What is ADSL? how does it work? how have BT made a version that is more expensive and goes slower than all original expectations lead me to believe it would be?

ADSL is a digital transmission technology that employs frequency division multiplexing to provide two separate simplex channels for transmission and reception of data. The technology is asymmetric in that the bandwidth allocated to the downstream link is higher than the bandwidth allocated to the upstream link. This is the case because generally, users will receive more data than they send. This fact makes ADSL less suitable for running a web site that most people would like. The standard voice telephone carrier is also multiplexed on the same line, usually at a lower frequency than the upstream and downstream links.

ADSL access rates start out at about 2Mbit/sec which is very similar to the data rate of 2.048Mbit/sec achieved by ISDN Basic Rate Access over a leased line. Cable modems by comparison should be capable of about 30-100Mbit/sec when initial problems with contention ratios have been resolved and capacities for each spur have been increased. Initially access rates will be around 2Mbit/sec on properly designed systems. It is important to remember that some companies rather short-sightedly did not anticipate this need for bandwidth when they installed their systems and consequently, are going to have to pay significant costs to upgrade all their systems.

ADSL data rate is dependant on the attenuation characteristics of the copper cabling used. Obviously, the further the client from the exchange, the greater the attenuation will be due to resistance. At a distance of about 2.5km from the exchange, data rates of 6.5Mbit/sec will be possible, falling to as low as 2.5Mbit/sec at around 5km from the exchange. This is the obvious limitation of the maximum data rate that this cabling will ever be able to provide and it calls in to question whether ADSL services are worth the investment for a technology that at most, has a lifespan of a few years. Really what is needed is a replacement of this obsolete copper wiring with fibre.

Cheaper versions of DSL use less expensive filters (only a few pounds less expensive) to separate out higher frequency ADSL signals from voice. This, given BT’s track record, it was always likely that this would be the version that BT would sell the British public. This means that BT can sell slightly faster versions for extra high premiums, it also means that ADSL will not have such an impact on BT’s high margin leased line market and in the future slightly faster versions can be sold to existing users without significantly changing the technology, making BT even more money. This cheaper version also saves BT backplane capacity on their exchanges, since presumably the exchange is responsible for packet switching the data on to the telecom companies internet backbone. Exchanges were never designed with this purpose in mind and it would be interesting to know what the total backplane capacity of an exchange actually is and how many ADSL users an exchange could support. Presumably the contention ratio is imposed on this basis and this is likely to be done quite conservatively to allow for greater than expected uptake.

ADSL works with a technique called Discrete Multi-tone Modulation (DMT). This is used to provide a number of channels between about 20kHz and 1.125MHz. To reduce interference, not all these channels are used. The spectrum between these frequencies is divided up in to 256 carriers spaced at 4.3kHz. To maximise channel capacity, binary data is converted in to quaternary data in pairs of 2 bits to improve data rates. Forward error correction is employed in ADSL to eliminate any errors due to noise.

Local loop unbundling is cited as a big issue. Really, it is probably not until BT replace the copper cabling with fibre. Telecom providers can only provide a limited amount of bandwidth over copper cable and BT exchange equipment probably can handle particularly high data rates anyway. There also has to be a considerable degree of compatibility in non BT equipment used and frequency spectra usage will have to be controlled carefully to prevent interference between devices.

Detailed technical documents by BT explain some of the issues behind UK ADSL implementation. BT's approach in these documents appears to be quite cautious and conservative. Perhaps this will also be one of the factors that limits the speed of BT DSL services. A BT paper discusses issues such as RF, crosstalk and various other forms of electrical interference. My overall view of this paper is that BT are taking a very conservative view of implementation of ADSL which may lead to slightly better coverage, but this will be at expense of service performance and competition in the sector. If local loop unbundling is done the BT way, there will be little scope for variation in the type of equipment which will be allowed to be used over existing cabling, and consequently little real competition or innovation in the sector. BT also seem to be looking at ADSL services with a long term view of potential problems which are more likely to occur if ADSL services were to become very widespread. I would seriously doubt that ADSL has the lifespan for this type of approach to be necessary, desirable, or cost effective. It is however reassuring that BT have looked at the issues in some depth and appear to be going about implementation of ADSL in a sound, if somewhat conservative (but unfortunately very anticompetitive) way. It is really Oftel's responsibility to make sure BT do not get any say in the technicalities of local loop unbundling. All technical decisions MUST, for the sake of competition in the UK telephone market, be made by BT's competitors and independent commitees.