Two prerequisites must be satisfied to have successful communication. The first prerequisite is understandability. The transmitter and receiver must speak the same language. It doesn't matter how big or how clean a pipe you have between the two endpoints. If they're not speaking the same language, you will not be able to understand the message. In the case of data communications, we've resolved these issues quite elegantly: We have software and hardware translation devices that can convert between the different languages that individual computing systems speak. In the realm of human communications, we're about to embark on that exciting journey as well. Through the use of advanced voice-processing systems, in the next five to seven years we should have the ability to do real-time foreign language translation as part of the network service.
The second prerequisite is the capability to detect errors as they occur and to have some procedure for resolving those errors. In the case of human communications, intelligent terminals at either endhuman beingscan detect noise that may have affected a transmission and request a retransmission, thereby correcting for that error. In the case of data devices, similar logic has to be built in to end devices so that they can detect errors and request a retransmission in order to correct for the errors.
If these two prerequisitesunderstandability and error controlare met, then communication can occur. We communicate by using data devices over what is generically termed a transmission line. There are five main types of transmission linescircuits, channels, lines, trunks, and virtual circuitseach of which has a specific meaning. The following sections describe each of these types of transmission lines in detail.
A circuit is the physical path that runs between two or more points. It terminates on a port (that is, a point of electrical or optical interface), and that port can be in a host computer (that is, a switching device used to establish connections), on a multiplexer, or in another device, as discussed later in this chapter.
In and of itself, a circuit does not define the number of simultaneous conversations that can be carried; that is a function of the type of circuit it is. For example, a simple, traditional telephone circuit is designed to carry just one conversation over one physical pathway. However, converting that to a digital circuit gives you the ability to extract or derive multiple channels over that circuit, subsequently facilitating multiple simultaneous conversations. So, the circuit is the measure of the physical entity.
There are two types of circuits: two-wire circuits and four-wire circuits.
A two-wire circuit has two insulated electrical conductors. One wire is used for transmission of the information. The other wire acts as the return path to complete the electrical circuit. Two-wire circuits are generally deployed in the analog local loop, which is the last mile between the subscriber and the subscriber's first point of access into the network. Figure 2.1 shows an example of a two-wire circuit.
Two-Wire and Four-Wire Versus Two-Pair and Four-Pair
Don't confuse the terms two-wire circuit and four-wire circuit with the terms two-pair and four-pair. Two-pair and four-pair refer to the number of wires in the internal cabling plan. Two-wire and four-wire have to do with the number of electrical conductors associated with a transmission line.
Figure 2.1 A two-wire circuit
Figure 2.2 A four-wire circuit
A four-wire circuit has two pairs of conductors. That is, it has two sets of one-way transmission paths: one path for each direction and a complementary path to complete the electrical circuit (see Figure 2.2). Four-wire circuits are used where there is distance between the termination points which requires that the signal be strengthened periodically. So, for example, four-wire circuits connect the various switches that make up the public switched telephone network (PSTN). Four-wire circuits are also used with leased lines, where a customer may be connecting locations of its own that are separated by distance. Also, all digital circuits are provisioned on a four-wire basis.
There are two types of four-wire circuits: physical four-wire and logical four-wire. In physical four-wire you can actually count four wires. In logical four-wire, physically there are only two wires, but you derive the four individual paths by splitting the frequency. Half of the frequency band carries the transmit signal, and the other half carries the receive signal. So you can't always tell just by looking what kind of circuit you're dealing with; the application dictates the type of circuit it is.
Using Two-Wire and Four-Wire Circuits
Whenever you release energy into space, it loses power as it's traveling over a distance. So, because networks were designed to carry communications over a distance, we need tools to augment signals that have been losing power as they have traveled across the network, which are called attenuated signals. These tools are called amplifiers and repeaters. An amplifier boosts an attenuated signal back up to its original power level so it can continue to make its way across the network. The PSTN traditionally used copper wires. Based on how quickly the signals flow through the copper wires, there's a certain distance requirement between amplifiers. The distance requirement between amplifiers is relatively short on copper wiresgenerally about 6,000 feet (1,800 meters). As networks were built, these distance considerations were kept in mind. (Repeaters are discussed later in this chapter, in the section "Digital Transmission.")
Network builders had to give some thought to another aspect of amplifiers: First-generation amplifiers were unidirectional. They could only amplify a signal moving in one direction, so any time you needed to provision a circuit that was going to be crossing a distance, you had to literally provision two circuitsone to amplify the information in the transmit direction and a second to amplify the information in the receive direction. Therefore, whenever a network was crossing a distance, it needed to use a four-wire circuit. But in building out the millions of local loops for subscribers, it was seen as being cost-effective to have to pull only two wires into every home rather than four. Therefore, the local loops were intentionally engineered to be very short; some 70% to 80% of the local loops worldwide are less than 2 miles (3.2 kilometers) long. Because the local loops are short, they don't need amplifiers, and therefore the subscriber access service can be provisioned over a two-wire circuit. However, the local loop is increasingly being digitalized, so as we migrate to an end-to-end digital environment, everything becomes four-wire. Figure 2.3 shows an example of a segment of a network in which two- and four-wire circuits are traditionally used.
