Study Shows Inequalities in "White-Spaces" Wireless

A person living in upstate Michigan may gain significantly more from the death of analog television than someone living in New York City–at least, as far as long-range wireless Internet is concerned, a study suggests. On November 4, 2008, the Federal Communications Commission voted to allow the “white spaces” in the radio spectrum that were freed up by the analog television switch-off to be used for long-distance wireless Internet connectivity. This spectrum will be unlicensed, meaning any standards-compliant device can use it.

The most detailed analysis yet of the potential of these white spaces for long-distance wireless Internet has now been published by researchers at the University of California, Berkeley. Their models illustrate how the interaction of population density, television stations, and economics will determine what consumers ultimately get.

The decision to free up white spaces was warmly welcomed by companies including Google, Microsoft, and Intel because the lower-frequency signals in the newly released spectrum travel further and penetrate buildings more effectively than existing wireless data connections like Wi-Fi, Bluetooth, or cell-phone links. (Google cofounder Larry Page has dubbed this kind of wireless connectivity “Wi-Fi on steroids.”) But unlike other commercial spectrum, the frequencies available to devices will not be the same in every location. Devices will instead have to use whatever spectrum is unclaimed by the digital television stations in their area. Although this is intended to make more efficient use of available spectrum, it makes the actual deployment of these white-spaces devices more complex, because they will need to avoid treading on the radio-frequency toes of local television broadcasts.

The FCC has drafted rules on the requirements for how devices might do that, with the final version expected late this year, and various electronics companies are already working on white-spaces devices. But the exact scale of the opportunity is still unclear, says Mubaraq Mishra, who conducted the new study with Berkeley colleagues Anant Sahai and Kate Harrison.

To get at what may be on offer to people who want to use white-spaces Internet, their model uses the FCC’s database of active television stations to calculate the usable white space available in each area. It also draws on census figures for the number of people living in each U.S. zip code. Including population data at the zip-code level makes their assessments unique, says Mishra, and more powerful. “White spaces are ultimately for people, not random locations, so we think it’s important to also incorporate that,” he says. “We wanted to know how many channels or how much bandwidth does a random person in the U.S. get.”

The answer is that it depends where that person lives. The number of television stations operating in an area constrains the amount of white space available, while the number of people living there determines how much white space can be dedicated to each individual.

The image shown top left, for example, shows the variation in personal bandwidth that might be available across the country in a scenario in which every 2,000 people use one white-spaces transmission tower in their area. The New York City area is noticeably darker than upstate Maine, for example, meaning that users there would experience slower speeds. “We see things like that, because where the people are, the TV broadcasters are too,” says Mishra. Some places have a good balance of population and television-channel density that allows for greater per-capita bandwidth.

“These analyses could be used for a range of business or technical decisions, for example to help companies figure out what kind of services or broadband offerings are possible and start answering questions about the size of the cells needed to do that,” says Mishra.

“This gives very interesting insights,” says Kiran Challapali, a researcher at Philips in New York and chair of the Cognitive Networking Alliance (CogNeA), a trade group working on standards for devices to use short-range white-spaces links inside the home. “It will be useful for those thinking about getting into white-spaces broadband.”

Kyutae Lim, associate director of technology at the Georgia Electronic Design Center and chairman of the only international standard yet for white-space devices, published by international standards body ECMA International in December 2009, agrees. “The commercialization aspect for long distance is less attractive than low power because you have to build towers, and spectrum availability is unclear,” he explains. “It’s risky to spend money to build a system, so analysis like this is needed.”