Nate Anderson over at Ars Technica wrote an in-depth analysis of the current white space debates. It's well worth a read.

White spaces, angry faces: Inside the battle over 'interference'

By Nate Anderson |
Published: September 23, 2008 - 11:30PM CT

High stakes

Every city in the country—even New York City—has a host of unused TV channels. Opening up that fertile field of spectrum to the seeds of innovation is a worthy-sounding goal that everyone can agree to in principle, but when it comes down to making the rules that govern access, and to certifying the devices that can operate, the debate hops on the express train to Nastytown.

Read more...

So say the broadcasters

How did a campaign to allow unlicensed access to TV "white spaces" turn into a "campaign of fear," a "political proceeding," and a series of "ridiculous assertions"? Because of what's at stake.

I sat down recently with several of the leading voices in the white spaces debate. The goal: to get beyond the political posturing in order to present the technical issues being argued about in Washington. What quickly became apparent is that the two aren't so simple to separate.

Black, white, or gray?

Despite the ferocity of the debate over mobile white space devices, one thing that becomes clear in talking to the participants in the conflict is that the issue isn't just a black-and-white engineering challenge. It's much more a calculation of risk vs. reward, except that in this game, the incumbents stand only to lose (from interference), while the new entrants stand only to win (by selling devices and creating more demand for their online services).

No wonder the broadcasters have dug in deeper than World War I trench fighters; in their view, the stability of broadcast TV is at stake, with little upside for them should the experiment go bad.

Take the question of spectral availability. While this might seem one of the easiest to settle—is more than half of US TV spectrum lying fallow or is it not?—the answer isn't obvious. The New America Foundation, a think tank that supports white space devices, claims that 25 to 80 percent of TV bandwidth is unused, depending on where you look. When the group hosted Google co-founder Larry Page at a DC event a few months back, Page also spoke of all that prime spectral real estate that was being "wasted."

But hold on a minute, says that National Association of Broadcasters. We spoke with the group's Senior VP of Science and Technology, Lynn Claudy, who told us that determining spectrum "availability" wasn't a simple task. "You can't just look at a channel lineup for a community, see that half the available channels are empty, and say that the white spaces are wide open," he said. (The group has an entire website devoted to pitching the idea that white space devices will create "interference zones.")



Instead, making the calculation is a matter of setting thresholds that have as much to do with politics as engineering. Because radio frequency signal propagation can "vary wildly over time," especially as one gets farther from the transmitter, the question really becomes, says Claudy, "How reliable do you want the signal to be at what percentage of time?"

The broadcasters want over-the-air signals to be bulletproof, without the slightest danger of being interfered with. After all, the signal is the bread, butter, jam, cup, plate, and silverware of local TV stations; if transmissions are fuzzy, snow-covered, or otherwise interference-laden, people won't watch and ad revenues will drop.

According to Claudy, when you factor in the reliability that broadcasters want to see, urban areas are actually "pretty clogged up." In major cities, interference possibilities become "to us more risky than we would want to see."



Ed Thomas, a former FCC chief engineer who now advises white space backers, like the Wireless Innovation Alliance, says the reality is that there is "a lot of vacant stuff out there," especially in rural America.

And Sascha Meinrath of the New America Foundation calls the entire spectrum availability issue "absolutely ridiculous." The spectrum is only booked up, he notes, if the FCC rules out sticking a white space device on an empty channel adjacent to a TV channel. In fact, so concerned is the broadcasting industry about interference that it has asked for a "third adjacent" rule on white space devices, should they be allowed; a device would need two empty channels on either side of its transmission channel before it could broadcast (for a total of five empty channels).

The issue is tremendously important. If white space devices truly need to find three (next adjacent) or five (third adjacent) available channels before transmitting, their usefulness would be curtailed dramatically.



To see if interference was a concern, the New America Foundation's Wireless Future Program commissioned a study from the University of Kansas Information and Telecommunication Technology Center (ITTC). Sounds technical and science-y, right? Onward, empiricism! How could anyone argue with the results?