Figure 2.3 Using two-wire and four-wire circuits
A channel defines a logical coversation path. It is the frequency band, time slot, or wavelength (also referred to as lambda) over which a single conversation flows. A channel is a child of the digital age because digital facilities enable multiple channels. The number of channels on a transmission line determines the number of simultaneous conversations that can be supported. Because we are becoming more digitalized all the time, you often hear people refer to the number of channels rather than the number of circuits.
Lines and Trunks
Lines and trunks are basically the same thing, but they're used in different situations. A line is a connection that is configured to support a normal calling load generated by one individual. A trunk is a circuit that is configured to support the calling loads generated by a group of users; it is the transmission facility that ties together switching systems. A switching system is a device that connects two transmission lines together. There are two major categories of switching systems:
CPE switchesThe most prevalent form of switch in the customer premises equipment (CPE) environment is the private branch exchange (PBX), which is called a private automatic branch exchange (PABX) in some parts of the world. A PBX is used to establish connections between two points. It establishes connections between telephones that are internal to the organization, and it establishes connections between internal extensions and the outside world (that is, the PSTN).
Network switchesA hierarchy of network switches has evolved over time, and the appropriate switch is called into action, depending on which two points the switches are connecting together. For example, in Figure 2.4 the CPE is on the left-hand side. Each individual single-line instrument represents a subscriber line. (Again, the fact that it's called a line means that it's a circuit configured to carry the calling load of just one user.) Above the single-line instrument is a business enterprise with a PBX. The connection from this PBX to the PSTN occurs over a trunk that is specifically configured to carry the calling load of multiple users. Beyond the PBX are multiple end users that are attached to that PBX. Each end user's connection would be referred to as a station line, again emphasizing that the line is carrying the calling load of one user.
Figure 2.4 Lines, trunks, and switches
The customer environment attaches to the PSTN, and the first point of access is the local exchange, which is also referred to as a Class 5 office (and in North America, as a central office). The traditional local exchange switch can handle one or more exchanges, with each exchange capable of handling up to 10,000 subscriber lines, numbered 0000 to 9999. The only kind of call that a local exchange can complete on its own, without touching any of the other switches in the network, is to another number in that same local exchange. Local exchanges are discussed in detail in Chapter 5, "The PSTN."
For a local exchange to call a neighbor that resides 10 miles (16 kilometers) away and who draws a dial tone from a different local exchange, the connection between those two different exchanges is accomplished through the second part of the hierarchya tandem switch (also called a junction exchange). The tandem switch is used to connect local exchanges throughout the metropolitan area. When it's time to make a toll call, one that is long-distance in nature, another switching center is called into actionthe toll center (also called the Class 4 office, transit switch, or trunk exchange). The toll center is responsible for establishing and completing national, long-distance communications.
The top of the hierarchy is the international gateway, whose exchanges are specifically designed to connect calls between different countries.
A trunk supplies the connections between the numerous switches within the PSTN, between customer-owned switches such as the PBX, and between the PBXs and the PSTN. On the other hand, a line supports a single user in the form of a subscriber line in the PSTN or an extension provisioned from the PBX. (Chapter 5 describes in detail the entities involved in managing local, tandem, and toll exchanges.)
Today, because of the great interest in and increased use of packet switching, most networks use virtual circuits. Unlike a physical circuit, which terminates on specific physical ports, a virtual circuit is a series of logical connections between sending and receiving devices (see Figure 2.5). The virtual circuit is a connection between two devices that acts as though it's a direct connection, but it may, in fact, be composed of a variety of different routes. These connections are defined by table entries inside the switch. A connection is established after both devices exchange agreement on communications parameters that are important to establishing and maintaining the connection and on providing the proper performance for the application they are supporting. The types of communication parameters that could be included are message size, the path to be taken, how to deal with acknowledgements in the event of errors, flow-control procedures, and error-control procedures. The term virtual circuit is largely used to describe connections between two hosts in a packet-switching network, where the two hosts can communicate as though they have a dedicated connection, although the packets may be taking very different routes to arrive at their destination.
Figure 2.5 A virtual circuit
There are two types of virtual circuits: permanent virtual circuits (PVCs) and switched virtual circuits (SVCs). The vast majority of implementations today involve PVCs. PVCs and SVCs are commonly used in packet-switching networks (for example, X.25, Frame Relay, ATM).
A PVC is a virtual circuit that is permanently available; that is, the connection always exists between the two locations or two devices in question. A PVC is manually configured by a network management system, and it remains in place until the user reconfigures the network. Its use is analogous to the use of a dedicated private line because it provides an always-on condition between two locations or two devices.
In contrast to PVCs, SVCs are set up on demand. They are provisioned dynamically by using signaling techniques. An SVC must be reestablished each time data is to be sent, and after the data has been sent, the SVC disappears. An SVC is therefore analogous to a dialup connection in the PSTN. The main benefit of an SVC is that you can use it to access the network from anyplace. The predominant application for SVCs is to accommodate people who are working at home, in a hotel, at an airport, or otherwise outside the physical location of the enterprise network.