Open Spectrum: The New Wireless Paradigm

prostoalex writes ""Almost everything you think you know about spectrum is wrong." - starts Kevin Werbach in his working paper Open Spectrum: The New Wireless Paradigm. He touches the possibilities of using open spectrum, and then dwells on such innovative products like software-defined radios, spread spectrum or cooperative wireless networking. Truly informative insight into where the U.S. government stands on the issues of wireless spectrum, where it should be, and how it will benefit society and individuals."

Comments

not everything

[dirvish]

One thing I do know about it that I am fairly certain is not wrong: the government will attempt to regulate it. It doesn't matter what shape it takes the government will make sure it makes a profit from it.

Too Little, Too Late

[WittyName]

Even if the FCC goes along with this, the telecom industry can not afford to roll it out.

They are heavily burdened by debt from Cell and cable modems.

No more!

[t0qer]

I swear I think all this extra radiation is adversly affecting my health. (places tinfoil hat on head)

Sure

[ArchieBunker]

This has about as much chance as the US switching over to the metric system tomorrow.

With public domain frequencies...

[kpansky]

there are some really interesting possibilities. Think of software-defined radio letting you join different wireless networks that are decentralized and encrypted using something ala freenet. If you were to couple unfettered access to wireless frequencies where people's ingenuity sets the standard, and not self-interested corporations.

I would personally love to see open hardware designed to utilize wireless technology available similar to the projects at OpenCores.

Re:With public domain frequencies...

[NanoGator]

"Think of software-defined radio letting you join different wireless networks that are decentralized and encrypted using something ala freenet [sourceforge.net]. "

I have a question: If you have two different sources broadcasting digitally on the same frequence, then you can seperate the two based on encryption standards or protocols, right? However, wouldn't having both transmitters sending data jam each other's signals? Seems to me like they'd have to hear each other and cooperate in order to work efficiently.

In other words, I don't think you can set up two 802.11 nodes with different SSID's and have them work at full capacity.

I'm really naieve here, so if anybody can explain to me if I'm wrong (or even right) I'd really appreciate it. My education on this topic is very basic.

Re:With public domain frequencies...

[chriso11]

Yes you could put two 802.11 nodes together - an 802.11a @ 5.7GHz and and 802.11b/g @ 2.4GHz.

As for spectrum coexisting, DSSS (discrete sequence spread spectrum) was originally developed for military usage. DSSS provides several advantages - extremely hard to detect or jam, and works very well even when other DSSS transmissions are occuring in the same region.

DSSS used to be extremely difficult to implement, but now is commonplace and cheap.

The whole reason that the radio spectrum is used is because it is an effective way to send information. Technology has provided a way to transmit significantly more information in less bandwidth, except that it is against the law...

An example would be: Suppose that the transportation industry was regulated in the same manner. In the old days, the only delivery option was horse-drawn cart. Now technology has developed trains, planes, and trucks. However, only a few routes are legally available for the trains/trucks/planes to go on - everything else is still regulated to use horse-drawn carts. Obviously, there would be significantly less deliveries across the country.

If I understand correctly

[drhairston]

While the author is correct in stating that modern advances in digital spread spectrum allow digital signal processors to place multiple signals in the same frequency, he ignores the impact this would have on existing analog technologies which are incapable of this feat. I personally do not wish to be walking down the street the day it starts raining model airplanes.

Re:If I understand correctly

[JaredOfEuropa]

He does suggest to limit this to certain bands, such as the current 2,4GHz band, or to do this in existing bands using extremely low power. Neither should affect existing analog technology such as model airplane transmitters.

The idea that spectrum is a limitless resource is a myth, however. The author does say as much somewhere, but mostly he talks as if the spectrum will bear an unlimited number of users, which is bollocks. Having lots of spread-spectrum devices in the same area and frequency band will affect the performance of these devices. If that was not the case, why the hell would we even need to follow the author's suggestion to open up more spectrum for such use beyod the current 2,4 GHz band?

With that said, current analog transmission technology does not make a good use of the spectrum, and assigning a whole band over a large area to one transmitter is terribly wasteful. While I disagree with the picture the author paints in his article, I endorse his plea to make more use of spread spectrum and to assign more bands to unlicenced use.

Re:If I understand correctly

[mesocyclone]

I wouldn't say he ignores it, but there is one assumption about underlay that is incorrect. The proposed mechanism of underlay is for the the underlay system to detect that it is causing interference and cease to do so within a few milliseconds.

