Rethink - An Open Spectrum License

Supercommons explains in much greater detail how the spectrum fallacy came about. Suffice it to say, people thought it impossible for for two operators to transmit on the same frequency without interfering with each other. As Werbach asserts: "This assumption is false".

If the property/lease model is false, then we must search for a new, more correct analogy. Happily, one exists in very air that surround us. Spectrum is like air, a medium. Consider a concert hall, and imagine the sounds are like radio transmissions.

In a hall, it is possible for a single source, say, the orchestra, to dominate the airwaves with their music. However through further inspection we may see that broadcasting music in a hall is not exclusive (nor is, by analogy and demonstrably through science, broadcasting in spectrum).

Before the concert ever begins the hall is filled with spectators, all carrying conversations in this shared medium. They speak simultaneously, creating interference waves. Despite the interference, the conversations are able to carry on unimpeded. Indeed, each speaker/transmitter modulates the power of their voice to be just loud enough that their neighbour can hear. In addition each listener's audio nerves are capable of discriminating a friend through hundreds of different voices.

During the concert, it remains possible for a neighbour to whisper quietly in the ear of another without disrupting the reception of the other listeners. This is accomplished by a pair of techniques. First, the transmitter whispers at a very low power level. Second, the transmission is highly directed, straight from the mouth to the ear. Anyone not standing directly between transmitter and receiver would hear nothing.

Anyone who has ever experienced a "whispering gallery" has observed further odd effects as sound travels through the air. These effects are also present through parabolic antennae, in the spectrum medium.

Although ancient radios were unable to discriminate in this way, modern wireless systems can. Digital cellular networks, for example, divide a single frequency channel up using a timeslice technique called TDMA (time-division multiple access). But TDMA is long since eclipsed by more advanced data techniques.

CDMA (code-division multiplexing) is also widely deployed in cellular networks. It applies a different code to each signal at transmission, and then a code-reversal at the receiver pulls the signal back out of the air. The code is analogous to the sound of a specific person's voice, which our brain is able to pull out of an air filled with many voices.

Newer digital technologies go further in mimicking the human voice. Mesh networking is a technique that allows radios that are close to each other to communicate at low power. Messages propagate across the network like frogs across lilypads. Each node communicates only with its closest neighbours. Such systems are being deployed in "MANET" -- mobile, ad-hoc networks.

The traditional spectrum "license" is actually a lease. By comparison, an open source "license" is a contract between a provider and any user. People who use open source software must first agree to the terms of the license in question, which may be the GPL (GNU general public license), the BSD license, the Mozilla license, etc.

The various contracts serve to create freedoms not so much for the author or end users, but rather for the software itself. Anyone who agrees to the license may freely use the software ... and they must ensure that the software remains free by following certain basic provisions.

The specific details of each license are important, but they are based on something even more important, the philosophical guidelines of the open source movement. These come in several forms - the Free Software Guidelines [2], the Debian Free Software Guidelines [3], and the Open Source Definition [4].

These definitions preceded the actual licenses in their creation. Therefore it makes sense to pause and try to set out what would define an open spectrum license.

1. Free use of the medium - the license shall not restrict anyone from using the spectrum, nor shall any fee, royalty or payment be required

2. Accept interference from others - the license shall require that all receivers accept interference from other sources

3. Minimize interference to existing users - the license shall require that all transmitters make reasonable attempts to minimize interference with other existing users

4. No Discrimination against fields of endeavour - the license must not restrict anyone from using the spectrum for any field of endeavour

A practical open spectrum license would embody the open spectrum definition into a practical license. The exact terms of the license must, of course, depend on the specific technology and frequency that it covers.

For example, a open spectrum license being offered for spread-spectrum radio technology in the 2.4 GHz band would be equivalent to part of the current ISM band "regime" that is in place in many parts of the world. This example license would place specific requirements on the broadcasting power allowed. If we are to mirror the current USA policy, the limit would 1 watt EIRP (Effective Isotropic Radiated Power) under normal circumstances. This is acceptable because part 3 of the proposed Open Spectrum Definition specifically requires that the license minimize interference to other users, and limiting power to 1W is agreed to be a good way to do that.

An open spectrum license for UWB (ultra-wide band) radios would be a bit more complex. The frequency range would be much broader, perhaps as broad as 150 MHz to 20 GHz (or higher). The section 3 interference definition would be basically, to operate below the noise floor of existing receivers. A broadly defined open spectrum license could be released along these vague terms and then different frequencies with more specific details about power limits could be released to provide a layered approach to introducing UWB.

One might cynically wonder why bother to create a new open spectrum licensing paradigm when there is already an existing spectrum management system. In this, the final part of this paper, I will attempt to justify the need for a change in open spectrum thinking.

From a regulatory point of view, it is well-accepted that open spectrum policy (in its current form) is a bestiary of complexity. Any method of simplification would bring a sense of greater control to the situation, help in resolving differences between open spectrum and amateur radio enthusiasts, and help open spectrum proponents to better make their case.

The situation is particularly critical in developing nations. There is often a lack of resources towards spectrum planning, such that sophisticated "unlicensing" schemes are not pragmatic. Open spectrum licensing, by contrast, is simple to both describe and put into policy.

The open spectrum license also addresses a particularly pernicious problem with those who stand outside the walls of traditional spectrum management agencies and industry. There are people who equate "unlicensing" with "deregulation" (which might be accurate in other industries). Open spectrum licensing clears up any potential misunderstanding that could occur in this area.

Although the open source movement can at times be controversial, it also embodies an almost unstoppable momentum. Open Spectrum shares many of the same ideals as open source. It seems reasonable to assume that an increase in open spectrum (presumably a goal of many readers) can benefit from the shared philosophy.