There has been considerable hype over the bandwidth growth of telecommunications networks driven by the phenomenal growth of the Internet. Overall the Internet has been growing at about 100% per annum for the last few years (except during 1995/6 when the World Wide Web appeared and much higher growths occurred) [1]. User volumes rather than user bandwidth demands drove the largest portion of Internet growth (most users were using dial modems with average session bandwidths in the 5–10 kb/s range).

Along with the euphoria that existed for everything Internet and communications generally was an assumption that if the bandwidth was delivered, services would materialise and generate new revenues. These additional revenues would pay for the provision of the bandwidth so producing a virtuous circle.

There is no doubt that the Internet, data and video services provide a huge potential demand for bandwidth. The problem is how to deliver that bandwidth in an affordable way. Bandwidth demand can easily outstrip the revenue growth that is needed to pay for the network investment and, not surprisingly, the access network provides the greatest challenges both in the provision of bandwidth and the necessity to reduce cost and expenditure.

Access is still the major bottleneck for broadband services. The vast majority of customer sites are connected via copper twisted pair, the majority of which have not yet been enhanced with DSL (digital subscriber loop) technology or fibre. A major decision for incumbent operators with large embedded assets in copper technology will be how far to upgrade using DSL technologies and when to replace it with fibre.