Broadband without broadband

Billions of dollars will be invested this year developing broadband technology to deliver video, voice, and the Internet to homes and businesses. Hundreds of companies are working in areas such as DSL, cable modems, wireless and optical communications. All this investment is needed because broadband aims to give users direct access to tens of gigabits of content. Today's DirecTV satellite broadcasting service, for example, carries a total of more than 3 gigabits of digital video per second.

Broadband deployment already is in sight. Within five to 10 years, almost all of us will have access to fast Internet connections and digital television.

Ethernet is the simplest anwer to the "last mile" problem

Or we could start giving customers these benefits next year with no further R&D. All the necessary technology exists to deliver video, voice and data economically to nearly everyone on Earth. You've probably got most of the technology in your own office. Ethernet is the simplest answer to the "last mile" problem. Plain old baseband 100Base-T Ethernet, based on copper cable, can connect most apartments and offices to a wiring closet in the same building. Robust, inexpensive copper wiring also will suffice for most houses; several homes can share a cheap Ethernet switch with a gigabit optical uplink to the nearest neighborhood "pedestal" (wiring box).

Some broadband technology—in the form of wireless radio-frequency links—is needed for the few customers outside the reach of wired networks. Current services such as Sprint Broadband, or new systems based on the terrestrial use of the same technology used for DirecTV, could meet this need.

Commodity Ethernet links are symmetrical—the uplink and downlink speeds are equal. This symmetry would allow the Internet to fulfill its true potential as a peer-to-peer platform, not just a way to deliver limited content choices from high-cost, high-bandwidth commercial Web sites. Indeed, the economics of a baseband system suggest that only uplink bandwidth needs to be metered; downlink service should be provided on a flat-rate basis.

The chips to build Ethernet-connected set-top boxes and telephones have been available for years. Such set-top boxes would be just like cable or satellite set-tops, but with the expensive radio-frequency tuners replaced by cheaper Ethernet interfaces. Ethernet telephones are more expensive than our familiar analog telephones, but only one is required to interface a home's existing phone wiring to the new system.

I'm not saying this baseband scenario represents how the world will, or even should be connected, though it does offer certain compelling cost and capability advantages over the broadband approach. I'm just saying that this scenario is possible. Given the way today's major players are investing their R&D funds in broadband solutions, baseband would have to come from a smaller, more agile company. But remember, there's nothing to invent here. Someone could begin today and deliver their first products next year. Maybe that someone could be you.

Peter N. Glaskowsky is a senior editor of the Microprocessor Report .  For more information on topics covered in this column, visit http://www.chipadvisor.com. Send Peter e-mail at chipadvisor@ideaphile.com.