Leased Lines

Leased lines have been used by businesses for many years to establish point-to-point, dedicated connections. These lines are generally leased from the provider, hence the name, and provide a fixed bandwidth for a fixed cost. Connecting the telephone system from a branch office to the main office, for example, can be more cost-effective if you purchase a leased line from the telephone company and pay a flat fee for its use rather than using the normal long-distance network. If your business environment already uses leased lines for its telephone systems, using those same lines to connect a computer network is a logical step up.

Basically, leased lines provide a point-to-point connection that represents a permanent circuit that you lease and do not have to share with others. The leased line might consist of a physical line that traverses the entire length of the connection from end to end, or it might be composed of connections at both ends to the local exchange carrier, with the two exchanges connected by some other technology.

Because of the point-to-point topology of leased lines, no call setup is required at either end. The connection exists and you can use it whenever you need tothat's why it's often referred to as "always on" technology. The physical lines also are specially conditioned by the carrier to minimize errors as compared to an ordinary connection to the local exchange.

The first type of leased line was based on analog technology, just as the voice telephone network was, and often used modems at each end of the connection. Digital leased lines now provide connections of up to 56/64Kbps and use a channel service unit (CSU) device and a data service unit (DSU) device, which also are used on other digital lines, such as T1 and fractional T1 services. Both a CSU and a DSU are commonly implemented in the same device (CSU/DSU), and are sometimes combined with a router through the use of modular insertion cards.

The CSU is used to provide the basic functions needed to transmit data across the line. Other basic functions provided by the CSU include the following:

• An electrical barrier The CSU protects the T1 (or other line) and the user equipment from damage that can be caused by unexpected electrical interference, such as a lightning strike.

• Keepalive signal The CSU transmits a signal on the line that is used to keep the connection up.

• Loopback capabilities You or the telephone company can perform diagnostics on the line using loopback facilities provided by the CSU.

• Statistical information Depending on the vendor and model, the CSU can provide statistical information useful to the network administrator. Some units have SNMP capabilities.

The DSU works with the CSU but also provides other functions. The DSU is responsible for translating between the data encoding used on the line, such as the time-division multiplexed (TDM) DSX frames that are used on a T1 line, and the serial data format used on the local network. A DSU usually has RS-232C or RS-449 connectors that can be used to connect to data terminal equipment (DTE), which then provides the actual physical connection to the LAN (see Figure 15.1). Each end of the line requires similar equipment.

Figure 15.1. The CSU/DSU provides the connection to a leased line or other high-speed service from the local provider.

Other important functions that the DSU can perform include these:

• Timing functions for user ports

• Error correction

• Handshaking across the line

Usually, the CSU and DSU are combined into one device. Typically, these functions also are incorporated directly into a router. If you use a router instead of a bridge (or another device), you can reduce the traffic that travels between the two connected LANs because only packets destined for the network on the other end are passed across the connection. In other words, you can save valuable bandwidth by routing only traffic destined for the far end of the network across the expensive leased line.

Analog leased lines are not as common as they once were. Most of the public switched telephone network (PSTN) between central offices has now been converted to digital lines because the service a digital line provides is much better than that of an analog line. Although an analog signal can be regenerated with amplifiers, noise on the line also is amplified, so the quality can deteriorate. Digital encoding with error correction techniques can deliver a signal over a long distance more accurately because the digital packets that are transferred can be corrected (in some cases where minor corruption has occurred) or retransmitted if the data cannot be recovered. Additionally, the conversion from digital to an analog signal and back again at the destination adds to the overhead of using an analog line. One of the disadvantages of a leased line is that it cannot be modified to give you a larger bandwidth. If you need additional capacity on the line, you must add another line or perhaps move up to another technology such as T1.

The T-Carrier System

In the past, leased lines usually would give you a bandwidth of up to 56Kbps. For larger bandwidth, you would need a larger data pipe, which is where the T-carrier or, if you are in Europe, the E-carrier system comes into play. Today, the term leased line can refer to various line speeds. The point is that the line is a dedicated link from one point to another, and you don't have to "dial" to set up the connection. The connection is dedicated and always on.

The T-carrier system was developed in the early 1960s by the Bell Telephone System in the United States and was used to digitally transmit voice communications. The first service offered was the T1, which can provide a transmission rate of up to 1.544Mbps. If you need more bandwidth than can be provided by a T1 line, you can contract for a higher level of service, such as a T3 line, which provides a 44.736Mbps connection. The range of transmission rates and number of channels for each kind of T-carrier service are listed in Table 15.1.

Table 15.1. T-Carrier Services in the North American Digital Hierarchy

Designation	Channels	Total Transmission Rate
FT-1/1	1	64Kbps
T1	24	1.544Mbps
T2	96	6.312Mbps
T3	672	44.736Mbps
T4	4,032	274.186Mbps

The T-carrier system is an all-digital transmission system. For voice systems that use a T1 line, the signal is sampled at a rate of 8,000 times per second, and the result is stored in 8 bits, or 1 byte. The T1 provides 24 separate channels that can be used to send voice or data from one place to another using two pairs of wires. Each of the 24 channels can transmit at a rate of 64Kbps.

Note

The European equivalent of the T1 line is called the E-1. Although the two use the same kind of technologies for transmission, the E-1 provides 30 channels and a total bandwidth of 2.045Mbps.

