Around the world, digital subscriber line (DSL) services are in varying states of maturity from early trials to full commercial use. In theory, the customer gets broadband services, the operator receives additional revenues, and everyone is happy. In practice, it is not so simple. Alun Lewis explains.
For once, the different parts of the telcoms infrastructure and customer demand seem to be falling into step with one another. After an intense ten years of technological and regulatory change, there is increasing consensus about the eventual likely shape of the telecommunications network at least for the next few years. Much of the emphasis of the last decade's research and development has been on the backbone network firstly with the digitisation of switches, then the introduction of fibre optics and SDH into transmission and finally and more recently, moves towards a fully packetised network using ATM and IP. Complementing this have been parallel developments on the software side, including the maturing of Intelligent Network technologies and, in the last year or so, the emergence of philosophies based on truly open and distributed network concepts such as SoftSwitches and SIP and related initiatives such as JAIN and PARLAY.
On the access side of the network the last few kilometres to the customer's premises things have moved far more slowly. Until very recently the best the average domestic or small to medium sized business user could get was ISDN. This soon faced competition from top range modems. Sure, if a customer was lucky enough to live in the right area, they might have the option of a cable modem, but speeds here would soon be affected as other people on the same infrastructure came on-line.
Enter Digital Subscriber Line (DSL) technologies in all their different flavoursa
Now, around the world, DSL services are in varying states of maturity ranging from early trials to fully commercialised offerings. Whatever form DSL comes in and there are acronyms aplenty to choose from it basically works by using part of the available spectrum of an ordinary twisted-pair telephone cable to transmit data at speeds far above those possible with modem technologies. The historic strength of modems has been in their ability to transmit data at relatively high speeds across ordinary voice networks. However, because these networks have been optimised to carry voice, they are fitted with filters that limit voice-grade bandwidth to 3.3KHz and so limit the total available spectrum despite the inherent capacity of a copper twisted-pair to carry frequencies in the MHz ranges. DSL solutions, by contrast, typically use a frequency range between around 400KHz to 1.1MHz to send and receive data what could effectively be called awireless over copper' in the same way that coaxial cable is used to transmit video feeds in CATV networks.
The use of these frequencies means that signals lose strength (attenuate) the further they get from the source. The basic rule of thumb is that the higher the data speed, the shorter the distance over which services can be delivered. In practice, downstream data rates to the customer using DSL technology can range from between around 1.5Mbit/s to 26Mb/s and from a few hundred Kbit/s to a couple of Mbit/s in the opposite direction.
Typical configuration
In a typical domestic ADSL configuration, the service offered by an operator consists of one standard analogue voice line and one high speed Internet circuit, linked to a PC or LAN through a network adaptor usually using an Ethernet connector. At the local switch, the analogue signal is retrieved using a frequency splitter and sent through to the traditional circuit-switched PSTN, while the data stream is directed via a DSLAM card to the carrier's packet network and onto the internet.
The result, in theory at least, is that the customer gets broadband services, the operator gets additional revenues and everyone is happy.
In practice, as is ever the case with telecoms, things are a little more complicated at every level. Not only are there major engineering issues concerned with rolling out mass-market DSL services, but the implications for almost every other aspect of a service provider's operations and wider commercial strategy are only just starting to resound around the industry.
Taking the engineering issues first, there are a number of complex factors that interfere with the basic maths of bandwidth. Much of the copper loop presently in use, particularly in Europe and North America, has been there for a while, either hanging in the air or buried underground and signal degradation can easily result from age or poor splicing carried out over the years. Additionally, many local circuits have also had loading coils added to them to improve voice quality by cutting capacitance effects over long distances. These loading coils actually serve to block the DSL frequencies and so re-engineering is inevitably necessary.
Secondly, the radio aspect of DSL can also cause headaches for engineers. The use of the high frequencies required for DSL can lead to significant interference between different circuits travelling on adjacent wires and this may be exacerbated in a highly deregulated environment where different service providers using slightly different DSL technologies have to share infrastructure. Finally, filters put in to existing externally mounted lines to remove the possibility of radio interference from AM radio stations can also act to affect the DSL signals and so these have to be eliminated before installation can proceed.
The result of all this is that operators have to be extremely careful about making broad promises to customers about the ready availability of services. Bad experiences all too easily translate into lost confidence with both technologies and service providers. This can seriously affect business plans and shareholder confidence.
