Due to its ubiquitous infrastructure, data communications has become strongly intertwined with the telephone system, despite its limitations as a high-bandwidth medium. For most of its life, the telephone network has served the single purpose of allowing people to converse remotely. Suddenly, in only a couple of decades, the uses being found for the phone system have exploded, and supplementary systems such as cellular phones and cable television are rapidly being installed. The rise of the Internet has fueled a burgeoning market for data communications over the phone network, an application for which it was not originally designed.
This week's Bookshelf installment, containing excerpts from "Introduction to Wireless Local Loop," by William Webb, published by Artech House Publishers, provides a detailed survey of the current scene, explaining the basics of the various connectivity options along with an analysis of their advantages and disadvantages. For the book, contact Artech House at www.artech-house.com.
Wireless Local Loop (WLL) is all about providing access from the home into the switched network. WLL is only one of a number of competing technologies that can be used to provide access. It is just one of many existing and proposed technologies that are, or might be, used to provide local-loop access.
The twisted pair can be used directly to provide voice communications. To provide data communications, it is necessary to make use of a modem. The telephone channel has a bandwidth of about 3 kHz. It also has a relatively good signal-to-noise ratio (SNR) of some 30 to 40 decibels. That means that although only some 3,000 symbols per second can be transmitted, each symbol can contain a relatively large amount of information. Instead of representing just two different levels, as is normal in digital modulation, it could represent, say, 16 or 32 different levels. The modulation used to achieve this is termed quadrature amplitude modulation (QAM).
The key advantages of voiceband modems are the following:
- The economies of scale achieved have resulted in a cost per modem of around $200 each.
- They can be connected directly to a telephone line with no need for the phone company to modify the line in any manner.
The key disadvantages are the following:
- They need a dedicated line for the time they are in use; hence, voice calls cannot be made or received on the telephone line.
- The maximum capacity is around 56 kbits/second, which is relatively slow for computer data transfer.
Integrated Service Digital Network (ISDN) basically is a framing format that allows data to be carried at a range of data rates across a bearer. ISDN makes use of the fact that twisted-pair cables can carry more information if the problems of crosstalk can be overcome. To provide ISDN access, the phone company first must remove filters on the line that prevent signals of bandwidth greater than 3 kHz being transmitted. There is an installation cost involved, which the user must pay. An ISDN modem is then installed at both ends of the line.
Not all lines are suitable for ISDN. Older lines, or lines over 3 km, typically cannot carry ISDN because the crosstalk with other lines is too severe or the signal attenuation too great. A test on the line is required before ISDN service can be provisioned.
ISDN is an international standard that provides a range of data rates. The lowest rate ISDN channel is 64 kbits/s, with a typical ISDN deployment providing a basic-rate ISDN access, or BRA. There are two basic (B) 64-kbit/s channels and one data (D) channel of 16 kbits/s. The data channel can be used to provide signaling information, while both basic channels are in use. Hence, a 2B + D channel provides 144 kbits/s. The advantages of ISDN include the following:
The area of digital subscriber line technologies is a relatively new one (the abbreviation xDSL refers to all the approaches to digital subscriber lines). The concept, like ISDN, is to use existing twisted pair, less any filters that may be in place, to provide significantly greater data rates through complex intelligent modems capable of adapting to the channel and removing any crosstalk that might be experienced. The term xDSL has come about to encompass a host of proposed different approaches, such as asymmetric digital subscriber line (ADSL), high-rate digital subscriber line (HDSL), very high rate digital subscriber line (VDSL), and doubtless more to come.
Research has shown that these technologies can offer up to 8 Mbits/s, perhaps more, depending on the quality of the existing twisted pair. Readers at this point may be asking themselves why on the one hand the twisted pair can provide only 56 kbits/s and on the other hand the same twisted pair can achieve 8 Mbits/s. The reason has to do with the manner in which crosstalk is treated. Voiceband modems overcome the problem of crosstalk by ensuring that none is generated. The xDSL technologies generate significant crosstalk but employ advanced technology to cancel its effects. It is that difference in approach, enabled by advances in digital signal processing, that has allowed xDSL to make such dramatic improvements in the data rates that can be achieved. HDSL is intended for business applications. HDSL signals can propagate only a few kilometers along twisted pairs. Most businesses, however, are relatively close to their nearest exchange, so that is not a significant limitation. HDSL typically is less suitable for residential applications, because homes may be at much greater distances from the local exchange.
