Networks are the heart of communications in most large businesses today. Learning the veins of these systems can help you improve the flow of vital information.
In Part 1 of this six-part series, we discussed using special types of conductors, cables, and hardware to help us get signals to their destinations without distortion or diminishment. This month, we’ll look at computer networks and their components: power sources, conductors, and loads.
A computer network is a collection of electrical signaling circuits, each carrying digital signals between pieces of equipment. A power source is a network device that transmits an electrical signal. The conductors are the wires that the signal travels over to reach its destination (another network device). The receiver is the load. Connected together, these items make up a complete network. We create networks with these elements so individual computers can share information and use programs from other computers. Let’s take a closer look at the fundamental parts of networks.
Connection to the computer. If the network allows other computers to access the information in one computer, the network must have some way of getting into it. It does this through a network interface card (NIC). The NIC is nothing more than a circuit card that fits into one of the expansion slots in the back of a personal computer. The card interprets information between the computer and network and feeds information in and out of the computer. This is the most vital link between the computer processor and the network.
Communication means. Once you have an NIC in place, you need some method to get the data signals from one computer on the network to another. You can use any of the following methods to accomplish this: twisted pair cables; coaxial cables; optical fiber cables; radio waves; infrared light; or electronic signals sent through power lines.
While twisted pair cables are the most common means of communication (or communications media) in a network, this method depends on the high frequencies of these signals.
The types and speeds of the signals you may send through a network reflect the details of that network. You must coordinate all parts of the system to send, carry, and receive the same types of signals. Usually, these details are greater concerns when the different parts come from different vendors. However, if you ever have trouble with your network, check these signals and make sure they are the correct types. Note: We are not concerned with the signals getting from one piece of equipment to another; we’re concerned with the type of signal making its way from one device to another.
Connection pattern. You can use several methods to connect computers together. Some of the most common methods are a ring pattern, star pattern, bus pattern, mesh pattern, the tree structure (a group of stars, connected to a bus), and star ring. We’ll take a closer look at the three most common configurations.
Ring topology. Connecting a group of computers together in a large ring forms a ring topology. We do this with fiber-optic, twisted pair, or coaxial cables. In this arrangement, each computer has two cables connected to it, one from the previous computer and one to the next computer in line. When you draw this arrangement, it looks like a simple ring, but it’s more like a chain run from computer to computer—with a final run back to the first computer (once it reaches the end of the line).
Star topology. A star topology is an arrangement where one computer is at the center of many computers. The computer in the center of the star is the file server or main computer. From this central computer, there are several (exactly how many depends on the type of network and application) cable runs to the various satellite stations (nodes). In actual practice, each computer has only one cable connected to it, with all of these cables feeding into one central location (i.e. the file server).
Bus topology. On a diagram, a bus topology looks as if all of the computers connect to one central cable (i.e. bus). In actuality, two cables connect to most of the computers. The cable “splices through” each computer to make them continuous. However, the computers at the end of the line have only one cable connected to them.
The rise of structured cabling. The idea of categories of data cabling began in the late 1980s. As computer networking quickly rose in importance, and end users needed, send more signals, most networks used a proprietary type of cable. Every manufacturer built and tested their cable differently, so comparing one cable to another was difficult.
Following the widespread acceptance of unshielded twisted pair (UTP) cable, a push to standardize cables and topologies for data applications resulted in what we call Structured Cabling, codified as EIA/TIA 568. Under the original specifications, levels of cable performance allow customers to select the most cost-effective cables for their applications. These levels include:
• Level 1: POTS—Plain Old Telephone Service.
• Level 2: ARCNET—Low speed computer terminal and network applications.
• Level 3: Ethernet and 4/16 MB/s (megabits per second) Token Ring cabling.
In 1992, the industry introduced two more levels, reflecting newer developments in high-speed networks:
• Level 4: For passive 16 MB/s Token Ring. • Level 5: For the copper wire versions of FDDI (Fiber Distributed Digital Interface) at 100 MB/s.
We now have new signal transfer methods. Affecting us are ATM (Asynchronous Transfer Mode), a new way of sending data in packets rather than as a continuous stream; Fast Ethernet; and Gigabit Ethernet. However, the 100 MHz limitation on Cat. 5 cabling is inadequate for the future of UTP.
There are new standards in development for Cat. 6 cable, and Cat. 7 cable. These add an expanded Cat. 5 specification (called 5e), and add Cats. 6 and 7 UTP specifications. Under the new proposals:
• Cat. 5 includes delay skew and powersum NEXT testing.
• Cat. 6 expands the frequency range to 155 MHz and tightens powersum NEXT specifications.
• Cat. 7 expands the range to 200 MHz and tightens powersum NEXT specifications.
Sidebar: Limitations of Copper Wire
Due to the electrical properties of copper wiring, data signals undergo some corruption during their travels. Signal corruption within certain limits is acceptable, but if the electrical properties of the cable cause serious distortion of the signal, replace or repair that cable.
As a signal propagates down a length of cable, it loses some of its energy. So, a signal that starts out with a certain input voltage arrives at the load with a reduced voltage level. The amount of signal loss is attenuation, which we measure in decibels (dB). If the voltage drops too much, the signal may no longer be useful.
Attenuation has a direct relationship with frequency and cable length. The higher the frequency the network uses, the greater the attenuation. Also, the longer the cable, the more energy a signal loses by the time it reaches the load.
A signal loses energy during its travel because of electrical properties at work in the cable. For example, every conductor offers some DC resistance to a current (sometimes called copper losses). The longer the cable, the more resistance it offers.
Aside from copper losses, copper cables have inductive and capacitive reactance. In an inductive reaction, a current’s movement through a cable creates a magnetic field. This field induces a voltage that works against any change in the original current. We measure inductance in Henrys. Capacitance is a property that two wires exhibit when you place them close together. The electrons on the wires act upon each other, creating an electrostatic charge that exists between the two wires. This charge opposes change in a circuit’s voltage. We measure capacitance in farads.
Resistance reduces the amount of signal passing through the wires; it does not alter the signal. Reactance, inductance, or capacitance distorts the signal. Reactance can distort the changes in voltage that signify the ones and zeros in a digital signal.
Sidebar: Cable Impedance
Cables in computer networks are rated by this characteristic. Characteristic impedance refers to the internal signal-transmission characteristics of the cable. However, you cannot measure the characteristic impedance of a cable with an ohmmeter, and the number does not refer to a number of ohms per foot or per hundred feet. You should never replace a cable with one of different characteristic impedance. For example, if you’re running 100-ohm cable, you cannot insert a piece of 150-ohm cable in the network. If you were to do this, the signal would reflect off of the 100 ohm to 150 ohm junction and distort the transmission signal. In short, the link will not pass signal if you mix cables of different impedances. In general, higher impedance cables mean less attenuation of data signals.
Characteristic impedance determines the amount of power able to transfer between devices on the network. Whenever you connect two electrical devices, the power transferring from one to the other maximizes when their impedances are equal. Any amount of impedance mismatch causes some of the power to reflect back into the source device. This is why you should never directly connect cables of differing impedance, and you should use impedance-matching transformers.