This article throws light upon the top six methods of data transmission. The methods are: 1. Twisted Pair 2. Shielded Twisted Pair 3. Baseband Coaxial Cable 4. Broadband Coaxial Cable 5. Fibre Optics 6. IEEE Standards.
Method # 1. Twisted Pair:
As must be obvious from the example, transferring data by copying it onto magnetic tape or floppy disk or compressed discs and transporting these tapes and/or disks by truck, the bandwidth of such a form of data transmission (one should really call this “data transportation“) is very good.
If the data are transported over short distances, the bandwidth can exceed several Gigabits per second for magnetic media, which is considerably superior to the bandwidth achieved in Asynchronous Transmission Modes (ATM), possibly the fastest standard method of data transmission available today.
However, the minimum time involved with such data transmission must be in minutes, if not in hours, whereas in communication technology we are accustomed to thinking in terms of microseconds.
Therefore, we have to look at other transmission media that can directly connect two or more computers together. The most common example of such a transmission medium is a twisted pair. A twisted pair is not a criminal-minded couple, as some of you might be tempted to think, but simply two copper wires twisted together. This twisting reduces the electrical interference from similar twisted pairs close by.
Twisted pairs are very widely utilised as telephone cables. They can carry data over long distances (sometimes over several kilometers) without requiring that the data or signal be amplified. In case of telephone cables, the lines coming out from a closely populated area, there are bound to be several such twisted pairs close by. These do not interfere with each other because of the twisting.
These twisted pairs can carry analogue as well as digital signals and over short distances they can usually achieve bandwidths of up to about 10 Mbps (megabits per second). In any case, the bandwidth that they can carry is a function of the diameter of the wire and the distance over which the data are transmitted.
These twisted pairs are available in many categories but Category 3 (or Cat 3 as the computer buffs and wiring mechanics usually call them—their vocabulary can be as twisted as the cables that they use) twisted pair is the category, most widely used in computer networks.
More recently, another category called Category 5 (or Cat 5, according to the twisted vocabulary of our computer buffs) is being increasingly utilised. Cat 5 wires have more twists per unit length than Cat 3 twisted pair and are also Teflon insulated. These differences improve their bandwidth capability as well as reduce the amount of cross-talk and interference.
Twisted pairs are classified under the heading of Unshielded Twisted Pair (UTP), although for the Cat 5 twisted pair, this does seem a misnomer since they are Teflon-coated.
More recently, other categories of Twisted Pairs have been developed (though not yet been given official approval through them being recognised by standards institutions). Table 6.1 gives the specifications of the existing standards and those proposed by their manufacturers.
Table 6.1 gives a list of existing as well as proposed specifications of various UTP cables. However, the ones widely in use today are described below and their details are given in Table 6.2 given later.
Method # 2. Shielded Twisted Pair:
These cables are twisted bulky, expensive and, as the name implies, they are shielded. Introduced by IBM in the 80s, they never really caught on and although some are still in use, they are invariably being used in some IBM installations.
But they have never proved very popular. Their bandwidth characteristic is not much better than that of UTP cabling and most users feel that they are both bulky as well as expensive. They have, therefore, practically gone out of the market.
Method # 3. Baseband Coaxial Cable:
Coaxial cables are usually of two types—baseband and broadband. Baseband cables generally come in two varieties, the “50-Ω cable” and the “100-Ω cable”.
The 50-Ω cable is normally used for digital transmission, whereas the 100-Ω cable is normally used for analogue transmission. The terms 50-Ω and 100-Ω refer to the impedance of the cable. The cable consists of a copper cable at the core surrounded by insulation material.
This is surrounded by a conductor, usually in the form of a mesh, and is finally surrounded by a layer of insulation. The computer buff and the wiring mechanic usually refer to this as “co-ax” (although this cable is not twisted, their terminology continues to be “twisted“).
The bandwidth available on such cables is dependent on their length; for distances of up to about 2 km, the bandwidth available can be in excess of 1 Gbps. With increasing distances the bandwidth drops down, often quite sharply.
These cables were quite popular in telephone systems, particularly over long distances.
However, with the increasing acceptance of fibre optic cables and their excellent characteristics, coaxial cables are gradually being replaced by fibre optic cables, in spite of their high cost. Coaxial cables, however, are still widely used for analogue transmission, such as cable TV. Some LANs also continue to use these, but in general, their use is shrinking.