However, the ideas of Open Spectrum, applied carefully and with good engineering understanding, can indeed significantly increase the usage of spectrum. However, unlike many of the subsequent posters on this thread will assume, it will *not* work without appropriate regulation (as the author recognizes). It is a different and superior method of spectrum management, not spectrum anarchy.

The idea of setting aside spectrum "parks" for the unlicensed services, and then applying strict technical regulations to those systems is the most promising. Setting aside channels 60-69 would free up 60 MHz of very useful bandwidth for mobile and portable applications. The appropriate standards would allow proper sharing of the underlying spectrum without licensing individual users or sites.

However, some of the techniques that allow this sort of operation may not be that inexpensive to create. They will require substantial processor power and probably power consumption. This will limit their use in extremely inexpensive uses (such as keychain transmitters for auto alarms). For these kinds of uses, different spectral parks may be required.

In other words, one may need some spectrum for dirt cheap devices (where an additional $.01 is a significant cost increase), and other spectrum for sophisticated devices where the value allows greater costs. Likewise one may want different spectrum and rules for wide area systems than local ones. Furthermore some systems can tolerate significant random interruption (remote meter reading, for example) while others must work well all of the time in real time (police communications, air traffic control, etc). These may again require different parts of the spectrum in order to be protected from inadvertent interference from nearby non-cooperating unlicensed systems.

Furthermore, one needs to make sure that failure modes of these devices don't screw up a whole area!

Hmmm... this starts to sound a lot different from just turning folks loose on unlicensed bands! It illustrates the complexity and the need for sophisticated standards and associated regulation.

Underlay is IMHO much more dubious than the author lets on.

In practice it can be very hard to do well. Existing narrow-band systems use techniques that *cannot be used* in wideband receivers. The techniques (such as very high Q low loss RF filters) allow the narrow band receiver to operate with very weak signals, signals which could not be adequately detected by a wideband underlay system, and which would then be interfered with by that system. These existing systems are engineered, and regulated, to use the minimum power needed - and thus are inherently susceptible to this new interference.

There are a number of physical factors that limit receiver sensitivity. They range from thermal noise in the receiver to exotic topics such as intermodulation, desensitization, and quantization noise. It is not possible to optimize for all of these in a frequency agile receiver to nearly the degree one can in a narrow band receiver.

In addition, some of the techniques are inherently expensive. Moore's law doesn't apply to the fabrication of precision metal resonators, for example.

What this means is that for an underlay service to be truly non-interfering, it requires either a very expensive, big and power consuming receiver, or it needs to be on a portion of the spectrum where these techniques are not applied by existing users of that spectrum.

Other approaches, such as using spread spectrum - or wideband as the autho prefers -(so you don't have to detect systems you are interfering with) have different problems. A wideband system distributes its power across a wide spectrum, but that power is still not zero. This means that if it is too close to a traditional narrow band receiver, or another wideband receiver it will cause damaging interference.

Overall, however, the Open Spectrum initiative is a good thing and can have enormous economic value. But it should not be viewed as a magic solution or one that can rely strictly on anarchy or unregulated cooperative development.

Re:If I understand correctly

[WolfWithoutAClause]

The author does say as much somewhere, but mostly he talks as if the spectrum will bear an unlimited number of users, which is bollocks. Having lots of spread-spectrum devices in the same area and frequency band will affect the performance of these devices.

No. I think he's right. Certainly, if you reduce the power you transmit to the absolute minimum to reach just the nearest nodes then the bandwidth goes up. Then if you route through the other nodes, the bandwidth goes up. Direction aerials, bandwidth goes up. Optimal filtering to make good effect of the reflections in the room, bandwidth goes up. I truly think that the available bandwidth is very, very, very large. Sure it's lots of extra hardware, and there's lots of processing, but they're getting cheaper and cheaper. Also, as the users, and if properly designed, the room, tends to absorb microwaves, then the amount of bandwidth scales with the number of users- the noise floor just simply doesn't keep going up, since the users get in the way.

Re:If I understand correctly

[chriso11]

Actually, you need to be aware that the unlicensed bands ARE regulated - for example, you can only use channels a certain width at max (say 1MHz), and need to use a spectrum conserving modulation, such as FHSS (frequency hopping spread spectrum). You also have maximium tranmission power requirements and so on.