However, the cost of having to send an engineer to check every circuit before offering DSL service represents an enormous overhead for service providers. With a potential customer base that could easily rise to millions in a comparatively short space of time, a lot of attention has naturally been given to automating this process as much as possible. A number of solutions have been developed for this. They range from line-test software that can be sent to customers on a CD for them to initiate a line-quality check themselves, to more automated systems that can be run centrally by the service provider.
The next major area of concern centres around the impact of local-loop unbundling and the drive, in most countries, to encourage alternative service providers to compete with the established incumbents aon a level playing field'. In theory, a new operator should be able to install their own access equipment in the incumbent's exchange to pick signals up from their customers out in the field and then send them onto their own backbone networks so-called aco-location'. In practice, incumbents usually seem to find problems with this hardly surprising since they are effectively being asked to open their front doors to let the competition walk in and treat the place like home. Delaying tactics vary from situation to situation, but usually involve objections based on engineering practice, power or air conditioning availability and space limitations. Needless to say, lawyers are lining up on both sides to get involved in this area while customers are reduced to continuing to rely on modems or ISDN.
In this context, it is hardly surprising that some new entrants are either setting up independent joint co-location hosting sites in major centres or looking to the alternative broadband access technology of Fixed Wireless Access. The latter solution has been given a major push in the last year in Europe with the award of licences in most countries for 26GHz (28GHz in the UK) spectrum, capable of delivering point to multipoint digital microwave circuits to customers, usually in multiples of 2Mbit/s.
As a minor aside, there are two spin-off effects of DSL that are just starting to change other parts of the industry. Firstly, since DSL is an aalways on' service, in that you do not actually initiate a call by dialling up, the vulnerability of connected PCs to attack by hackers increases by an order of magnitude. The result is that domestic IT security suddenly becomes a very relevant issue and there is a corresponding increase in the marketing of firewalls and similar packages for the home market.
Secondly, the combination of bringing broadband into the home and the increase in homes with multiple PCs now justifies the case for the development of home broadband networks of some kind. Since few families want cables stretching around the house, there are a number of home networking industry initiatives under way to replace these with alternative methods. Systems currently available or under development include wireless networks using Bluetooth or Wireless LANs and those that employ the domestic power circuit as a LAN equivalent.
The final area of concern about introducing DSL and easily the broadest involves the whole aback office' environment of the service provider. Traditionally the least aglamorous' if such a word is appropriate part of telecommunications, the Operations Support Systems (OSS) department is currently undergoing what can only be termed a renaissance. With competition becoming an increasingly harsh reality for incumbent and insurgent new operator alike, attention is shifting from the ago faster' parts of network engineering towards those systems and processes that actually turn an infrastructure into a viable business.
When the customer is faced with a choice of service providers each offering broadly similar services at roughly similar prices they are going to make their decisions based on factors such as quality of service in both technical and human senses, time to provision service, time to repair and others such as flexibility in billing or self-provisioning. OSSs have been designed and built in the past to deal mainly with a far simpler world where basic dial tone was all that was on offer and, generally speaking, aone size fitted all'.
The challenge facing the service provider and supplier communities now is how best simultaneously to integrate and automate many previously labour-intensive operations to do with service creation, provisioning, assurance and management in what is a far more complex world. This need is further accentuated by the fact that, as each day passes, less and less telecoms traffic remains on just one operator's network and typically will pass across a number of different networks to its destination. Things are compounded further by the implications of number portability for the DSL world.
For the service provider looking to capture the lucrative corporate market through offering IP-based VPN services, issues soon arise of how they can provide consistent levels of service to a growing population of distributed users many of whom may be using DSL to link from home. Alternatively, the DSL circuits may be being used by an ISP to reach its own customers. Any provisioning or performance problems here will soon rebound onto the relationship between ISP and telecoms operator.
Looking to what is currently the most mature DSL market North America those operators who have been most successful in this space are those who have closely examined the wider implications of DSL right from the start. The complexity is mind-numbing even for the telecoms industry and involves examining right from the start how best to integrate the voice and data switching platforms, the DSL and other outside plant, all the necessary Quality of Service and bandwidth control functions, the billing and customer care systems, plus all the relationships with the other operators involved in delivering service.
While there is no doubt that DSL in all its variations is going to play an increasingly central role in the development of truly networked economies and societies around the world, it must also be regarded as a deeply subversive technology that will challenge familiar world views. In this context, every aspect of the continually evolving telecommunications value chain will be affected. The parts of the jigsaw are there it is up to us now to join them together.
