Getting Data over the Telephone Line: CSUs and DSUs

team lib

When you lease a T1 line for frame relay or Switched 56 service from either the telephone company or alternative service provider, you receive access to one of the company's connectors, to which your network attaches. (In telephone company jargon, everything beyond the connector is called customer premises equipment, or CPE.)

However, you can't attach your router to the telco's connector by simply plugging a cable with the right pins into each device. To ensure that a network interfaces properly with a telephone line, several functions must be performed. CSUs and DSUs were created to handle those functions.

The Role Of The CSU

A channel service unit, or CSU, is the first device the external telephone line encounters on the customer premises. As recently as the early 1980s, CSUs were always owned by the telephone company, which leased the devices to customers. But during the course of telecommunications deregulation , users were allowed to buy and install their own CSUs.

One of the principal functions of a CSU is to protect the carrier and its customers from any weird events your network might introduce onto the carrier's system.

A CSU provides proper electrical termination for the telephone line and performs line conditioning and equalization. It also supports "loopback tests" for the carrier, meaning the CSU can reflect a diagnostic signal to the telephone company without sending it through any CPE, so the carrier can determine if a problem is one it needs to correct itself. CSUs often have indicator lights or LEDs that identify lost local lines, lost telco connections, and loopback operation.

When CSUs were provided only by the telephone company, they were generally powered by the telephone line itself. Nowadays not all telephone lines supply power. So, a CSU may need to have its own power supply, and perhaps a backup power supply, at the user site.

The DSU's Duty

Data service unitssometimes referred to as digital service unitsor DSUs, sit between a CSU and customer equipment such as routers, multiplexers, and terminal servers. DSUs are commonly equipped with RS-232 or V.35 interfaces. Their main function is to adapt the digital data stream produced by the customer equipment to the signaling standards of the telephone carrier equipment, and vice versa.

To get a better grasp of a DSU's function, we need to delve a bit into the mysteries of the Physical layer. For those of us who aren't electrical engineers , it's easy to fail to give the appropriate respect to layer 1 and take the attitude "It's just electrons on the wire."

In fact, a large part of a telephone company's investment is devoted to delivering a Physical layer that's compatible with many years of legacy hardware, but also able to interoperate with state-of-the-art components such as fiber optic cable, Switched Multimegabit Data Service, and ATM.

The digital streams produced by many customer devices especially those with throughput less than 56Kbits/secare asynchronous, which means that each byte is distinguished by start and stop bits and that the time interval between bytes is arbitrary. However, the preponderance of customer devices in the public telecommunication infrastructure uses synchronous signaling, in which senders and receivers coordinate local clocks with each other in order to identify the boundaries between units of data.

In this case, the DSU may be called upon to parcel out incoming asynchronous data at the stable rate the carrier line expects, and to wrap start and stop bits around incoming synchronous data before passing it along to the user network.

The signaling techniques of the telephone network are quite different from those used by many customer premises devices. It's more or less natural to think of digital signals as positive voltage for a 1 and a zero voltage for a . This type of signaling is known as unipolar nonreturn to zero (see Figure 1A).

click to expand
Figure 1: With unipolar signaling, the mark signal, usually representing a 1, is indicated by a positive voltage or current, while zero voltage or current represents a space or 0. Polar signaling represents marks (or1s) with a positive voltage or current, while spaces (or 0s) are represented by a negative value. Return to zero signaling requires that each mark signal start and finish at zero within the alloted time for a bit.

There are several objections to a unipolar nonreturn to zero approach from the point of view of the telephone system, even though it works for RS-232 equipment. Unipolar nonreturn to zero signaling tends to build up a direct current (DC) signal component on a line over time. This signal component can be blocked by some types of electrical componentsfor example, transformers . Furthermore, a more or less random level of DC on the telephone wire interferes with the task of providing power to devices, such as repeaters and CSUs, that derive operating power from the line.

Using a signaling method where each 1 or returns to zero makes it easier to detect the correct digit, but does nothing to compensate for the DC (see Figure 1B). Using a polar signaling method where 1 s are positive and s are negative can diminish the DC buildup (see Figure 1C). But lengthy strings of 1 s or s will still have that result. Like unipolar signaling, returning each bit to zero doesn't solve the DC problem (see Figure 1D).

The solution is to indicate a with no voltage and indicate a 1 with alternating positive and negative voltages (see Figure 2A). No DC voltage can build because any residual charge from a positive-going 1 will be canceled by the succeeding negative-going 1 . Failure of a 1 pulse to have the opposite sign of the preceding 1 can be readily detected by simple circuitry and is known as a bipolar violation (BPV).

click to expand
Figure 2: Bipolar signaling indicates a mark or 1 with alternating positive and negative valves alternate mark inversion (AMI)and spaces or Os are indicated by zero values. Two successive marks with the same polarity constitute a bipolar violation, or BPV.