Re:If I understand correctly

[argoff]

I think you're thinking of it wrong. You need to think of it like 50,000 people talking in a stadium. Yes it will be more hard to talk to your neighbor, yes you may half to yell a little louder, or talk directly torard his ear, but you can still do it and exchange information without regulating everyone in the stadium. Also, when talking, you likely have plenty of incentive to yell and direct just loud enough to make sure your neighbor gets your message without alot of wasted effort.

The FCC's attempt to regulate this is awfull. It is like no one is alowed to make a peep - accept the people with the FCC megaphones. If people did that with speech, we would see it like one of those pro fidel castro rallies, but when we do it with spectrum (which physics wise is the same) people just take it on faith that without the FCC - we would all be ruined. It is very disheartening.

Not unless they refactor the laws for tort ...

[LL]

(tort is the legal action against harm within common law jurisdictions). Let's consider a future where individuals have wireless wheelchair, interacting with their environment (like doors/cars/etc). How do you prevent individuals being careless (cf case of leaving a concrete block on road for car to hit)?

Courts (in torts against trespass to chattels w.r.t. deciding spam cases) have rules that having an electronic signal impacting on the computer system is sufficient to be "in contact". Now extend this to a generic wireless world and you can see the potential combinations of potential problems. If my wireless car activates someone elses' garage leading to a theft then are you liable? Medical instrumentation are a major concern, as are anything which records ownership (cf person entering building with wireless and downloading trade secrets).

Wireless will change how we interact provided we can sort out how social responsibilities and obligations are partitioned.

LL

Re:Not unless they refactor the laws for tort ...

[chriso11]

If your wireless car opens someone else's garage and they get robbed, then yes, someone is liable - the manufacturer of the garage door opener. Unless you send the correct sequence, that garage door shouldn't open.

What is your point on the wireless wheelchair? I think that interference concerns are overblown. Circuits need to be designed with care to insure that they aren't susceptible to interference, but they need to be designed that way now. You have to deal with cell-phone towers, microwave ovens, and even fluorescent lights.

Re:is it just me

[Waffle Iron]

No, only people who use "paradigm" in a business or technology-related manner. It's a perfectly valuable philosophical construct when used properly.

So, if I understand you correctly, you're saying that paradigm is ok under the old paradigm for paradigm, but the new paradigm paradigm is unacceptable.

Contradictory

[SiliconEntity]

The paper goes on and on about how WiFi shows that open spectrum works, there's no problem with congestion, no need for ownership or frequency allocation. Then in recommendation 2 we read this:

Improving existing unlicensed bands isn't enough. Most are so narrow and congested that their utility for open spectrum is limited.

So here he's saying exactly the opposite, that congestion is a serious problem for open spectrum! Which is it?

It's also bogus to claim that WiFi proves that open spectrum works. The truth is that WiFi is so sparsely implemented that congestion hasn't yet been an issue. For all the hype, my town of 175,000 people has no wireless public access points. Even in the big cities they're not so close that congestion is a problem.

The article could just as well have used cordless phones and baby monitors (which use the same frequencies) as evidence that open spectrum works. The only difference is that they don't score as high on the hype meter. All these examples prove is that the technology works when the range and distribution of the transmitters is sufficiently limited, which everyone knew already.

Re:Contradictory

[WolfWithoutAClause]

So here he's saying exactly the opposite, that congestion is a serious problem for open spectrum! Which is it?

It's a bit of both. The existing protocols lack some features you really need to make best use of the bandwidth, for example, node routing and power control are both absolutely critical, and WiFi does neither out of the box.

Human Nature and The Status Quo

[Iguanaphobic]

It is time to question our long-held assumptions, and explore new policy approaches that could generate tremendous benefits for the American people.

He had me until this sentence. Getting the government to act in the best interests of the American people is not on the agenda. The current licensing scheme guarantees profit for the few. The few, in turn, guarantee $$$ to the government. Anything that threatens this simply will never happen.

Software Defined Radio

[Anonym1ty]

Software-Defined Radio Every radio can be tuned to pick up a certain range of frequencies and it takes some amount of time to change the tuning. Traditionally, these characteristics are fixed in the radio hardware. Thus, for example, the same radio can't pick up both FM radio and mobile phone transmissions, or both 2.4 GHz and 5 GHz wireless LAN signals. Software-defined radios, by contrast, can tune dynamically over a wider range of frequencies. A software-defined radio can receive or transmit different kinds of wireless transmissions automatically. If it is a so-called "agile radio," it could adapt to the local environment and seek out open frequencies to communicate. Even in licensed bands, most of the spectrum is empty most of the time. Agile radios could take advantage of that empty space, moving out of the way when another transmission appears.