But when the report came out in early 2007, "argument" is exactly what happened. Lynn Claudy says that the NAB looked at the report, and the broadcasters approved of its message that "you must have protections for the adjacent channel." If you don't, there's the risk of interference to at least some of the millions of over-the-air TV receivers out there.

Meinrath, of the New America Foundation, was incredulous when I read him the quote. "He lied to you," he said. New America, which commissioned the report, had numerous conversations with the University of Kansas researchers, and Meinrath remains convinced that he understands the report perfectly. Its conclusion, he says, is exactly the opposite: properly designed devices of 100mW or less did not cause interference to TV signals on adjacent channels during the testing.

When the two sides can't even agree on how to read the results of a technical report, it's obvious that no mere survey, study, or test can, on its own, put the issue to rest. But that doesn't mean that the engineering details aren't important; convincing the FCC's engineers, at least, becomes a crucial part of the entire campaign.

Spectrum sensing

To deal with interference issues, current proposals on the table could require white space devices to use three separate techniques to identify open spectrum before broadcasting: spectrum sensing, geolocation, and beacons. This tripartite approach wasn't initially favored by white space backers, but companies like Google and Philips and Microsoft have been willing to compromise in the hope of dialing back the opposition from groups like wireless microphone manufacturers.

Spectrum sensing was the first proposed solution and is simple enough to understand: The white space device would scan the airwaves before transmitting. It would only broadcast on a channel found to be empty, assuming that the spectrum sensing is accurate. The NAB worries that it might not be.

The reasons are threefold. First, there's the possibility of bad engineering, in which the device in question simply isn't properly sensitive to TV signals (especially distant, low-power ones). This is what the FCC's Office of Engineering Technology has been testing in its labs and in the field. Results from the first round of testing last year were promising, but problems remained.

Some of those problems were later attributed to broken prototype devices, which brings us to the second broadcaster concern: reliability. It's all fine and good if the devices work when they're released, but what happens when devices break or have "power issues" like the prototypes in question did? The FCC is not testing device reliability; in fact, it isn't even working with production-ready devices.

The third concern is that, even if everything works, interference might still result due to things like the "hidden node" problem.

Arthur, Merlin, and the case of the hidden node

Imagine two hills, separated by a valley. On top of the left hill is a TV transmitter, and on top of the right hill, a man watching football on his TV. We'll call him Arthur. In the valley between them lies the home of Arthur's nemesis, a white space devices user named Merlin.

Arthur's TV reception is great, except when Merlin turns on his white space box to get himself some hot wireless broadband action. The box works perfectly, it uses spectrum sensing, but it still causes problems. Why?

Because the hills attenuate the signal. When Merlin's box scans the airwaves, it sees nothing because it's in the valley, and the TV signal falls below its detection threshold. It transmits on the same band that the TV station uses, and when it does so, Arthur's football game gets scrambled. The same thing could happen more easily in an urban canyonland like Manhattan.

Lynn Claudy, of the NAB, says that problems like hidden nodes mean that spectrum sensing is simply based on "a false premise." Just because a channel is clear at the moment, and from a particular location, doesn't mean that other viewers won't be affected by a broadcast on that frequency.



Wally is about to be flagged for interference

Of course, this is the point at which the debate becomes contested once more. Ed Thomas says that the hidden node issue is real, but the only thing a hidden node does is create a lower-level signal. One way of addressing the issue, which white space devices do, is to bake in 30dB of margin.

The threshold for digital TV visibility is -84dBm; below that power, no signal can be displayed. But white space devices detect the presence of signals all the way down to -114dBm, and Thomas says that the huge cushion of margin will take care of hidden node issues. The devices could potentially do sensing at even lower power levels, but at that point, the noise starts to generate false positives.

New America's Meinrath also points out that the transmitting and receiving nodes on a white space network need to indicate a clear channel before data transmissions will begin. Even if one node is located in a hidden node, the other node it must connect with probably won't be.

The odds of interference in this situation are "astronomically small," Meinrath says, in the same realm as worries that "a TV station could get hit by a meteorite." It might, and that would make for some compelling TV news footage, but it's hardly the sort of stuff policy makers worry about.

