Technology debates are almost universally viewed as debates about products or standards. So does that mean wireless broadband is a battle between 3G and WiMAX, or maybe between CDMA and GSM? Actually, the future of wireless broadband is probably best expressed as the battle of the densities. And the most important tool in the wireless strategist's toolbox may be an old-fashioned compass to draw circles.
Demand density tells us a lot about the overall economic quality of a market area. In fact differences in demand density explain why some countries (Japan and Korea, for example) can offer customers 50 Mbps DSL or FTTH, and why in the U.S., Verizon has one strategy (FTTH/FiOS) and AT&T has another (DSL/IPTV).
Demand density skips specific service requirements
One thing the raw number doesn't take into account is that every possible service doesn't draw on the total potential ARPU. For example, if you decided to offer a customer something like the old ISDN service with 128 kbps of full-duplex data capacity, it's pretty clear that high-definition (HD) video is not on the table as a revenue source. Most planners don't fall into that trap. But a surprising number fail to consider service-specific revenue opportunities when it comes to wireless, and that can lead to a major problem down the road. The reason is a second kind of density -- bit density -- and how it relates to demand density.
Any wireless broadband service tends to have a simple hub/circle topology. The wireless cell is fed from a central antenna point by some wireline technology. From that point, radio frequency (RF) access is projected in a roughly circular pattern for as many feet or kilometers as the particular technology allows. The circle you get by setting your compass to the service radius of the technology (we could use 10, 20 or 40 miles for WiMAX, for example) is the area we want to use to compare densities. Here's a simple example. If we assume a provider is going to stick a wireless antenna on a pole and use WiFi to reach a community of homes, we could set our compass at 100 feet (being generous) and describe a circle around the pole. Using high-school math, the area of that circle is about 31,000 square feet, which is about a thousandth of a square mile. If we draw that circle in an urban area, it would intersect about 32 households. In the suburbs, it would cover between one and five households, and in rural areas, there's about a 10% chance that it would hit any households at all. For any given assumption on ARPU at a household level, it's pretty clear that WiFi in rural areas isn't a natural winner.
Is WiFi a wireless winner?
But is WiFi a winner anywhere? That's where the second kind of density—bit density—comes in. WiFi is about a 50 Mbps technology, and like all wireless services, the capacity of the cell is shared by all the members inside it. Thus we're sharing our 50 Mbps with 32 families in that urban setting, giving us what would be the equivalent of mid-range DSL service (1.5 Mbps per household), and forgetting for the moment how we'd meter the families to that level. HD video wouldn't work any better at this rate than it would on ISDN. So that means that while WiFi might have enough coverage to support a reasonable number of urban households, it couldn't give them the service they'd need to support IPTV, unless they somehow managed to stagger their scheduled viewing.
So what about WiMAX density?
Let's use the same technique to respond to some published comments on WiMAX. An example is that 26 WiMAX cells could cover the whole city of San Francisco, or about 1.5 million households. In geographic coverage terms, that's probably accurate. But the capacity of 26 WiMAX cells is 1,300 Mbps. If we divide that by our 1.5 million households, we get even less per-household capacity than we had with our WiFi example, less, in fact, than a 1200 bps dedicated modem connection to each would provide.
It is true that all data applications are bursty, meaning that they don't consume 100% of the bandwidth 100% of the time. But it's also true that the premium applications that service providers are staking their infrastructure investments on these days aren't nearly as bursty as email or web surfing. In fact, consumers who regularly consume free online video from YouTube or other sites also produce pretty non-bursty traffic. This type of traffic, when downloaded over wireless bandwidth shared among tens or hundreds of households, creates more bit-density requirements than the wireless technology can support.
Working around wireless capacity issues
What about making the cells smaller? AT&T's U-Verse assumes you need about 25 Mbps DSL to get IPTV to work properly. You can give that to two users using 50 Mbps WiFi or WiMAX. 3G technology, which limits cell capacity to about 2 Mbps, can't support that rate at all. Two households per cell means that if everyone is a prospect for broadband video, WiFi and WiMAX need half the number of wireline feeders to cells as you'd need wireline broadband connections to the home.
This isn't to say that wireless broadband isn't a good thing for some applications, because clearly it is. What it does say is that it's not equivalent to wireline broadband, and that every wireless broadband application admitted onto the network draws bits from a pool that everyone in that magic access circle shares. Planners need to keep that in mind when they consider wireless roll-outs.
About the Author: Tom Nolle is president of CIMI Corporation, a strategic consulting firm specializing in telecommunications and data communications since 1982. He is a member of the IEEE, ACM, Telemanagement Forum, and the IPsphere Forum, and the publisher of Netwatcher, a journal in advanced telecommunications strategy issues. Tom is actively involved in LAN, MAN and WAN issues for both enterprises and service providers and also provides technical consultation to equipment vendors on standards, markets and emerging technologies. Check out his SearchTelecom networking blog Uncommon Wisdom.