Digital Network Transmission Media - How Computers are ConnectedDescription of digital network transmission media (cables and wireless) and how they transmit digital data between two points.
Irrespective of the format of the signal to be transmitted throughout the network there must be an infrastructure provided to transport the information. There are currently three broad categories that provide this function.
Copper Cables
The simplest form of providing a connection between two devices is to use two insulated copper wires running in parallel. A signal voltage applied to one end will also be present at the other end. The receiver will detect either the voltage or current while the transmitter will control the amplitude of this voltage or current. There are two basic methods used with this type of transmission line Unbalanced and Balanced.
In the unbalanced mode, one of the lines is used as a fixed reference point, (usually 0 volts - earth), while the second line carries the varying voltage or current. This will provide a data link over a maximum distance of only a few metres and with a maximum bit rate of only a few kbps (Kilobits per second). The main problem is that the line carrying the information will pick up extraneous voltages or currents from other similar information-carrying lines and other sources of electromagnetic energy. This is similar to how a radio uses an aerial to pick up a broadcasting station. This is a cumulative effect and eventually, the additional pick-up will corrupt the data signal to a point where it will become unusable. As just mentioned, the data signal will be radiated from the line as it travels from the transmitter to the receiver. This will not only cause interference on adjacent lines but is, of course, a loss of energy (attenuation). Hence, this simple method of connection results in the data information getting smaller and the interference (noise) getting bigger, as it travels along the line.
A performance improvement can be obtained if the line is operated in a balanced mode. In this case, both lines are used to carry the information by using the difference between the voltage (or current) on each line to vary. Neither line is connected to the earth, and the transmitter will cause opposite polarity signals to appear on the lines. Instead of applying +5 volts to one of the lines (as in the previous example), one line is supplied with +2.5 volts and the other is supplied with -2.5 volts. The receiving terminal will still detect the same 5 volts between the ends of the transmission line. However, any external interference source will cause approximately the same noise signal to be picked up by both lines. This will cause the voltage (or current) on both lines at the receiving terminal to rise or fall equally but leave the difference the same. If a noise spike causes both lines to rise by 1 volt, the receiver has +2.5 volts plus 1 volt = +3.5 volts on line one and -2.5 volts plus 1 volt = -1.5 volts on line two. The difference between these two signals is still the correct 5 volts. (3.5 - -1.5 = 5 volts).
Reducing the noise pick up on the lines in this way, both the distance and data rate can be increased. The attenuation remains much the same and the cross-talk from adjacent information-carrying lines will still cause problems. Further improvement can be obtained if the two wires are twisted together instead of running in parallel. In this way, the relative position of each line is constantly being reversed. Any noise picked up by one line will have an identical signal picked up by the second line, once half a twist has been completed since the lines have now taken up each other place. Enclosing the twisted pair(s) inside a screened outer covering can reduce the noise pick-up still further. This consists of an earthed braid or foil layer which blocks external electromagnetic radiation before it can reach the transmission lines. Even if the best quality twisted pair cable is used in conjunction with sophisticated transmitter and receiver equipment, the data rate is still limited to a few Mbps (Megabits per second) over a maximum distance of about 100 metres.
Another type of cable has been developed known as a coaxial transmission line. It is a development from the unbalanced line and the screening techniques described above. If the earthed line in the unbalanced pair is flattened out and formed into a tube, then the signal line can be placed in the centre of this tube. Since the signal line is now enclosed inside an earthed tube, there is little chance of external signals causing interference and energy cannot be radiated from the signal line. Although this type of cable is more difficult to manufacture it can be used with data rates of tens or even hundreds of Mbps and over distances of several kilometres. The main limiting factor is the losses in the insulation required to support the signal line in the centre of the tube. Ideally, the space between the inner and outer conductors should be a vacuum or air. Since neither of these provides any mechanical support, clearly this is not possible. Various insulating materials each with its characteristics provide physical support. The usual trade-off is between losses (attenuation) and robustness. Low-loss cables have very little mechanical strength and may not have sharp bends whereas cables that can be self-supporting or follow sharp bends in ducting, will inevitably have greater losses.
All of the above cable types have their origin in the analogue telephone system and have been adapted for use in data transmission. Many of the characteristics are common for either use and place similar limits on the operation of the system. Firstly, the speed of transmission for an electrical signal along a copper wire is not instantaneous. Typically, the speed is about 200,000 km/s or 2/3 the speed of light. This presents little difficulty for people using a telephone system but for a computer system, this may represent a significant time delay.
Another feature of a transmission line is termination. The electrical signal will move from transmitter to receiver but will reflect towards the transmitter if incorrect termination is encountered. If a bit stream on the route from the transmitting terminal to receiving terminal was to encounter a reflected bit stream travelling in the opposite direction, the result would be catastrophic. Every cable has a characteristic impedance, which is the ohmic value of an infinitely long cable of a specific type. When a finite length of cable is terminated in its characteristic impedance, then it will appear to the electrical signals as an infinitely long cable. Since the signal would take forever to reach an infinite distance, there can be no reflection back along the line. To ensure that there will be no reflections in a digital network; all cables must be of the same type and be correctly terminated. Where there is to be a change in the transmission path, from coaxial to twisted pair for example, then a matching unit must be employed to ensure reflection-free transfer.