Geolocation

Geolocation is another potential solution to the interference problem, and it is essentially GPS on steroids. Each white space device would take a location reading and consult a database; the database would tell the device what channels are currently open for use in the device's particular location.

Spectrum sensing would still be used before transmit, but the geolocation system would curtail the hidden node issue. It's not a perfect solution, though; how does the device access the database without having Internet access yet? Would the database be built into devices? Would it be properly updated? Who would host the online, fully-updated database? How far from a transmitter does a white space device have to be before the channel is considered "safe for use" by the database? Etc.

But it's a technique that everyone agrees would work. Lynn Claudy says geolocation gives the devices "a better chance of operating in a benign way." Ed Thomas says that his group believes that sensing alone is good enough, but "politics" showed them that the FCC needed additional options.

If geolocation is required, Google has offered to host the database of available channels (sorted by latitude and longitude) free of charge, though some versions could also be loaded into devices to get them up and running.

But here's where the New America Foundation parts company with white space backers like Google. Meinrath's worry is that geolocation creates a network dependency (to check the database), but the "future of networking is opportunistic" and ad-hoc. Nodes on the edge of a network need to have the right to make their own connections with other nodes, he believes, without checking it against a central server.

That freedom is key to disruptive innovation, in his view; just imagine how WiFi would have done if every wireless router had to go through some sort of geolocation process before it began to function.

Beacons

Ed Thomas says that we can think of spectrum sensing as the belt in the white space ensemble, while geolocation is like a backup pair of suspenders. Following this metaphor, local beacons would be... the garter belt? The cummerbund?

White space backers are now proposing a third system to prevent their devices from transmitting in ways that would interfere with TV. Wireless mics, many of which are (illegally) unlicensed and operating in churches and theaters around the country, have the potential of being disrupted by white space devices transmitting on "clear" channels.

An argument has been raging in DC over this issue for months, largely on the grounds that the mics in question are being used illegally, and hey, we don't give amnesty to anyone who breaks the law, do we? But that approach has proven politically infeasible, and one new idea for making the mics coexist with white space devices is local beacons (Motorola is one of the biggest backers of the idea).

Venues could install these inexpensive transmitters to warn white space devices in the area that specific channels are off-limits. The beacons work by mimicking a DTV pilot tone, making the channel look occupied.


Page at New America

Despite the antagonism between mic makers and white space backers, the wireless mic debacle has actually provided a talking point to the white space crowd. It's one that Larry Page himself used on his visit to DC earlier this year: if all these microphones have been operating for years in the TV spectrum without a license (and without reports of interference), why not allow white space devices, too?

The NAB has an answer to that, of course. Wireless mics exist only in certain spaces—theaters, churches—not the apartment on the floor above you and the apartment on the floor below you. Putting these devices in the hands of consumers, allowing them to be installed willy-nilly, it all smacks of reckless disregard for the broadcasters and their businesses, businesses that (after all) alert the public to tornado warnings and show waterskiing squirrels on the local news.


Don't tread on me

Off-the-record conversations with wireless mic makers, like Shure, indicate that they also hate the white spaces idea with a passion usually reserved only for wired microphones. (You can see Shure's many, many thoughts on the issue at its website, but the company has a basic disregard for the RF engineering prowess of white space backers and worries that millions and millions of dollars of hardware could be disrupted.)

The FCC has already ruled that wireless mics in the 700MHz band can't be sold after the digital transition, but some groups estimate that as many as a million devices are already in the wild.

Interference with... cable?

Fortunately, all of this only applies to over-the-air TV, the sole TV source for a mere 13 percent of US households, and cable has nothing to fear, right? Wrong, says Big Cable; white space devices could wreak havoc on cable systems.

The National Cable & Telecommunications Association filed lengthy comments with the FCC on September 10, concluding that "unlicensed TV band devices, as currently proposed, will interfere with cable service." Not only can home cable wiring be subject to direct pickup (DPU) interference, but local cable offices around the country could find their high-gain local antennas affected (which could mess up local TV channels served to all cable subscribers in the area).



