Open Spectrum: Free the Airwaves

akb writes: "Most of the RF spectrum in use is licensed for exclusive use. What do we get? Inefficient use through spectrum hoarding, political finagling to abuse the regulatory system to gain competitive advantage and access to the airwaves for only a few players. A good article over at CNET picks up on the example of 802.11b in using spread spectrum technology and unlicensed bands and proposes that model be applied to the rest of the spectrum. For the hardcore check out NYU law professor Yochai Benkler's writings, particularly this article (pdf) and Durga Satapathy's papers for the tech end of things."

Comments

Re:Didn't you read the article?

[libre lover]

Apparently spread spectrum technologies deal with this problem.

Not quite. Spread Spectrum is like pseudo-random radio "noise", constrained within a defined spectrum. Such transmitters have the effect of raising the "noise floor" in the spectrum they're using (and thus the error rate of the channel). This places an upper limit on the number of spread spectrum users a given channel can support. It's also a problem in spectrum (i.e. Amateur Radio and Astronomy) where weak signals are important.

One regulatory change that can be made is to allow use of spread spectrum on airwaves currently allocated for narrowband use where the narrowband noise margins are typically fairly good. In theory at least the two types of transmissions would not interfere with each other.

Oversimplistic

[mesocyclone]

Current radio regulation is far from efficient, but removing the regulation entirely is foolish, and ignoring frequency sharing won't work. There are engineering realities that the writers and lawyers don't understand that limit this.

While bandwidth is an oversimplistic way of either looking at things, or regulating them, it is a fact that any communications system can be analyzed (roughly) in terms of bandwidth. And this means that any communications system can interfere with any other communications system if they share frequencies in any sense.

For a real world example of why you can't ignore bandwidth, try running WiFi in a house where you have some 2.4GHz phones. It may work. But sometimes it doesn't - the reason - radio frequency interference. They share the same bandwidth.

Ah, you say... so they don't use good enough systems... or aren't broad-band enough... or something! Not true... there are hard physical limits that no amount of scheming will get around. The rest of this post discusses that in more technical detail.

All signaling systems (INCLUDING Time-Modulated Ultra-Wide-Band)require a separation of signal from noise. Noise is either natural (thermal, atmospheric, solar, etc), incidental (power line leakage, etc) or other radio systems. Regardless of what kind of signaling system is used, it has a limit as to the amount and kind of noise that can be tolerated in any given situation. The other limits described below affect the amount of noise reduction/signal enhancement that is possible.

Limits to processing gain. WiFi and other modern technologies (CDMA cell phones) use spread spectrum to reduce the effects of interference. Unfortunately, this does not eliminate interference. In engineering terms, it is the equivalent of adding gain to the desired signal. The gain is roughly the bandwidth occupied by the transmitted signal divided by the bandwidth required to send the signal without modulation (the baseband bandwidth). This value is measured in decibels, and is typically 20-30 dB, although it can increase. But the higherhe data rate, the lower the processing gain!

The effect of distance - radio signal energy decreases by an inverse square law. This means that a nearby interference source can have a much stronger signal, proportionally, than the desired signal from a farther source. Some numerical examples:

1. A receiver at room temperature will have an inherent noise level of -174dBm (10E-20.4 Watts). This means that if you want to send a 1HZ signal, you must generate more than -174dBm in the receiver. This sounds like a tiny number, BUT...
   
2. A hand-held cell phone operates up to about 600mW which is +27dBm.
3. Now, let's transmit that signal a few miles. The antenna has roughly no gain on the handset. The receiver antenna might have a capture are of 1/4 meter. At 3 miles, the 600mW is distributed across the surface of a 3 mi radius sphere, giving a signal strength of 6.5*E-10 Watts at the receiver (-62dBm).
4. The baseband signal of this cell phone is about 2KHZ. This means that the -174dBm requirement is upped by a factor of 2000 to -140dBm. But we also need a signal to noise ratio of, say, 15dB to receive that signal well, so now we are at -135dBm. So - we have a roughly 43 dB margin.
   
5. Now add 40 dB of path loss from buildings in the way and you have a 3dB margin... your signal barely makes it adequately.
   
6. Let us fire up another cell phone on the same frequency band (I am assuming we are using spread spectrum). Let us assume a reasonable spreading gain of 1000 (30dB). Put that cell phone 100 yards away, and guess what: It gives you an effective signal of .6/1000/125000/4 = 1.2x10E6 milliwatts or -59dBm. Our desired signal is at -62dBm, so it is wiped out!
   

This illustrates that a signalling system, by itself, will not prevent interference - defeating the main argument. Specific factors are:

Imperfections in equipment. Real equipment will not reach theoretical levels of performace.

Limited dynamic range. If you have a 100,000 watt transmitter 3 feet from your receiver, there is a good chance that no matter what its technology, it will not be able to pull out the desired signal. In digital terms, this is the equivalent of running out of bits in your integer! If a number is too big, you either overflow your math, or you scale it down, losing the little bitty number you wanted.

