I see several drivers for increased Internet traffic, but the biggest (and longest-term) is Internet television.

Consider that today you can listen to thousands of radio stations around the world using the Internet. "Near CD-quality" audio is now streamed at 64 kbits/second using the latest codecs.

The same model could be applied to video, but it requires much more bandwidth. While there are many video sources on the Internet today, most are broadcast in low- to medium-quality streams at up to 500 kbits/s. Digital cable and satellite broadcasts are typically encoded at about 1.5 Mbits/s.

For the consumer, watching television on a computer screen is not very compelling. But it would be easy to add Internet capabilities to a TiVo-like device that could then display downloaded or real-time programming on a television set.

That arrangement could solve the video-on-demand problem by providing access to movies and other archived programming. Ultimately, it would allow anyone with an Internet server to become a TV station.

For service providers, there is already a working business model for providing television programming to the home: Many consumers pay $40 per month or more for that service and would probably pay even more for video-on-demand.

An Internet service provider could compete against cable and satellite TV by negotiating to distribute a comprehensive package of programming. Television networks and other content owners could use the technology as an additional distribution channel.

This new paradigm will require the ability to receive multiple broadcast-quality video streams reliably and without significant delays. Emerging technologies like fiber to the curb will be needed.

That additional bandwidth capacity will ripple through the rest of the Internet. In the United States alone, if 100 million people were watching a 1-Mbit/s Internet stream, it would generate 100 terabits per second of traffic.

That number seems huge, but if Internet traffic continues to double every year, it will reach that level in less than 10 years. If television moves onto the Internet, that growth rate is assured, driving new chip and equipment sales.

Linley Gwennap is Founder and Principal Analyst of the Linley Group (www.linleygroup.com).