According to NCTA analysis and some FCC testing, cable systems can pick up noise from 4.3mW white space devices at a distance of two meters. But current white space device proposals want the ability to use 100mW of power. Here's how the NCTA describes the potential problem:

"To put this in perspective, every time a consumer in a single family home uses a personal/portable TV band device as currently proposed, its signal output will interfere with cable services. For example, a family member using a TV band device in one room for home networking could foreclose another family member from watching a particular TV channel in another room. The affected channel would go blank or be seriously degraded. The idea that a consumer could simply put more distance between the device and the television set is not a solution. As our studies have shown, even in a single family home, the distance that the consumer would have to maintain between the device and the TV receiver to ensure non-interference is impractical."

The problem could get even worse in apartment buildings, and the group points out that even Ed Thomas has conceded that the issue "needs to be looked at."

NCTA also notes that many cable headends rely on high-gain antennas to pull in distant TV signals, in some cases from more than 100 miles away. This means that the TV signals in question are quite weak and susceptible to interference.

The NCTA says it doesn't want to be obstructionist, it likes the basic idea, but it does want to protect its members' billions of dollars in investment. To address the problems, the group would like to see unlicensed white space devices limited to 10mW, or one-tenth the amount of power that white space backers want to use. As for headends, NCTA wants to use the geolocation database to clear a swath of space around them that would be free of white space devices.

New America's Meinrath doesn't buy it. "Cable systems are supposed to be closed loops—they're not supposed to bleed (either into the surrounding RF environment or in from that environment). If cable installers have failed to properly terminate their lines, that would facilitate RF pickup. Likewise, if providers have used such cheap gear that they are not properly shielded, that too would cause a problem. In either case, the failures lie with the cable industry—foregoing a useful new technological innovation (in essence, rewarding the profit-maximizing, corner-cutting that the cable industry alleges is a problem) doesn't make any sense whatsoever."

From the laboratory to Broadway

FCC testing, not usually of great popular interest, has become surprisingly fascinating this time around. We're not just talking nerds in the parking lot any more; instead, the FCC has traveled to NFL stadiums, Broadway theaters, the Patapsco Valley State Park, and even an observation area at Baltimore/Washington International airport.

Microphones have been one of the key concerns, despite questions about the legality of many applications. The FCC's Office of Engineering Technology recently wrapped up white space trials at a Broadway musical at New York's Majestic Theater and at a football game at Maryland's FedEx Field to see whether the devices could properly detect occupied channels. Both sides claimed victory.



A white space test device

Tests were also run at urban and suburban sites around the DC area to test both the spectrum-sensing capabilities of the protoypes and their potential for interference. Motorola and Philips have both said publicly that the tests were successful, but companies like Verizon have continued to warn that the technology is unproven.

As the FCC testing went on for week after week, the agency repeatedly warned news outlets and participants that the results would not be presented as "pass/fail." Which brings us back to our initial point: doing well in the technical trials is necessary but not sufficient. Assuming that the final OET report on the white space prototypes doesn't contain language like "rampant interference," "fry the brain," or "scramble every television set in a five-mile radius," the real battle to open the white spaces for unlicensed use truly begins.

That battle will be quick and it will be brutal. In other words, it will be political.

The technical has become the political

And it won't be "political" in the sense that a freshman seat on the student council is "political." The powerful broadcasting lobby is lined up in opposition, for one thing, but various other large spectrum users are also expressing uneasiness. On the other side, Google, Microsoft, Philips, Motorola, and a host of other companies are pushing white space devices forward, and that's saying nothing about the public interest groups that support the idea.

You can get a sense of how the battle is shifting by looking at Google's example. While the company has, for quite some time, collaborated with white space backers in the hope of creating more access to broadband (remember, it offered to host the geolocation database), it recently moved into an explicitly political phase of the campaign with a new "Free the Airwaves" website and petition. As we noted a few weeks back, the petition attracted more than 13,000 signatures in its first week. More importantly, it marked Google's first official use of its name and financial resources to enlist users in political action.