Um, like uh ok.

Are you likening a radio with a wide receive and transmit to a um Win Modem?

I understand what you mean but I really hate your terminology. Software defined agile radio, OMG lets just make all this kewl stuff sound like a pansy made it.

This really isn't anything new the idea in some forms exists now and has for some time and has been implemented. The only new thing is trying to convince the FCC to allow you to work in unused TV channels and other under-utilized bandwidth because your equipment is smart enough to know there isn't anything in there... ...kewl

Coulda said that in one line rather than such a verbose article.

This makes sense..but probably won't happen....

[Newer Guy]

Why? Our whole enterprise system is based upon monopolies. Look at the FCC these days... they've been making the monopoly of the airwaves BIGGER! In most cities, instead of having 15 or 20 different radio owners, it's now down to three or four. The U.S. radio spectrum policy is based as much (if not more) on politics as on physics and technology. As long as it stays this way (and it will, trust me), the status quo will prevail. The unlicensed bands are few compared to the licensed (read: monopolized) ones...they are less then 1 % of the electromagnetic spectrum. This isn't likely to change, either.

Yes, the RF can be quite bad

[Ed Avis]

I'm afraid it's true. Open Spectrum beats current wireless networking technologies hands down.

The clock frequency of the Z80 processor is well-optimized for generating long- and short- range RF waves. With the Spectrum case Open, the highly advanced rubber keys act as efficient waveguides, particularly helped by the printed-on aerials (square shapes with one half or one quarter cut out, depending on the phase of the waveform).

The only snag is that due to an unfortunate typo, the Sinclair Microwave communication device ended up as the bizarre Sinclair Microdrive.

Obligatory Simpsons Quote!

[VistaBoy]

Writer: "'Paradigm' and 'Proactive'...aren't those just buzzwords that stupid people use to sound smart?"

*Blank looks from the managers*

Writer: "I'm fired, aren't I?"

Manager: "Yes."

Open Spectrum means chaos

[Dolphinzilla]

If you ever work with Satellite communications systems (spread spectrum included) you will quickly find that even with all the so called "Draconian rules" in place, there is a high probability of someone stepping on your signal and satellite channel - the same has been my experience with other more conventional line of site RF systems - The ONLY way that whis would ever be feasible is if we took all the existing systems and simultaneously destroyed them before wheeling out the "Open Spectrum" systems. I for one don't believe that "free for all" RF spectrum is ever going to be practical or desireable. Does anyone else remember the days of the CB radio arms race and home built 100 W linear amplifiers ? As hard as it is to accept government regulation this is one area we don't want to let go to anarchy - if we do all we'll hear on our fancy Spread Spectrum radios will be static.

Humm

[Wyatt Earp]

I read the .pdf and then I popped over to the New America Foundation and looked at the Senior Staff bios.

Wow I am so stunned at why anyone gives a rip what these jokers put out.

The head guy, the dude the Washington Post profiled last year, Ted Halstead - President and CEO, he doesn't do research. I work with people that don't do reseach, I call them the Grounds Crew.

He went to Harvard woohoo, Presidents Bush and W Bush went to Yale, having exclusive school deploma doesn't mean one is a genius. Harvard Business School didn't consider anything Internet to be "business" until about 1998-99.

Ooh Hollywood types and CongressCritters like them. Another nail in the coffin of respectablity.

The head joker at NAF is the buttmunch that told Warren Beatty to run for President.

"Previously, Mr. Halstead was Executive Director of Redefining Progress, another public policy institute that he founded to promote new approaches to economic and environmental policy." - Thats alot of words to say "He sat around and talked about cloud-cookoo-land."

"Kevin Werbach" - the guy that wrote the paper linked here - "is a technology consultant, author, and founder of the Supernova Group." He also has some 'leet HTML skillz - http://werbach.com/home.html - He uses a Mac, a point in his favor.

They throw out buzz words and do 20 pages and we are suposed to care why?

Re:No it is true!

[IIRCAFAIKIANAL]

*insipid childrens show music*

Welcome to the Slashkids Fun-Filled Fallacy Post!

Slashkids:Yay!

Today we're going to learn about The Post Hoc (Ergo Propter Hoc) Fallacy!

Slashkids:What's that?

Well kids, that's Latin for "after this therefore because of this."

For example, you could easily say that eating breakfast causes car accidents, because most people that have had a car accident had breakfast that day.

