Throughput speeds are radically different in Chicago and Rural Town, North Dakota.

When it comes to availability of economical high bandwidth Internet access, there’s clearly a disconnect between those who live in cities and suburbs and those who reside in rural areas. If your ZIP code is 60602 (downtown Chicago), for example, you can choose between DSL services provided by the local phone company and broadband service from the local cable TV provider. Of course, that’s in addition to three or four wireless broadband services currently being offered. And if it’s not there already, most likely you will soon have the option of getting “fiber to the home” service for truly eye-popping data speeds.

On the other hand, if you live in ZIP code 58620 (Amidon, N.D.) chances are the only practical means of Internet access available to you, aside from a satellite link, is through a dial-up service delivering perhaps 30 kbps download speed.

This contrast of extremes illustrates the so-called “digital divide” separating urban and rural levels of Internet access. A number of sociologists and economists worry that as high speed Internet service becomes ever more critical to our daily lives and livelihoods the digital divide will become a real problem for those whose work or living preferences put them beyond the reach of broadband networks.

There is growing hope that wireless data networks will fill that breach, especially with the recent auction of 700 MHz spectrum. That’s the hope, but there are plenty of technical and economic hurdles that will have to be overcome.

THE HURDLES
As with any wireless data network intended to provide general coverage (as opposed to localized “hotspots”), the engineering of a rural system will require a delicate balance involving available RF spectrum, robustness of coverage, throughput bandwidth and capacity. In rural areas, however, there will likely be one or two additional factors that make things more challenging.

Consider, for example, the issue of backhaul. Rural base stations may be coverage- rather than capacity-limited, but they still have to be provided with sufficient backhaul bandwidth to support peak channel throughput rates. If a site has three sectors, each with a peak throughput of 5 Mbps or so, the required backhaul will probably be about eight T-1s or their equivalent.

That’s a lot of backhaul capacity for a rural base station, particularly since it will probably have to be carried quite a distance to the nearest node where that much traffic can be connected to the Internet backbone.

It’s always nice to be able to deploy data network base stations where high bandwidth fiber is available for backhaul, and that’s often practical in urban areas. But siting of rural base stations is pretty much dictated by the morphology of the areas to be covered, leaving little room to fudge things in order to reduce backhaul costs.

THROUGHPUT
Another huge challenge that will have to be dealt with in engineering rural wireless data networks will be balancing coverage, capacity and minimum acceptable individual throughput rates. These factors will affect performance in both the uplink and downlink, but it’s generally recognized that customer perception of data service quality is mainly determined by downlink speed, so let’s concentrate on that.

Since the service we are contemplating is intended to bring the benefits of “broadband” to rural users, it seems reasonable to assume that we would want to reliably provide users with a minimum downlink speed of about 500 kbps.

For the individual user, observed downlink throughput will depend primarily on RF channel quality and the number of other users simultaneously demanding service on the same channel. If we assume that the channel in question is 5 MHz wide, 500 kbps will require a very modest throughput efficiency of 0.1 bps/Hz.

In a noise-limited situation, for a practical user terminal receiver, and allowing adequate margin for fading and other path impairments, that will probably require a receive signal strength in the neighborhood of -95 dBm. That sounds very doable for a typical rural environment, particularly given the lower path attenuation found on the 700 MHz band.

Unfortunately, that’s not the whole story, because whatever throughput a channel can deliver has to be shared with all simultaneous users. What’s more, total channel throughput isn’t a fixed number but rather will depend upon the channel quality provided to each contemporaneous user. That is, a user downloading at 500 kbps under poor channel conditions may consume as much channel capacity as a user downloading at five times that speed but with excellent channel quality. And, to top it off, geometry is an enemy when it comes to average channel quality. If user distribution is fairly homogeneous, as is often the case in rural areas, there will be a lot more users relatively far away from the serving base station, suggesting that it’s likely more will have poor channel quality than excellent quality.

So, what it all boils down to is this: Given what a carrier can practically charge for rural broadband Internet access (assuming no subsidy), how many users will need to be served by each base station in order to make the economics work out? And, given that number of users, can the base station provide sufficient capacity to deliver acceptable speeds with reasonable reliability?

One thing that would certainly help make rural wireless broadband more practical would be to equip users with enhanced transceiver setups. Compared to a device with an integral antenna that sits on a desk indoors or plugs directly into a computer, a rooftop-mounted directional antenna could easily improve average channel quality by 10 dB or more. That would yield huge increases in individual throughput rates and total channel capacity. Of course, the downside would be loss of mobility, but I suspect the good folks in Amidon would accept this limitation in order to close the digital divide.

Drucker is president of Drucker Associates. He may be contacted at edrucker@drucker-associates.com.