Cheap, reliable, high-speed broadband access anytime anywhere. How often have you heard that? This wireless access will be provided by GPRS, EDGE, CDMA 2000 1x, UTMS/WCDMA, CDMA 1xEV-DO, WiMAX, 4G, you name it. It’s already here, or just around the corner, and it will only get better and better and better …. or will it?
The promise of mobile broadband wireless access is exciting and large amounts of money, time and effort have been spent to make it happen. The results are quite impressive. Using a laptop with, say a 1xEV-DO card, you can get internet access in many locations. But mobile broadband wireless has a dark secret which is seldom discussed in any detail. Namely – it does not scale well. The technology works well when the number of broadband users is small, but as that number grows, performance will rapidly deteriorate. The reasons for this are both technical and economic. The following is a somewhat simplified explanation of the problem.
One of the important characteristics of any wireless system is its spectral efficiency – the ratio of the peak data rate it provides to the bandwidth it consumes. A system delivering 10MBPS in a 5MHz slice of spectrum has a spectral efficiency of 2. Some examples: EDGE has a spectral efficiency of 1.92, WiMax 3.75, 802.11a/g 2.7. These numbers correspond to peak data rates. In reality the average data rate will be significantly lower, so the effective spectral efficiency is significantly lower (this deserves a more detailed explanation in a separate post).
The spectral efficiency of a system is limited by the signal to noise ratio (SNR). The higher the spectral efficiency you want, the higher the required SNR. In cellular systems the SNR is quite low because of interference between cell sites and between different sectors in the same site. The average SNR in such systems is on the order of 2 – 4 dB, which limits the spectral efficiency to 3 – 4. Again, the effective spectral efficiency can be much lower. The spectral efficiency of wireless systems has been gradually improving from 2G to 3G and beyond. Further improvements are possible but they will be relatively small because of this SNR constraint. What this means is that the total data rate available per cell site for a given spectrum will keep increasing as technology improves, but future increases will be modest at best.
This is what happens on the supply side – the supply is increasing slowly and is rapidly approaching a plateau. On the demand side the situation is quite different. While a voice channel requires 13 – 15 KBPS, a data channel requires an order of magnitude larger data rate. Users who want to browse the net need, say 200 – 300 KBPS to get the kind of browsing experience they are used to having at home or at work. Currently the number of data users is relatively low (at least in the US), so the supply meets the demand. However, as the number of data users will increase, as will their desire for high data rates, the demand will increase and outstrip the supply. This is likely to happen sooner than later. The recently introduced iPhone is a case in point. The iPhone is probably the first truly mobile device which allows reasonably convenient web browsing, and thus encourages the user to demand relatively high data rates. Not surprisingly users have already been reporting problems with frustratingly slow net access. As more such devices are introduced the demand will quickly exceed the supply.
What can be done to increase the supply? We either need more spectrum, or we need to increase the density of basestations by an order of magnitude, thus reducing the number of users per basestation. The first is virtually impossible because of the scarcity of “good” spectrum. The latter is prohibitively expensive and is not economically viable because it will require large increases in revenues from data users. System costs are roughly proportional to the number of basestations which is increasing, while revenues are proportional to the number of users which is not increasing. Users are not likely to be willing to pay much more than they are paying now.
So cellular companies are caught “between a rock and a hard place”. Mobile wireless access is available, and it is a technological marvel, but if people will really start using it seriously, the current infrastructure will not provide the increasing demand, and upgrading the infrastructure is too expensive. Newer technologies (WiMax, 4G) can not solve the problem, only provide some temporary relief.
Is there a solution to this problem? In my opinion any such solution will have to rely heavily on WiFi – and I will elaborate on that in a future post. Of course there are always the two traditional solutions:
(1) Legislation: we should push through as soon as possible a bill to repeal Shannon’s law. This outmoded law has been a major stumbling block for developing high speed wireless access in the United States, and has long outlived its usefulness. Anyone interested in joining a political action committee to push this bill, please let me know.
(2) Innovation: if we only put our minds to it, let our creative juices flow, do some serious brainstorming, and think out of the box, our good old American know how will solve this problem. In fact I can reveal that I have recently developed a revolutionary invention which can instantly quadruple the capacity of any wireless system. I can not provide too many details for obvious IP reasons, but I can say the following. The new system uses square waves instead of the conventional sinusoidal waves. The key to this technology is a special filter which passes square waves but eliminates almost completely sinusoidal waves. This filter allows us to re-use the spectrum occupied by old-fashioned sinusoidal communication systems. These square signals have other advantages: they enable communication using lower power than traditional systems, have much better building penetration capability, and are easier to implement because they can be generated digitally. This technology, which we named SMax is still under development, but preliminary tests have been very promising. Please contact me if you want to learn more.