Interest in dynamic spectrum sharing seems to be increasing
Jim Carlson, president and CEO of Carlson Wireless and Preston Marshall, who currently is working for Google in the area of wireless network development, are kindred spirits these days. Carlson has been working for several years on solutions that are capable of taking advantage of TV white spaces spectrum. Meanwhile, Marshall served for seven years as a program manager for the Defense Advanced Research Projects Agency (DARPA)—a unit of the U.S. Department of Defense—leading a project that worked to develop a cognitive radio that costs less than $500.
Carlson and Marshall now are focused intently on dynamic spectrum sharing as a means of improving spectrum utilization—a vitally important task, given how demand is exploding just about everywhere one looks, driven by the rapid expansion of smartphones, tablets, machine-to-machine devices and the smart grid.
Indeed, the report issued last July by the President’s Council of Advisors on Science and Technology (PCAST) cited one statistic that predicts the number of devices connected to mobile networks worldwide will grow tenfold by 2020, from 5 billion to 50 billion. To meet this demand, PCAST recommended that President Obama direct the Commerce Department to identify 1000 MHz of federal spectrum that could be leveraged to create “shared-use spectrum superhighways.”
However, relocating existing licensees to free this much spectrum is wholly infeasible, Marshall said.
“Relocating people is increasingly difficult—it is comparative to ice building up in front of a plow as you move forward,” he said. “And we’ve built up a lot of ice—some of these people have been moved once or twice before—and it’s only going to get more difficult.”
Dynamic spectrum sharing technology offers a potential solution to this problem, according to Carlson and Marshall. It works something like this: a company, such as Spectrum Bridge or Telcordia, would maintain a database—updated daily by the FCC—that indicates which frequencies are available for use in a given area at a specific time of day and what power can be used at a particular time. An enabled device then would query the database to discover what it is allowed to utilize at that particular moment, given its GPS coordinates; once it has this information, the device automatically selects an available frequency and adjust its power accordingly, so it doesn’t interfere with incumbent users in the band.
“It all happens in less than a millisecond,” Carlson said.
The flip side is that such devices “are far from ready to go,” said Carlson, but he added that prototypes are being tested in the UHF band using experimental licenses. While Carlson envisions that dynamic-spectrum-sharing devices ultimately will operate in 3.5 GHz and/or 4.9 GHz spectrum in the next few years, the UHF band was selected for the test for a couple of simple reasons, according to Carlson.
“There’s a contiguous 240 MHz there, and the VHF band is split into two pieces, which is much more problematic—there’s impulse noise and skip issues with VHF that you don’t have with UHF,” he said. “Also, internationally, VHF is not universally used in a lot of countries—it’s all UHF.”
According to Carlson, a device capable of dynamic spectrum sharing is easier to develop, from an engineering perspective, than software-defined radios that also look for slivers of VHF and UHF spectrum between TV broadcast channels known as “white spaces.”
“A white space device is going to be much more complicated if it’s going to have significant RF power behind it,” he said. “That’s because, to go from 470 to 800 [MHz], it’s close to an octave of bandwidth—and try to do that with one watt, and do that with your side lobe emissions being 56 dB below the carrier. That makes it an extremely challenging radio. Whereas if you’re doing that in narrower bands up higher, that are only 3% or 5% of an octave, then it becomes easy and simple.”
Another advantage to dynamic-spectrum-sharing technology is that it can utilize shorter hops than white-spaces devices, which makes it possible to coordinate small femtocells that offer cost and logistical advantages. While many more are needed to produce the same coverage provided by a white-spaces device, they are less expensive to produce and don’t require expensive infrastructure to deploy them—they can be placed on buildings, utility poles, water towers and the like. For areas with user densities above 50 per square mile this allows a workable solution.
However this all plays out, Marshall stressed that the traditional thinking regarding spectrum utilization must change, citing the 3.5 GHz band—used primarily by the military—as a good example.
“Its use is dominated by a single high-power system, and NTIA originally said that it’s not very useful, because if you put high-power LTE in there, it really can’t operate,” said Marshall, who believes that the sweet spot for dynamic spectrum sharing is between 2.7 GHz and 3.7 GHz.
“But, if you take a look at alternate architectures, like small femtocells, then all of a sudden this band looks really attractive. If you change your vision of what might be deployed there and think about a different mix—a much more heterogeneous mix of systems—then suddenly spectrum sharing makes a lot of sense.”
It also makes dollars and cents, according to Marshall.
“PCAST believes that there’s about a gigahertz of spectrum that could be shared under one arrangement or another,” Marshall said. “But not just to create unlicensed bands. You also could create a type of market where you could sell the right to share spectrum, and you could make that as much a right as an auction, hopefully with more liquidity.
“I might buy the spectrum around a hospital, because I want to have high QoS for my medical instrumentation. Or someone might want to do carrier offload and buy the spectrum in their apartment. You can imagine lots of different models for spectrum access and innovation that don’t fit the current licensing model.”