Limited bandwidth - there is a limited amount of bandwidth, useful for a given purpose, at any place and time. This bandwidth, for many purposes, is between 1GHz and 25GHz (although for ionospheric radio, it is only 30 MHz). This means that if someone is generating a strong signal in the bandwidth you are using, there may be no other bandwidth you can jump to.

Intermodulation. Any nonlinearity in the system, including incidental nonlinearities such as a nearby rusty pipe, will cause all the RF signals impinging on them to be mixed, and the mixing products re-radiated. Receivers have inherent nonlinearity, which unfortunately gets worse as the power used by the receiver is reduced.

Leakage. You may have a great receiver, but an interfering transmitter that is close enough may leak through its plastic case and get into an intermediate stage of your receiver.

etc.

Without regulation, some other system must arise to arbitrate needs for radio spectrum, or chaos will result

Re:Just another thought

[Steveftoth]

You are confusing the HTTP protocol ( how a browser gets information ) with HTML ( how a browser displays information ) and in the case of both there are central controlling bodys that manage them. HTML is managed by the W3C and HTTP is an RFC. The diference between this and the airwaves, is that you can't get arrested for starting a non-compliant server / browser. Like with radio.

This is really comparing apples to oranges. As well. You can't mess up someone who isn't viewing your webpage with a messed up http implementation. While you can really mess up the radio spectrum if you send out bad signals.

Open spectrum is part of the FCC mandate...

[pongo000]

...interestingly enough. Over the years, however, various interests have hijacked much of the spectrum on an "exclusive use" basis. Maybe it's time the FCC returns to its original ideals (through the gentle prodding of the federal courts?)...

151. Purposes of chapter; Federal Communications Commission created

"For the purpose of regulating interstate and foreign commerce in communication by wire and radio so as to make available, so far as possible, to all the people of the United States a rapid, efficient, Nation-wide, and world-wide wire and radio communication service with adequate facilities at reasonable charges, for the purpose of the national defense, for the purpose of promoting safety of life and property through the use of wire and radio communications, and for the purpose of securing a more effective execution of this policy by centralizing authority heretofore granted by law to several agencies and by granting additional authority with respect to interstate and foreign commerce in wire and radio communication, there is created a commission to be known as the "Federal Communications Commission", which shall be constituted as hereinafter provided, and which shall execute and enforce the provisions of this chapter."

Present IEEE standards issues

[dtmos]

The author is out of touch with the issues now facing those of us now working on IEEE 802.11, .15, and .16 standards. The primary problem 802 has at the moment is that almost all of its draft wireless standards (e.g., 802.11g, 15.1, 15.3, 15.4, etc.) are being planned for the 2.4 GHz ISM band, due to its combination of near-worldwide unlicensed availability, suitable (i.e., relatively wide) bandwidth, and technical practicality (small antennas, possibility of cheap CMOS RF implementation, etc.). The major exceptions are in 802.16, the WirelessMAN(tm) Metropolitan Area Network standards, which typically employ such a high data rate that even the 2.4 GHz band is too narrow; however, even there, the 802.16b task group is developing a standard for the unlicensed 5-6 GHz band.

The difficulty is coexistence, or how all these standards will affect each other when networks using them are placed into service. This concern started as a Working Group issue, and was addressed by coexistence task groups (e.g., 802.15.2, 802.16.2a), but has now bubbled up to the 802 LMSC itself, with the recent formation of the 802 COEX coexistence study group. 802.11 has become the 800-lb. gorilla in the 2.4 GHz band, microwave ovens included, and it is far, far from the truth to say that just because every system involved is spread spectrum the band may automatically be shared among many users.

Spread spectrum offers protection only to the extent of its processing gain which, for direct sequence systems, is defined as the ratio of chip rate to data rate. Present FCC regulations for the 2.4 GHz band specify a minimum of 10 dB processing gain; this requires a chip rate that is 10x the data rate. As one can see, to get significant processing gain one either (a) raises the chip rate, and the associated current drain of the product, to a high value, or (b) reduces the data rate to a low value. Neither of these is attractive when one considers that even a ratio of 40 dB (10,000x) is insufficient in many interference scenarios; worse, the FCC is proposing to eliminate the 10 dB requirement completely so that OFDM (Orthogonal Frequency Division Multiplex) signals, like those proposed for 802.11g, may be used.

CFR 47 15.247 devices, like 802.11 and .15 devices, are sold under the condition that they must accept interference to them caused by other devices. This was essentially a regulatory passing of the buck to the "free market," which has a spotty record in telecom (cf. U.S.' multiple cell phone standards vs. GSM). Since 802.11b has the largest installed base, any standard that follows that produces interference with 11b devices will have a hard time gaining marketplace acceptance; at the same time, brute force technologies to avoid interference, such as the use of processing gain, are insufficient. This leads standard and product designers to design ad hoc coexistence mechanisms to identify and avoid specific, predetermined interferers, an inefficient, piecemeal approach that places later, next-generation devices at a disadvantage over existing ones. The result is that 802.11 derivatives are going to defacto own the 2.4 GHz band in most corporate and (later) home environments; anything new in the band must carry the coexistence burden with it.

So, if 802.11b is the model for "Free airwaves," it's a poor model; it's more MS open spectrum than linux open spectrum.