Data Communications

In 1997, for the first time in the history of telecommunications, data carried across the networks was at an equal share as voice. Although voice revenue still exceeds that of data, that will change as more and more information is sent and bought in digital form. Many elements combine to share data with the outside world.

Elements of Digital Communication.

The fundamentals of data communications include:

• Standards
• Architectures
• Protocols
• Error detection
• Plexes
• Multiplexing
• Compression

Standards

The purpose of standards is to help sellers develop programs that will work with one another. Two types of standards exist: de facto and de jure. De facto means “of fact.” If more than one seller builds or complies with a specific technology, we can refer to that technology as a standard. De jure means “of law.”  Sometimes governing or regulatory agencies require specific standards for data transmission.

An example of a de facto standard is Microsoft Windows. Most software providers build programs to work with Windows because most people use Windows.  Ethernet modems follow de jure standards set by CCITT, an international committee that specifies the way modems relay data. 

Architectures

Once information is in digital form, it can be manipulated in an ever-expanding number of ways.  Telecom engineers must develop an overall design when planning a communications environment. Several communications architectures now exist.  IBM’s SNA and DEC’s DNA are the most widely known. The newest and most rapidly growing architecture is TCP/IP protocols as used on the Internet.

            Data architectures originally emulated voice architectures because data transmission initially traveled across voice networks.  Currently, however, many data devices conform to specific data transmission architectures, with which digitized voice signals are compatible.

Protocols

Protocols are key parts of communications architectures. Architectures are guidelines on how environments connect multiple devices in a network. Protocols provide the rules of communications among those devices.

Transmission Protocols (Synchronous versus Asynchronous)

Lower level data communications falls into two categories: synchronous or asynchronous. The words themselves stem from the Greek word chronos, meaning time. Data transmissions are either “in” time (synchronous) or “out of” time (asynchronous). Data communications depend on timing or clocking. Equipment at both ends must coordinate to agree on the beginnings, middles, and ends of transmissions.

            Asynchronous transmissions are clocked one byte at a time, while synchronous transmissions are clocked in groups of bytes. Asynchronous transmissions have a start/stop communication characteristic. Every byte has one bit signaling the start and at least one signaling the stop bits.

            In synchronous transmission, bytes travel in groups of one or more with each group preceded by a minimum of two bits to begin clocking. Clocking is controlled by data communications equipment (DCE).

Error Detection

In data transmission it is critical that data be sent and received in the same form.  There are two steps in error detection:  detecting transmission errors and triggering retransmissions when errors are detected.

            The solution is to send added information (data about data) to corroborate the information received.  The principal methods used to detect data communication errors are:

• Parity bit, or vertical redundancy checking (VRC)
• Longitudinal redundancy checking (LRC)
• Cyclic redundancy checking (CRC)

Using parity bits, the likelihood of detecting an error is about 65 percent. A better method, longitudinal redundancy checking, is about 85 percent accurate. Cyclic redundancy checking methods are more than 99.99% accurate. The following is a brief outline of these different error detection methods.

Parity Bit/Vertical Redundancy Checking (VRC)

The parity bit approach to error detection is simply adding a bit to every byte sent. The parity bit is set to 0 or 1 and calculated by the sending device and recalculated by the receiving device. If the calculations match, the associated byte is accepted, otherwise it is rejected.

Longitudinal Redundancy Checking (LRC)

LRC is similar to VRC, and goes one step further. LRC works by producing a parity bit across the longitudinal track of a byte. In earlier times, LRC had the advantage over CRC of needing fewer computing resources while providing reasonable accuracy.  With more powerful computers, CRC checking is now practical and routinely used.

Cyclic Redundancy Checking (CRC)

CRC comes closest to 100% error detection. CRC relies on a calculation, referred to as Frame Check Sequences, sent immediately after the original block of data bits. The length of the FCS is chosen in advance, based on a designated degree of error detection.

            The FCS results from taking the original data block bit pattern and adding to its end additional binary 0’s. The exact number of 0’s will be the same number as the bits in the needed FCS. The resulting binary number, including the trailing 0’s, is divided by a previously selected divisor (referred to as P in the algorithms).

P has certain characteristics:

• It is always one bit longer than the desired FCS.
• Its first and last bits are always 1.
• It is chosen to be relatively prime to the FCS; that is, P is divided by the FCS would always give a non-0 remainder. In practice, that means P is normally a prime number.
• The division uses binary division, a much quicker and simpler process than decimal division. The remainder of the division becomes the FCS.