Speed is needed, not intelligence

By NEIL JAMES*

Back in 1997 David Isenberg, an engineer then working at AT&T Bell Labs, wrote an essay titled The Rise of the Stupid Network.

Isenberg challenged the direction telecommunications companies were taking in developing their networks and services. He argued that while an intelligent network, which had a knowledge of the information it was carrying, might have been a good choice in the past (because of expensive and scarce infrastructure), this was not now the case.

Designing a network that is intelligently tuned for particular types of service, such as television delivery or a voice service or the transport of "bursty" data, inevitably limits the applications it can support.

If, on the other hand, the network is designed simply to move bits from one address to another as fast as possible, it can be used for the widest variety of communications activity. Such an open network encourages innovation because the way it is used is left in the hands of the users.

The telephone voice network is an example of a network designed for a specific purpose. In recent years this network has also been required to do double duty and act as transport for internet access for large numbers of people. This has caused distortions in use that have endangered the voice network.

Some telephone companies have been responding by retrofitting their networks with technologies such as ADSL. This technology is limited both in terms of speed and reach, and is expensive. It requires specialised equipment to be embedded in telephone exchanges.

In Isenberg's view such approaches are "crippled compromises at best" and are being used to extend the life of the established telecommunications business model.

To move beyond the old model for communications we need to embrace a new vision. We should develop reliable, simplified, and affordable networks that run at speeds much faster than current telephone technologies. The technologies required are becoming cheaper, making the dream of virtually unlimited bandwidth achievable.

Several initiatives around the world, including the Fibre to the Home (FTTH) Council (www.ftthcouncil.org/index.shtml) in the United States, and Canarie (www.connect.gc.ca/en/290-e.htm) in Canada, have just this aim.

"Stupid" fast networks can be used for a wide variety of communications purposes, including voice calls, video connections and data transport. The intelligence needed to handle and encode the various forms of data is added at the edge of the network, and may well be implemented by devices owned by the customer. For example, two parties wishing to video-conference need only a network connection, one of the many available video conference units, and the relevant IP addresses.

There is no requirement for intermediation of a telecommunications supplier.

How can New Zealand future-proof its communications? The most important single step is to acknowledge that our nation's vision for communications must be set very much higher than, for example, that embodied in the Probe Project. We need to understand that retrofitting an old telephone network will not take us into the future.

In 1999, the Swedish ICT Commission presented a vision of a future-proof IT infrastructure for Sweden with a fibre-optic network reaching everyone by 2005 and providing permanent connection to the internet for no more than the price of a bus pass. Further, it envisioned the speed of the connections doubling each year.

This would be a good starting point for New Zealand's vision for the future of communications in a knowledge society.

* Neil James is assistant director of the information services division of the University of Otago.

Email: neil.james@stonebow.otago.ac.nz