After HDSL came ADSL, which provides more data in the downstream direction than in the return path. This asymmetry meets the requirements of Internet access well, where more information is passed to the home than is sent into the network from the home. By restricting the return path to lower rates, less near-end crosstalk (NEXT) is generated. NEXT is interference from the return signal that contaminates the received signal. Because the return signal is at a lower rate, the effect of NEXT is reduced and higher downstream rates achieved. ADSL promises to provide up to 8 Mbits/s downstream but only tens of kbits/s upstream. Current trials are achieving around 1.5 Mbits/s downstream and 9.6 kbits/s on the return path. ADSL works by dividing the transmitted data into a number of streams and transmitting the streams separately at different frequencies. This approach is known as discrete multitone (DMT) in the fixed-line community; however, the technique has been used for many years in mobile radio normally known as frequency division multiplexing (FDM) or orthogonal frequency division multiplexing (OFDM). Indeed, this is the technique proposed for digital audio broadcasting and digital terrestrial TV broadcasting. This approach has the advantage that each transmitted data stream is narrow-band and does not require equalization.
ADSL is more appropriate for residential applications. By reducing NEXT, the range achieved is greater than that for HDSL, allowing long residential lines to carry ADSL successfully. Also, the asymmetrical signal typically is suitable for residential applications such as video on demand (VOD), in which more signal is sent to the home than received from it. It is estimated that up to 70 percent of all residential lines in the United States could be suitable for ADSL operation.
Finally, VDSL has been proposed where fiber to the curb (FTTC) has been deployed. In that case, the copper run to the subscriber's premises is very short, typically less than 500 m; hence, higher data rates can be supported. Using the most advanced technology proposed yet, it is suggested that VDSL could achieve data rates of up to around 50 Mbits/s, although that is still far from being proved. Current plans suggest 10 Mbits/s downstream and 64 kbits/s on the return path. VDSL cannot be used in networks in which FTTC has not been implemented. xDSL will be expensive to implement, even though the local loop will stay relatively unchanged. The phone company will need to install n\ew optical cable from the switch to a new cabinet in the street. Modems for xDSL are predicted to cost around $500 each, although the price in the coming years will depend heavily on the success of the technology and the economies of scale achieved.
A problem with all the xDSL technologies is that the data rate that can be achieved depends on the length and the age of the twisted pair. As the length gets longer, the data rate falls. As yet, it is not clear what percentage of lines will be of sufficient quality to accept xDSL signals. Figures quoted in the industry vary from around 60 percent to 90 percent. Due to the technology's relative newness, texts on xDSL are hard to find and tend to be limited to chapters in books. The key advantage of xDSL is the potential extremely high data rate on existing ubiquitous lines.
The key disadvantages of xDSL are as follows:
- The modems are relatively expensive.
- The technology is unproven.
- It is unlikely to work for all homes.
Cable operators have implemented what often is known as a tree-and-branch architecture. Cable networks vary in their composition. Some networks are entirely coax, others use fiber optic in the backbone (the trunk of the tree) but coax in the branches. The latter networks are FTTC or hybrid fiber coax. Some postulated networks are composed totally of fiber, termed fiber to the home (FTTH). At present, the economics of FTTH are not favorable.
FTTH has been something of a Holy Grail for phone companies and the cable industry because of the assumption that it represents the ultimate possible delivery mechanism, capable of delivering gigabits to the home. However, it has been pointed out that FTTH is akin to each branch connecting around 200 homes.
While fiber has a virtually unlimited bandwidth (on the order of gigabits per second), coax has a bandwidth of up to around 750 Mbits/s in the existing installations. With an analog TV picture requiring some 8 Mbits/s of bandwidth, that still allows numerous TV channels. With just one analog channel, some 50 Mbits/s of data can be transmitted using similar QAM techniques to voiceband modems. Cable, then, offers much higher capacity than even the xDSL techniques over twisted pairs.
This is not quite the whole story. For each home, there is one (or two) twisted copper pairs running from the switch right to the home. In a cable network, all homes share the backbone resource and the resource of the branch of the tree to which they are connected.