The goal of all these signatures isn't technical; it's designed to put pressure on the FCC Commissioners who will vote on the matter once the OET report is finalized. And Google is not the only one trying to bring this sort of pressure. Eight Congressmen last month sent a letter to FCC Chairman Kevin Martin, telling him that they "look forward" to the agency setting up rules within the next 90 days "to open the white space spectrum and provide adequate protection to existing users."

Assuming that unlicensed white space access becomes a reality in some form, there's no guarantee that it will provide a major new source of broadband competition. Backers certainly hope so, especially for poorly-served rural areas that rely on wireless solutions, but prominent analysts, like Blair Levin of Sifel Nicolaus, aren't convinced.

"There’s not that much left to be disruptive," Levin said at a recent conference, according to Telephony Online. "White spaces could be in rural areas, and a little bit in broadband, but I don't think so."

The beauty of unlicensed devices, though, is the possibilities for unexpected innovation that they create. When inventors don't need to ask permission or pay millions of dollars for licenses before proceeding with a cool new application for the technology, we've traditionally seen tremendous growth (think Internet apps and WiFi, to name two similar areas).

Will the TV white spaces get colored in with innovation? We'll know soon enough. The OET's report is expected within the next few weeks, and it's possible that an FCC vote will happen later this year.

Not that a vote will be the end of the matter. The broadcasters have previously shown themselves willing to go to Congress to have FCC decisions overruled. This time, though, they will be facing some influential and deep-pocketed tech titans, rather than community activists, so any Congressional fight could be bruising and inconclusive.

"Interference" or "competition"?

Judging motives is tough, but one of the complaints commonly heard about the broadcasters is that—despite the huge mound of technical documentation filed in the docket so far—the basic issue isn't about interference at all. It's about competition.

The thinking goes this way: even if white space devices work flawlessly, the broadcasters still stand to lose out. Video content migrating to the Internet (think about what Hulu and iTunes mean for the long-term viability of local, over-the-air broadcasters) means that the creation of a new, unlicensed pipe in one of the premium spectrum bands isn't in the broadcasters' interest. Critics have also long assumed that the broadcasters want nothing better than to shut down cheap broadcasting in their band, but opening up the TV band to unlicensed use could make this easy (nothing says that a white space device has to transmit only TCP/IP packets).

We've been here before. Battles about "interference" can easily be little more than holding actions designed to shore up an incumbent business model; our resident media historian, Matthew Lasar, points out that the same debate happened over "low power FM" (LPFM) broadcasting. LPFM was designed to allow the inexpensive licensing of local, low power stations, but it was vehemently opposed by radio broadcasters on interference grounds. The NAB won the fight by convincing Congress (not the FCC) to pass a "third adjacent" rule that effectively neutered LPFM stations. Congress is revisiting the issue after further engineering studies cast doubt on the interference claim, but the bill is nowhere near the finish line.

Then there's the current fight over a free, nationwide, smutless, wireless Internet service. A battle is now raging at the FCC over whether the service, to be offered in the 2-2.2GHz band, might interfere with some of T-Mobile's own spectrum holdings. That may be true, and the FCC has just finished looking into it, but airing the controversy certainly has the side benefit of stalling the proceeding.

None of this is new. As Lasar reminds us, you can go back to the days just after World War II, when the FCC relocated the entire FM band on short notice and gave much of it to TV. The worry was that sunspots might interfere with radio broadcasts (long story), but the radio folks suspected the foul handiwork of TV executives.

A recent article in The Wilson Quarterly nicely describes the aftermath of the decision, which required the public to buy entirely new receivers.

"The FCC said it wanted to boot FM up-spectrum 'before a considerable investment is made by the listening public in receiving sets and by the broadcasters in transmitting equipment.' In truth, a considerable investment had already been made. As of mid-1945, there were 53 FM stations, and they were broadcasting to a half­ million receivers. The change of technical standards meant that broadcasters' and consumers' equipment would become obsolete soon and suddenly.

"Perhaps coincidentally, perhaps not, the chair of the FCC resigned to become general counsel at NBC, a division of RCA, after the spectrum decision. It later emerged, too, that RCA had sent FCC commissioners free televisions during the FM hearings."

The line behind "interference concerns" and "competitive worries" can be a thin one, and even technical battles are political ones in Washington.