Method # 4. Broadband Coaxial Cable:
Broadband coaxial cables are generally used in analogue circuits and although it is sometimes used in computer networks, these cases are usually rare. Since they typically transmit analogue signals, they need to have devices at either end to convert analogue signals to digital signals and vice versa. The bandwidth allowable also depends on these modulators/demodulators.
Since they normally transmit over long distances, they need amplification devices in between. In computer networks, however, this poses a problem, since this amplification device implies that transmission must be unidirectional. This means that if a computer is downstream and wishes to transmit to a computer upstream, it normally cannot do so.
In order to overcome this problem, two cables are used, one to receive data from and one to send data on.
Both the cables are parallel and transmission takes place to the head end of the transmitting cable or the cable that is used to send data on. If the transmission is meant for a computer that is upstream of the sending computer, then the head end transfers the signal to the other cable, which then carries the data on to the upstream computer.
This is usually referred to as a dual cable system. There is another method of achieving this; that is by means of frequency allocation. In this, a lower frequency is used for transmission to the head end, whereas a higher frequency is used for transmission from the head end. This dual cable system is illustrated in Fig. 6.1.
However, since we are only concerned with computer networks, we shall not dwell on the applications to which Broadband coaxial cables are put (in TV transmission, for example). Suffice it to say that they can be used in computer networks, but are not widely used.
Method # 5. Fibre Optics:
While the Information Technology industry takes pride in the rapid technological improvement—and justifiably so—they often forget that the technological improvements in the communications industry, to which they are inexorably tied, far outstrip their efforts to the extent that modern communications technologists tend to think of computers as slow devices.
While the speed of operations of computers have gone up ten times in the last twenty years (compare the speed of a CDC6600 with that of a CRAY3 or even with that of a PAR AM 10000), the bandwidth available for transmission have gone up by a hundred times, in the same period (compare the speed of 56 kbps for the ARPANET with that of 1 Gbps available in modern communications) and this issue becomes absolutely certain.
This limit of 1 Gbps is by no means the limit reached; 100 Gbps speeds have been achieved in the laboratory.
This limit of 1 Gbps has been imposed because we are unable, at the moment; to convert electrical signals to optical signals and vice versa, any faster. But the horizon seems unending and we are looking at the possibility of almost limitless transmission speeds (of the order of several Terabits per second) and almost zero transmission errors.
Transmission errors, which were already very low (of the order of 1 bit in 10-5) have since then significantly reduced. All this has been possible because of the introduction of technology to handle light in transmission technology. This has led to the development of Fibre Optics.
Any Fibre Optic transmission requires three essential elements. These are a light source, a transmission medium and a detector. Typically, a pulse of light indicates a 1 and the absence of light indicates a 0.
The transmission medium is an ultra-thin glass fibre. By attaching a light source at one end and a detector at the other end of a medium, transmission of data can be achieved over this medium. This unidirectional signal is transmitted after receiving an electrical signal and converting it to a light pulse.
This light pulse is reconverted into an electrical signal at the receiving end. The light ray does not get dissipated because of the proper use of the principle of refraction. When light travels from one medium to another, it gets bent at the boundary of the two layers, according to the laws of refraction.
The amount of refraction depends on the index of refraction of the boundary between the two media. Thus, if the principle is utilised properly, between air and the glass fibre media, because of the light ray being bent, it can be bent back into the glass fibre thereby not being lost. It can, therefore, be transmitted for mile after mile without virtually any loss or dissipation.
This fact is very helpful and it has made the transmission of data by light pulses a practical reality. Currently, speeds of up to 30 GBps have been achieved for distances up to 30 km, but in laboratories, with the use of lasers, distances of up to 100 km are easily possible, without any intermediate repeaters. Further research continues and greater breakthroughs are bound to come soon.
At the present rate of development, communications technology researchers have left Information Technology researchers far behind. All this may be very interesting, but I fear we have to leave it behind. Readers who are interested in this technology are advised to read T.
Fibre Optics can be utilised in LANs as well as for communications over long distances and, in fact, they are being used in long-distance telephone communications. One advantage of using Fibre Optic-cabling that is obvious is that tapping into a Fibre Optic transmission line is far more difficult than on a line based on Ethernet.