Slashkids:Ha ha ha!

Or that you started a new job, and then your hair fell out.

Slashkids:Then what caused the hair loss?

Who knows? It could be something in the water, it could be any number of other factors. How do we avoid this type of thinking?

Slashkids:With control group studies, double-blind and random tests!

That's right!

Remember Slashkids, sequential patterns != causation AND correlation != causation!

Next lesson: The ad hoc hypothesis

Slashkids:Yay!

*insipid childrens show music, credits, cut to commercial*

(Sorry, I'm a prick and I couldn't resist :)

Learn from amatuer packet radio ...

[Tjp($)pjT]

There are a host of problems with ad-hoc networking schemes such as this. I have a few solutions to them, but mostly they would be overpriced in the consumer market. One such beast that will start to bite you on the * really hard is hidden transmitter syndrome . This is where your controlling nodes overlap, and say node A is visible to base B and base C but base B and Base C are out of each others range. If B and C don't "hear" each other they can't work in unison without a third party. So when Node A (mobile presumably) hunts for a connection, both B and C try to talk at the same time to A, thus hampering the usable bandwidth. Lots more lessons can be learned by just searcjing for the syndrome and reading the other problems mentioned around it.

Re:is it just me

[Jester99]

No, only people who use "paradigm" in a business or technology-related manner [talk out of their asses].

Hm. I'll have to disagree with you there. Webster's defines "paradigm" as "An example; a model; a pattern." When you think about it, a lot of both business and technology follows patterns. The giving-out-handles-selling-blades method that worked so well for Gilette, frankly, is a "business paradigm." That business model's been tried again and again since then (Compare to today's /. article on selling XBoxen as loss leaders to gain $$$ on selling games).

And in computer science, there's many different ways of writing programs. For instance, take "object oriented programming." If you use OO principles when engineering your software, you're using an "object oriented paradigm." I see no harm in a scholarly discussion of computer science using the term in this manner.

Yeah, sometimes people throw it around as a buzzword. But don't dismiss a term's applicability to two entire fields just because a few people try to sound impressive by repeating it.

Protocol smarts can solve this.

[Ungrounded Lightning]

... say node A is visible to base B and base C but base B and ase C are out of each others range. If B and C don't "hear" each other they can't work in unison without a third party. So when Node A (mobile presumably) hunts for a connection, both B and C try to talk at the same time to A, thus hampering the usable bandwidth.

But mobile node A can BE the "third party". If base stations B and C have unique (or at least non-colliding) identifiers then mobile node A can say, as part of its transmission, "I'm talking to B." or "I'm talking to C."

There are variants that work even if mobile node A is saying "Who's out there for me to talk to?". Look at the IP address resolution protocols for working examples. For starters, the a similar case arises on an Ethernet when a machine that wants to be booted uses RARP, broadcasting its Ethernet address and asking potentially redundant servers for its IP address. (Granted the servers COULD hear each other. But the protocol works even if they can't.)

That misses the point of the paper

[Chazman]

Yes, if both transmitters are broadcasting at the exact same frequency with the same modulation scheme, they interfere with each other. The point of this paper is to ask why are we cramming ourselves into just a few tiny nooks of the spectrum, thus creating this interference problem, when the spectrum is so large and underutilized. The paper points at the various incarnations of 802.11 as efficient use of a small patch of spectrum. If we did that with a few more patches, we'd easily have enough space for colocated wireless networks to work around each other -- change their frequency utilization automatically based on what other networks are nearby to minimize interference. 802.11b has such an option already; it's called channel agility. Unfortunately, since there are only three truly non-overlapping channels (1, 6, 11) within the USA/FCC mandated spectrum for 802.11b, areas of heavy wireless activity can easily find themselves without a clear channel to switch to. However, with say, a few dozen nonoverlapping channels to choose from, channel agility becomes a powerful and effective tool in colocating many wireless networks with a minimum of interference and degradation.

The rules of and assumptions underlying the way the FCC is carving up spectrum are based on 1930's technology. It assumes transmitters and receivers have poor filters, and cannot tolerate adjacent or overlapping signals. It assumes no spread spectrum or channel agility / frequency hopping technology. Fast forward 70 years. Technology has marched on. Spread spectrum and channel agility are cheap and commonplace. Transmitters meet much stricter tolerances for sideband and out-of-band emissions. Receivers can pick up weaker signals, and much more successfully distinguish their signal from other overlapping or closely adjacent signals. Thus we can now pack several times the data per unit of spectrum than the current rules assume we can. Yet the rules prevent us from doing this on a large scale because the unlicensed bands in which we can operate are so few and small. Users of the licensed bands (most of them anyway, cell phones being the one big exception) have little incentive to deploy these technologies and make maximum use of their spectrum because the rules guarantee them enough free spectrum that they can use older, less efficient technology with abandon, and still get done everything they want to.