FTTH is both expensive and problematic in that, unlike current telephony, power cannot be supplied along with the signal. Now that the phone companies are facing competitive situations and investments increasingly have to be justified, it seems unlikely that FTTH will be implemented in the next 20 years or more.
With cable all homes on one branch are connected to the same cable, whereas they all are connected to their own individual twisted pair. That is fine while cable is delivering broadcast services, to be watched by many viewers simultaneously, allowing 50 or more TV channels. However, if each user on a branch wants a VOD service, then only 50 users could be accommodated on one branch, unlike twisted pair using xDSL, where as many users as needed could be accommodated.
Such a sharing of resources causes even more problems in the return direction. Not only is the return path shared among all the users who require it, significantly reducing the capacity, but further, each user introduces noise onto the return path. The switch sees noise from across the entire network, significantly reducing the SNR and hence th content that can be received.
Despite all those problems, cable modems are being put on trial with a downstream capability of 30 Mbps and an upstream capability of 10 Mbps for an expected price of around $500.
TV signals currently are broadcast via terrestrial transmission, satellite, and cable. Terrestrial broadcasting uses about 400 MHz of radio spectrum in the UHF frequency band (typically 400 to 800 MHz in most countries). However, the same spectrum cannot be used in adjacent cells, because interference would occur. Satellite is in a similar situation. It has more spectrum, around 1 GHz in total, and does not need to divide that spectrum among different cells in the same manner as terrestrial.
Moreover, neither TV or satellite delivery mechanisms can offer a return path. Service providers for both mechanisms postulate that they might use the twisted copper of the phone companys for the return paths, but they would get limited revenue and would have difficulty truly integrating their services. So, as an access method, both leave a lot to be desired. Their main role, with the advent of digital TV, will be the delivery of broadcast data in relatively large volumes.
Cellular systems are available in most countries in the world. Compared to the other delivery mechanisms discussed so far, the key difference for cellular is that it provides access to a mobile terminal as opposed to a fixed terminal. But all cellular systems suffer from capacity problems, even when they are providing voice-only traffic. Cellular systems also suffer with voice quality. Even digital cellular systems provide a voice quality inferior to existing fixed-line quality. The dropoff is tolerated only because of the advantage of mobility.
Some cellular operators have attempted to attack the fixed-access market using initiatives such as cut-price evening calls and high levels of indoor coverage. Invariably they have discovered that they cannot match the phone company on price and that the cost of providing excellent indoor coverage is extremely high. When recently interviewed, all the U.K. cellular operators stated that the percentage of traffic they were taking from the fixed operators was minimal, certainly accounting for less than 5 percent of their overall call base. Although that is likely to rise in the future, it now appears clear that cellular will not be a substitute for the fixed network but a complement when mobile.
The most successful cellular system, the global system for mobile communications (GSM), offers voice or data, with maximum data rates of up to 9.6 kbits/s. Future enhancements to GSM might raise the maximum data rates as high as 64 kbps and improve the voice quality, but that rate is slow compared to the other access technologies and is expensive in call charges.
Cellular systems have many disadvantages as an access technology, particularly low capacity and high cost. In developed countries, it can be expected that for some time they will continue to fulfill the role of providing mobility but will not present significant competition to the other access methods. That is not as true in developing countries, where cellular operators can use the same system to provide mobile systems and WLL systems and thus realize economies in equipment supply, operation, and maintenance.
Cellular-based systems have a role to play in WLL networks but fixed mobile integration also may affect the selection of an access method. Cordless systems are similar to cellular but typically are designed for office and local area use. The key difference between cellular and cordless technologies in their application as an access technology is that cordless technologies such as the digital European cordless telephone (DECT) offer a higher data rate than cellular, up to some 500 kbps, with lesser limitations on spectrum and hence expense associated with the call. But, because cordless technologies typically provide coverage only in buildings and high-density areas, they will be unlikely to have coverage for most access requirements. It is that lack of coverage that makes them inappropriate as an access technology. However, when deployed as a WLL technology, cordless becomes viable as an access technology.
INTRODUCTION TO WIRELESS LOCAL LOOP (0-89006-702-3) BY WILLIAM WEBB IS PUBLISHED BY ARTECH HOUSE PUBLISHERS (WWW.ARTECH-HOUSE.COM).