This may also be a disadvantage, because in case of networks where mixed connections are required, the interfacing can become more difficult. For this, either a passive interface or an active repeater can be used. Using a Fibre Optic LAN, a ring topology or a star topology can both be used. It has several obvious advantages over copper wires.
1. Fibre Optics has much higher bandwidth than copper cables.
2. Due to low attenuation, it requires much fewer repeaters than copper cables—about 15% repeaters compared to copper cabling. This is because in the best copper cabling, repeaters will normally be required every 5 km, whereas in case of Fibre Optics, repeaters are required every 30 km.
3. Fibre Optics is not affected by atmospheric conditions and surroundings, such as corrosive chemicals and dampness.
4. Copper cables weigh much more than Fibre Optic and over somewhat longer distances this poses additional problems for the company laying down these lines. A 1-km length of Fibre Optic will weigh about 100 kg or about 10% of an equivalent copper cable that will weigh close to about 1000 kg.
5. Fibre Optics is not affected by electromagnetic fields or power surges.
Fibre Optic cables are very similar to coaxial cables. In the centre is a glass core (actually quartz) through which the light pulses propagate. This is surrounded by a glass cladding, the whole thing being encased in a plastic sheath. This is illustrated in the cross-section shown in Fig. 6.2.
The cable has two ends, each with a different equipment, performing different functions. The sending end obviously has a light-emitting source and the receiving end, a decoding source. Decoding is done by using a photodiode, which emits an electrical pulse every time a pulse of light strikes it.
Thermal noise may also be present and in order to overcome any problems from this, the light source has to be sufficiently powerful. This, incidentally, also reduces the error level.
Because of the low-error characteristics, longevity and ease of use, Fibre Optic cables are becoming more and more popular, in spite of their high cost. They also provide high security, since they cannot be tapped into and are not affected by electromagnetic fields.
They are also not affected by corrosion. Signal losses are few and repeaters are required at some 30 km intervals (compared to a maximum of 5 km for copper wire).
Their cost continues to come down as more and more research breakthroughs occur and it is likely that in the not too distant future, they will become standard for all network applications, certainly for applications in which distances of more than a few metres are involved.
Method # 6. IEEE Standards:
IEEE has established certain standards for LAN and MAN transmission protocols and networks. These standards are known as IEEE 802 standards. Of these, 802.1 gives an introduction to the set of standards and defines the interfaces over which these standards are to be applied. ISO has also incorporated these standards as 8002 standards. 802.2 describes part of the “Data Link Layer”.
The 802.3 standard refers to Ethernet. This remark needs some clarifications. While 802.3 is a protocol to which Ethernet also belongs and is sometimes used interchangeably with Ethernet, that is an error because 802.3 protocol also includes other networks.
Incidentally, Ethernet was developed by Xerox to connect a certain number of computers in Hawaii and later on adopted by the networking world as a de facto standard. But this 802.3 standard deals with Carrier Sense Multiple Access (CSMA) and Collision Detection (CD) protocols.
Nowadays most people refer to the cabling in a LAN as Ethernet rather than mention the necessary protocols. While several types of cables are used in such LANs, the 10Base5 cables are commonly known as thick Ethernet. Vampire taps are generally used to connect Ethernet cables. These cables are coaxial, use baseband signalling, operate at 10 Mbps and the signal fidelity is good for 500 metres (hence the name 10Base5).
Similarly 10Base2 cables, also known as thin Ethernet cables that use BNC connectors are good only for 200 metres. Of course, both thick Ethernet and thin Ethernet have limitations on the number of nodes that can be attached to each segment of the cable.
While thin Ethernet is much cheaper, only up to 30 nodes can be attached in each segment, while in thick Ethernet, which is more expensive, up to 100 nodes can be attached to each segment.
These cables, however, suffer from the problem of detection of faults and cable breaks. To avoid this detection problem, the concept of a hub is used. In this, the cable from each machine is attached to a central hub. For this an even cheaper cable is used.
These cables are called 10Base-T or twisted pairs. However, in their case, the distance limitations are even more severe and each segment has to be limited to 100 metres.
Finally, we have 10Base-F or fibre optic cables. Admittedly, they are the most expensive, but performance-wise, they are also far superior. In fibre optic cables, you may attach up to 1024 nodes and a segment can easily extend up to 2 km. The details regarding each type of cable are summarized in Table 6.2.