By making each bit return to zero, digits are easier to detect (see Figure 2B). In addition, this method allows wider separation of repeaters than other signaling methods . Bipolar signaling is also referred to as alternate mark inversion or AMI. Return to zero is often abbreviated RZ or RTZ, while nonreturn to zero is NRZ.

Keeping It Together

Another problem with the synchronous telephone network is the need to ensure that synchronization is maintained across any possible circuit. Telephone company devices, including the repeaters that rejuvenate the signal along the way, receive their clock, or synchronization, cues from the stream of bits.

One general rule for interconnecting older devices is that no more than eight bit times should pass without a signal. But it's not too rare for a stream of data to have eight or more successive s and thus provide no heartbeat for these devices. Without synchronization, transmission is impossible .

One solution has been to dedicate one bit out of every eight to a control role. If all the other bits in an eight-bit string are s, then the control bit will be set to 1 and synchronization will be maintained. The need for this control bit is the reason that dataphone digital service (DDS) lines and Switched 56 service, which after all run over 64Kbits/sec DS0 circuits, provide only 56Kbits/sec service to users. The eighth bit keeps the equipment running, so only seven out of eight bits are available to users.

An alternate solution to the synchronization problem, which has been widely adopted since the mid-1980s, involves forcing a deliberate bipolar violation to maintain the clock function while indicating that the BPV pulses are not to be interpreted as data. This technique, binary eight zero substitution , is called B8ZS. DS0 circuits that support B8ZS from end to end can carry 64Kbits/sec of user data.

Digital signals sent over telephone lines also need to be framed. Framing makes it possible to maintain the separation of multiple data streams that have been merged into a single circuit, with each stream getting its own time slot. Special overhead bits serve to demarcate the beginning of a frame. For example, on a DS1 (or T1) circuit, every 193rd bit is a framing bit, followed by 24 DS0 segments of eight bits each.

In light of these functionsAMI signaling, zero substitution, and framing requirementswe can return to the DSU. The role of the DSU is to convert the unipolar digital signal from the user device to one with the properties demanded by the telephone circuitry.

Mixing The Two

DSUs are often built into other devices, such as multiplexers or channel banks, and are often combined with a CSU to form a single unit referred to as a CSU/DSU or a DSU/CSU. CSU/DSUs may have built-in compression, and they may include analog or ISDN dial-up ports for backup.

Roughly speaking, a CSU has a similar role to that of an NT1 on ISDN lines, while a DSU is comparable to an ISDN terminal adapter. Sometimes CSU/DSUs (and ISDN terminal adapters) are described as "digital modems." This terminology is misleading and crude because CSU/DSUs don't modulate or demodulate anything.

Modulation is the process of using one signal (either analog or digital) to modify another analog signal, known as the carrier. The three fundamental ways a carrier can be modulated are known as amplitude modulation, frequency modulation, and phase modulation. For AM (amplitude modulation) broadcast radio, an analog audio signal modulates a carrier in the range between 540KHz and 1640KHz. For FM (frequency modulation) broadcast radio, an analog audio signal modulates a carrier in the range of 88MHz to 108MHz. Broadcast TV modulates a carrier using AM for the luminance or black and white video, using FM for the sound, and using a form of phase modulation for chrominance or color information.

Modems modulate an analog audio carrier that can pass through the POTS analog telephone line with digital signals. The modulation technique used by modern modems is a sophisticated combination of phase and amplitude modulation.

When CSU/DSUs are used on switched lines, they need to signal the destination of the call to the telephone company switch. In these cases, it sometimes makes sense for the DSU to use the AT command set originally employed by Hayes modems because software developers are familiar with the commands. Other than this minor overlap, and the fact that modems sit between a serial interface and the telephone network, the functions of modems and CSU/DSUs are totally different.

A modem is designed to transform digital signals into analog audio signals that fit into a circuit designed for voice transmission. CSU/DSUs adapt one kind of digital signaling to another type that is capable of fitting into the digital telephony system.

This tutorial, number 92, by Steve Steinke, was originally published in the April 1996 issue of LAN Magazine/Network Magazine.

 
team lib


Network Tutorial
Lan Tutorial With Glossary of Terms: A Complete Introduction to Local Area Networks (Lan Networking Library)
ISBN: 0879303794
EAN: 2147483647
Year: 2003
Pages: 193

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net