Time-division multiplexing, which allows each channel only a small amount of time to transmit (5.2 milliseconds), is used to combine all 24 channels into one signal. With each channel transmitting at 64Kbps, the total bandwidth on a T1 Channel is 64Kbpsx24, or 1.536Mbps. The difference between the 1.536Mbps and the full bandwidth of the T1 pipe (1.544Mbps) is due to the overhead used for managing connections (8Kbps).

Fractional T1

In many cases, the full bandwidth provided by a T1 line is more than the end user requires. Yet a slower 56Kbps leased line might not provide enough bandwidth. To handle this situation, the communications provider allows several users to use the full T1 bandwidth by allocating each user one or more of the 24 channels that T1 provides. This is called Fractional T1.

Diagnosing Problems with T-Carrier Services

When purchasing a T1 or T3 service, the local provider must check out the actual physical line and provide "conditioning" to be sure that it can transmit data at the expected rate with minimal errors. Conditioning means making the line stable enough to provide the service you contract for. Bridge taps and load coils that are normally found on voice-grade lines can't be used because they can cause the electrical pulses to be slightly out of shape, which makes them unrecognizable by devices on each end of the line. Inadequate grounding of the copper cables and physically defective cables are other sources of problems.

Bridge Taps and Load Coils

Bridge taps are places along the copper wire that have been tapped previously to provide service. Unfortunately, when services are disconnected, the telephone company doesn't always go out and remove the tap. After all, when someone moves out of a house in a residential neighborhood, the line is used again shortly, so after it's tapped, the line usually stays in place. For voice-grade service, bridge taps don't distort the signal enough for you to really notice anything. However, because no telephone is connected to the bridge taps, they are unterminated and can cause all sorts of problems for digital signals that travel over the line.

Load coils are another animal altogether. Because copper wires on the PSTN were originally used for just voice communications, it wasn't necessary to use much of the bandwidth that copper wires provide. In fact, voice service is usually provided by just the first 4MHz of the total frequency bandwidth that can be used on a copper wire. Higher frequencies on a wire tend to leak out and interfere with other wires. To solve this problem, load coils, which are basically low-pass filters (they allow the lower voice frequencies to pass through), are used to attenuate, or block, higher frequencies on the line. For ordinary telephone service, this is important; however, these higher frequencies are becoming very important today for providing sufficient bandwidth for digital services.

As you can imagine, bridge taps and load coils can also cause problems for xDSL technology, discussed in the next chapter.

When you request a T-class of service from a local carrier, usually you must wait a few weeks to a few months before the service is operational. If the telephone cables in your area were put in place many years ago, the carrier might have to condition the line by finding and removing bridge taps. New wiring also might have to be run in places where the original cables have degraded over time. All these functions are labor and time intensive.

The distance from the central office to your site also is important when trying to condition a line for any digital service. For example, twisted-pair wiring is normally used in the "last mile" from the central office to your business or home. The farther you are from the central office, the more the electrical signal attenuates. Because of this, the farther you are from the central office, the less bandwidth the wire can provide by the time it reaches you. To solve this problem, many telephone companies have been running fiber-optic cabling out into the field and installing a digital-services box closer to homes and businesses. Because fiber-optic cabling can carry a signal much farther with less attenuation of the signal than can copper wire, this effectively lets the telephone company put a mini central office out in the field. From this digital-services box, ordinary copper wiring can be used to connect to your location.

The loopback capabilities provided by the CSU/DSU unit are used by the provider to check the signal quality on the line. One of the simplest methods for checking the line is the use of a bit error rate tester (BERT). This provides a simple test to determine whether specific bit patterns transmitted by the test equipment can be received back with no distortions. BERT usually is the first test performed and is used to qualify the line as functional after the physical cables have been installed.

When a T1 line is installed, usually it is checked to ensure that all circuits are correctly terminated, which includes checking the user's equipment (such as the CSU/DSU). Signal loss can indicate that the connection is broken somewhere along the line, the signal being transmitted is too weak, or a connector is faulty.

For more information about BERT, attenuation, and other technical terms commonly used when discussing telephony, see Chapter 6, "Wiring the NetworkCables, Connectors, Concentrators, and Other Network Components," and Chapter 49, "Network Testing and Analysis Tools."

One problem that can occur is called timing jitter. As defined by the ITU-T (Telecommunication Standardization Sector of the International Telecommunications Union), timing jitter refers to "short-term variations of the significant instances of a digital signal from their ideal positions in time." The signal that is transmitted is a wave form; when viewed on an instrument such as an oscilloscope, you can see the rising and falling edges of the wave. If the wave form is slightly out of sync with the clocking mechanism, the signal might be interpreted by the receiving equipment incorrectly. All T1 circuits have a small degree of jitter, caused mainly by multiplexers or devices along the line that are used to regenerate the signal. Jitter also can be caused by electrical or atmospheric noise (as in the case of microwave transmissions).

Testing the line by using BERT or other instruments that depend on knowing the bit patterns that will be transmitted is called out-of-service testing. Obviously, this can be used only before the customer takes over the line for use. In-service testing, sometimes referred to as quality of service (QoS) testing, cannot make measurements based on expected bit patterns because the data transmitted by the customer can be roughly assumed to be random. Instead, tests performed when the line is already in service involve checking for things such as framing errors, parity errors, or checksum errors, depending on the kind of traffic carried by the line.