An article that I have had on my list to write, but never quite got around to it is
What Spectrum Shortage?!?!?! For a long time now, strongly influenced by the ideas of Dewayne Hendricks and others, and later bolstered by the Open Spectrum movement (formalized as the Open Spectrum Project -
http://www.law.nyu.edu/ili/spectrumpolicy.html) I believe that the monolithic, static policy we have in the US for spectrum allocation is very, very broken. That it's broken certainly isn't news, but the proposed solutions by the policy wonks trend, all too depressingly to various regulatory, legal, or financial / economic remedies. I think what is happening is that the true nature of spectrum is just so out of the realm of the wonk's collective experience that they simply are unable to grasp a basic, fundamental truth about spectrum:
Usable spectrum is entirely a creation of advanced technology; the more technology you apply, the more usable spectrum is available for use. Making more spectrum available is a
technological question,
not a regulatory, legal, or financial question.
Here's an example that's been on my mind lately (I'll try to make this brief). Several years ago I learned of a company called Triton Network Systems. Triton made 38 GHz radios, called Invisible Fiber, that operated at 100 Mbps transmit and receive simultaneously. Invisible Fiber radios were used two to a building, and in a metro area, they were connected in a ring; something like a wireless version of a SONET network with dual, counter-rotating rings of data. If any one radio were to fail, the data can still find its way back through the reverse direction. The key characteristic of the Invisible Fiber radios was that they had astonishingly good dynamic transmit power control (literally a range of picowatts to watts). They transmitted
just enough power to reach their partner radio, whether that radio was 100 feet away on the building next door, or 1 mile away. Quite simply, unlike all other 38 GHz radios (that I'm aware of), the Invisible Fiber radios
didn't "pollute" the RF path! That made it possible to reuse a single 38 GHz channel over, and over, and over in a metropolitan area. In contrast, the power level on conventional 38 GHz radios is set staticly, to the highest power level expected to be required in the worst weather conditions. The effect of that static power setting (or comparitively crude dynamic power control) is to "pollute" that channel, on that path, so that no other radios can use it.That's been the accepted wisdom and why you need lots of channels. Using the Invisible Fiber radios, in effect, new spectrum was "created" - spectrum that was previously unusable was now usable because the radios were smart enough not to "pollute".
Because the older and current radios generally aren't sophisticated enough to tolerate interference (or prevent the pollution from occurring in the first place), spectrum had to be "coordinated" so that interference didn't occur. The FCC doled out "licenses" to organizations who were then the only entity allowed to transmit on a particular band or channel in a particular region (and went so far as to specify power levels and other technical requirements). Initially the FCC doled out spectrum as necessary / on demand / on merit. At some point, there wasn't enough spectrum for all who asked, and so the FCC instituted lotteries. All the entities who were awarded a license in a spectrum lottery thought that the lottery system worked well; the losers were substantially less enthusiastic about how well spectrum lotteries worked.
Then came spectrum auctions, touted as the highest and best form of spectrum allocation - capitalism at its shining best. If you bid a lot for your spectrum and won, you were incentivized to begin using your new spectrum quickly and productively to recoup your investment. But that's not quite the way it worked out. In the majority of cases of spectrum auctions, the bids for the most popular spectrum reached levels that literally had no hope of the bid amount ever being recouped. The other extreme was spectrum that was relatively unpopular, snapped up for paltry bids. In some cases, spectrum was bid for, won, paid for, and deliberately not used as a strategic move against competitors (making sure that the competitors can't use the spectrum). In other cases, spectrum was bid for, won, paid for, and and not used because it was judged to be an "investment"; someday, someone would value such "warehoused" spectrum quite highly, and it would only then be sold at a hansome premium above the original investment.
With the above as context, you might be able to appreciate my growing sense of dread at a proposal by two FCC staffers, and glowingly endorsed by two academics formerly affiliated with the FCC, dubbed "The Big Bang Spectrum Auction". One story on this is at http://seattletimes.nwsource.com/html/businesstechnology/134476462_spectrum17.html. Generally, the idea is that all available spectrum should be offered for auction simultaneously, including spectrum that is currently assigned to commercial entities. Somehow, the spectrum bidders will "sort things out" and "get the spectrum that they need". Mentioned promiently in this scenario is that the US Government would be able to "buy back" spectrum that's already been sold to commercial entities and then offer it for license-exempt use similar to the current 2.4 GHz, 5 GHz, and 60 GHz bands.
To me, such a mega-auction for spectrum will merely exacerbate a bad situation to "completely dysfunctional". There will be a "land grab" for spectrum by speculators intent on "investing" and holding spectrum in reserve until "the price is right". If that were to happen, I feel there would simply be no tolerance for manual spectrum allocations and spectrum users simply begin to fend for themselves, ignoring the (at that point) completely dysfunctional FCC and its guidelines and spectrum allocations.
I feel strongly that technology is the solution to the spectrum allocation mess. Any new spectrum that is made available for use (like the newly-proposed "W Band") should be made license-exempt with minimum spectrum sharing technology mandated, and no promise of interference protection... much like Part 15 or UNII rules already in existence. With the example of the Invisible Fiber radios, the 38 GHz band could gradually be deregulated into license-exempt use by mandating the phase-in of dynamic power control capability similar to what Triton was able to implement; thus permitting 38 GHz to be much more widely used than it is currently.
There are a mind-numbing number of ways to technologically "create more spectrum". Better ("smart" / beam forming) antenna technology; spectrum overlay, spread spectrum, etc. ad nauseum, all of which leverage Moore's Law because RF technology is increasingly based on computation, where Moore's Law reigns supreme. Faster DSPs are emerging all the time. Computationally intensive modulation techniques such as OFDM are only possible with cheap, fast computional horsepower. Beam forming antennas have to perform millions if not billions of calcualtions per second to know what combination of phase, power, etc. to send to each element of a beam forming antenna.
I'll close with the observation that radios are now capable of making spectrum allocations decisions in microseconds. The equivalent capability, managed by humans, takes years. Yes, literally, Years.
My standard disclaimer: I consider this a significant development in the Broadband Wireless Internet Access industry. I'm way overdue for writing What Spectrum Shortage?!?!?!, as well as going into more background as to how I came to my conclusions, which will appear exclusively in my newsletter FOCUS On Broadband Wireless Internet Access.
Steve Stroh
Copyright (c) 2004 by Steve Stroh. This article originally appeared on my original Broadband Wireless Internet Access Weblog hosted on Radio Userland.