Ok, I see the problems with that PSI post... my fingers did not capture my thought. Perhaps I can explain better.
Let me start with the premise that unlicensed spectrum is exactly analogous to the classical macroeconomic "problem of the commons." If nobody has to pay for it, they are going to use up as much as they can, as fast as they can, before someone else does.
The problem here is similar to the shared nature of an HFC architecture. Subscriber "gluttons" degrade the service to the detriment to the average user. If the operator does not place limitations on usage, then it is forced into expensive upgrades to maintain a minimum quality of service.
However, at least in the HFC situation, the operator controls its destiny. It can (1) limit service or (2) charge more to everyone or (3) charge more to heavy users.
Yet the situation for the commercial operator using unlicensed spectrum is worse. Legitimate users of the spectrum - and one must expect there will be a multitude of them - who are NOT his customers, degrade the service to his customers. The operator has no control over other legitimate users of the band. Lets say he starts out offering dedicated service at T1 speeds. As the bands loads up, that wireless link degrades due to increased interference. Throughput drops and eventually the link fails. The customer has done no wrong; the operator has done no wrong, it just doesn't work anymore. Who pays? Will the customer accept higher charges for the exact same throughout and quality of service he has grown to expect? Or, in the alternative, can the operator afford to re-engineer the link (install new base stations) and then roll a truck and re-align a multitude of customer antennas to the new, geographically closer, base station which provides the original link margins?
At least, in the licensed spectrum scenario, the additional interference is being generated by his paying customers. Consequently, the capital equipment expenditures required for cell splitting are being paid for by the enlarged subscriber base.
Significantly, fixed wireless is different from mobile wireless in that the fixed wireless operator has to look far into the future on initial deployment and carefully plan for anticipated cell splits. The problem here is a directionalized antenna at the subscriber location, which - again - mandates a service call and downtime, if a cell split is required. (For example, at 28 Ghz the beamwidth of the antenna is about 2 degrees. This requires precise alignment and a skilled technician. Reviewing the Clearwire spec sheet, while the subscriber antennas are also directional, they certainly do not have such a narrow beamwidth hence the alignment problems are mitigated, but in no way eliminated if you have to install a new base station. Please, give me a low-cost, remotely-steerable, phased-array.(g)) Also, the effects of cell splits are not usually limited to the overloaded base stations and its first adjacents. In terms of frequency planning, the effects will ripple throughout the system. (In mobile, direct sequence and frequency hopping mitigate some of the ripple effects, but not all. IMO, they provide far less benefit in a point-to-point deployment.)
Again to contrast with the unlicensed spectrum, the operator has no control over other legitimate users of the band. The operator does not even know where they are. All he sees is that his customer's link margins are degrading. This can happen gradually, or instantaneously, e.g., someone turns on a transmitter in the building next door, or maybe in the next adjacent office. The customer doesn't understand or care about these problems, he just expects it to work.
Having said all that, my point in suggesting that the commercial operator in unlicensed spectrum has to ultimately migrate is that he is getting something for nothing, i.e., use of the spectrum. Eventually, the "free" spectrum will cost him more in capital equipment cost and operating expenses (those truck rolls) than the money saved on spectrum. The trick is when. Wait too long, and maybe you lose the customer(s), or maybe there is no spectrum available, or maybe the price is astronomical. (Look at the WNP/Nextlink deal; spectrum purchased for $187 million flipped for $695 million in just 11 months.)
ww |
