that is available in several "flavors."

Signaling

Analog versus Digital Transmission. Historically, telegraphy (the digital form of message communication) did not have the widespread penetration into research and education facilities and businesses that the telephone and its analog transmission network did. Although both the telephone and the telegraph have contributed to the data communications field, it was the telephone network that provided the early transmission media. Because digital transmission has significant advantages over analog, the telephone networks are adopting digital techniques. An example is ISDN, which provides digital service over existing telephone local loop circuits.

Data Flow

Simplex, Half Duplex, and Full Duplex. Digital circuits and some analog circuits cannot support simultaneous data transmission in both directions on the same wires. For ease of reference, the communications industry has defined the terms simplex, half duplex, and full duplex to describe both the circuit's capabilities and the communication subsystem's transmission requirement. A simplex circuit is capable of carrying data in only one direction at all times. A duplex circuit is capable of carrying data in two directions: a full-duplex circuit can handle both directions simultaneously, whereas a half-duplex circuit can operate in both directions, but in only one direction at a time.

Synchronization

No matter what technique of data transmission is used, the receiver must synchronize itself with the transmitter. The receiver needs to know when to pay attention to the signal (because the signal is stable) and when to ignore it (because the signal is changing). Inside the computer, a clock signal is sent on a separate circuit and all transmissions are synchronized with that signal. When transferring data between two devices, each has been set up to operate at a fixed speed. The reality of electrical utilities and manufacturing techniques is that two devices set to operate at the same speed will actually not be operating at precisely the same speed. In fact, the probability is that neither is absolutely correct, except by accident. Two modes of handling this issue are used: asynchronous transmission and synchronous transmission.

Asynchronous Transmission. In asynchronous transmission, the assumption is made that the two devices are relatively close to each other in speed and that if only a small burst of data is sent at a time, synchronization will not be a problem so long as the receiver gets a chance to synchronize with the beginning of the data burst. The amount of data normally transmitted in one burst is an octet, which is preceded by a "start bit" to permit the receiver to "mini-synchronize" to the transmission. It is followed by one to two "stop bit" time periods of idle state to allow the receiver to store the data and check for framing and/or parity errors. The large majority of computer-to-terminal hookups are asynchronous.

Synchronous Transmission. In synchronous transmission, a larger amount of data is sent in each block of data and the block contains or is preceded or accompanied by signaling, which permits accurate co-synchronization of the devices. Nearly all network backbone traffic is synchronous.