THAT is the point of this paper. We shouldn't be asking what if both nodes are at 2.4GHz. We should be asking why does the guy at 2.3GHz get to be so wasteful with his bandwidth when technology now makes it cheap and easy for him to get more done with less, and we're all crammed in here together at 2.4GHz?

Won't Work

[zentec]

A couple reasons come to mind as to why this Open Spectrum nonsense won't work and won't be applied.

First and most importantly, the federal government rakes in tons of money from spectrum auctions and licensing fees. However arcane, that simply won't be eliminated because something "better" has come along that's in the interest of the people. Reducing taxes are in the interest of the people, and we know how resistant the government is to that!

Despite what the author believes, spectrum allocations are a sane way to managing RF. Granted, spread spectrum doesn't interferre with other transmissions *when technically sound methods are used*. But when left to their general devices, the public sometimes eschews technically sound ideas and does stupid things. No matter how robust spread spectrum claims to be, when the front-end of the receiver is overloaded because of a dirty transmitter down the block, things quit working.

I've never had to DF (direction find) a spread spectrum transmitter, but I suspect that it's a far cry more difficult than finding a spur created from a faulty paging server.

What about radio astronomy?

[davecl]

With greater use of the spectrum, and the potential for software defined radios to use any frequency they want to transmit on, we're going to close out the possibility of ground based radio astronomy. This is not a good thing!

Radio astronomy produced many of the basic technologies that todays wireless communications revolution depends on, but is seeing none of the (financial) benefits and is gradually getting squeezed out of its own very limited parts of reserved spectrum. Maybe there should be a 1% levy on all radio licenses dedicated to help astronomers get around this problem and properly police their parts of the spectrum. Or maybe all the money raised from spectrum auctions should be dedicated to establishing space-based astronomy in the radio - probably on the backside of the moon to get away from all the noise!

There are needs for regulation to protect the other users of spectrum that wireless networkers forget about. Total spectrum freedom is not possible or a reasonable goal.

For more information see:

AAS webpages.

Re:What about radio astronomy?

[Celandine]

If you read the document carefully, you'll see that he doesn't advocate complete deregulation, and that radio astronomy is explicitly mentioned as a service that needs protecting.

As extremely broad-band radio astronomy receivers (think GHz rather than tens of MHz) are on the way (e.g. the plans for the EVLA) radio astronomers are going to have to abandon the idea of completely RFI-free bands (already a myth at 1.4 GHz and below) and concentrate on ways of automatically detecting and removing it instead. Of course, large numbers of small, frequency-agile transmitters are pretty much the worst nightmare in this sort of scenario...

With clever use of SDR, they might not notice

[Gerry Gleason]

Particularly with the 'underlay' concept, if you could quitely start using spectrum in ways that don't interfere with existing uses, who would know? In the regulatory space, spectrum license holders are always arguing against new uses because they *might* interfere with their bands even though there is precious little evidence that there would be problems.

Of course, this is the big conceptual problem with Open/Free Source SDR in general. Recieving is fine, but as soon as you want to transmit, the FCC wants to regulate the device. Now you are getting into the same problems you have with Open Source and DRM, if the end user has control, they can circumvent any controls by taking them out and recompliling. Buggy or malicious software could interfere in a big way, although the range of the interference would be limited by limiting the power of the transmitter which couldn't be easily overriden.

I like the paper's analogy to ships on the ocean, and I think it is very accurate for spectrum use. Most of the time you can put on the autopilot and sleep without much danger, but you'd better be on watch to adjust your course when you're near the shipping lanes. In any location, most of the spectrum will be silent most or even all of the time, and most modern recievers are pretty good at cancelling interference, so nobody will even notice if you use nearby frequencies.

The TV bands have all kinds of space in them, and there is all sorts of interference already unrelated to unlicensed transmitters. I'm still using my TV antenna in an urban area, and the big problem is multipath interference from all the large steel frame buildings. I wouldn't have any way to know that unused channels were being used for underlay digital comm., and I wouldn't care much either.