Fibre-optic Cables
Light sources and sensors have been used for several decades for information transmission. However, the transmitter and receiver had to be close to each other. An example of such would be the sound system used in cinemas where the "soundtrack" causes a light beam to be intensity modulated before being converted into an electrical signal by a light-sensitive device. The development of optical fibres allowed the light beam to be carried within the glass and hence follow a specific route. The problem with a simple glass fibre is that much of the light energy that enters the cable will "leak" out from the walls and what remains will be attenuated by impurities in the glass. These are overcome, at least in part, by using a glass of extremely high purity and different refractive indices. The central core is surrounded by a second layer such that a boundary is formed where the refractive index will change.
Any light attempting to cross this interface will be bent in such a way that it will return to the centre. In this way, most of the light will remain in the glass fibre. The light which enters along the fibre axis will not attempt to leave but will travel directly to the far end of the line. The light which is slightly off the centre axis may be bent a few times along the cable length and will, therefore, arrive after the direct beam. Light entering at a significant angle will undergo many bends back into the centre and so will arrive at the far end after the two previous cases. Light at a large angle from the central axis may not be bent sufficiently to return it and so will be lost from the cable. This type of fibre-optic transmission is known as the "Multimode Stepped index".
An improved method uses a central core surrounded by several layers each having a different refractive index. This has the effect of bunching the various light beams and so minimises the dispersion of the received light. Due to its complexity, this type of cable is more expensive and will, therefore, be used only if the additional cost can be justified. This type of fibre-optic transmission is known as the "Multimode Graded index". If the core of the optical fibre is reduced in size until its diameter is the same as the wavelength of the light being transmitted, then multiple paths are prevented and so there is no dispersion. However, the wavelength is only 3 to 10 pm and manufacturing down to this scale is very expensive. Again, this additional expense must be justified. This system is known as a "Monomode" and can provide data rates of hundreds of Mbps. Because light is used instead of an electrical signal, the speed of transmission is about 50% faster than a copper cable system.
Radio Systems (Wireless Transmission)
In the case of cable transmission systems (copper or fibre optical), the provision of the physical infrastructure is a major financial outlay and will require ongoing maintenance. It will also provide communication between fixed locations only. By using a radio link, it is possible to communicate between two mobile terminals and requires no expensive cable structure. The main problem is in obtaining the required space in the radio spectrum. The only available bandwidth would be in the UHF and above regions. This only provides "line of sight" communication and can, therefore, be used only over small distances unless many repeaters are employed. The use of communication satellites can overcome this lack of range and bandwidth. The satellite is usually positioned in a geostationary orbit where it can "see" almost half the surface of the earth and can, therefore, link two terminals located within this area. The radio frequency used is in the GHz range where bandwidths of around 500 MHz are available. This is sufficient to provide many high-speed data links over each channel. However, the cost of building and launching these satellites is prohibitive. They are usually "owned" by a consortium of financial institutions, and leased to an operator who will then lease out blocks of capacity which in turn will be sold as individual communication channels. The cost comparison between a satellite link and a cable link depends on the distance between terminals. Short distances are best suited to cable systems while long distances are more economically provided by satellite. However, if mobility is of paramount concern then this can be provided by the satellite, irrespective of cost. Even though the radio wave travels at the speed of light, which is faster than the electrical signals on a cable, the distance to and from the satellite is about 80,000 km and so the time delay is much greater.
Terrestrial radio systems using similar techniques to those of the satellite system are known as microwave links. Large system providers may operate these point-to-point systems and again lease out space to a network provider. The frequencies and bandwidth are comparable to those used by satellites but the distances are much less. Even if high towers located on hilltops are used, the maximum distance for "line of sight" is only a few tens of kilometres. A repeater station can receive and retransmit the signals to give an extended range but require a large capital outlay for each repeater station and will require constant maintenance. These systems usually supplement a cable system rather than replace it and have the advantage of a lower route delay as compared to a comparable copper cable.
The cellular radio networks provided for mobile telephone use can also be used as data links. In this case, many base stations must be provided and the same frequencies used in such a way that interference does not occur. However, the frequency allocation costs the system provider a massive amount and then the base stations must be equipped and maintained. The cost of even a very low data rate channel over this system will remain high when compared to other systems. It may have occasional use where a mobile terminal is required and the low data rate can be tolerated. There are frequencies available for very low-power systems similar to the cellular mobile structure, which can be used within an office building. It is never a cheap solution but may be required where cabling is not practical. Due to the very low power, the range is only a few tens of metres and will not extend much beyond the site. Such systems are also prone to interference from outside sources and may not always be reliable.
The latest radio system used for data communication uses frequencies in the 2.4 GHz range and is capable of transmitting and receiving packets of data at 11 Mbps and does not require a line of sight, although anything in between the source and destination will interfere with the signal. The range is in the region of 20 metres, so is only practical for small-area networks unless a large number of repeaters are used.
