FCC Chairman Wheeler made a speech last week about the lack of broadband competition in the country. As part of the speech he released four bar charts showing the percentage of households that have competitive alternatives at the download speeds of 4 Mbps, 10 Mbps, 25 Mbps and 50 Mbps. His conclusion was that a large portions of the households in the US can only buy broadband from one or two service providers. I was glad to hear him talking about this.
But unfortunately there is a lot of inaccuracy in the underlying data that he used to come to this conclusion, particularly at the charts showing the slower speeds. The data that the FCC relies on for measuring broadband is known as the National Broadband Map. While the data gathered for that effort results in a Map, it’s really a database, by census block, that shows the number of providers and the fastest data speed they offer in a given area.
A census block is the smallest area of population summarized by the US Census. It is generally bounded by streets and roads and will contain from 200 – 700 homes (with the more populated blocks generally just in urban areas with high-rise housing). A typical rural census block is going to have 200 – 400 homes. The National Broadband Map gathers data from carriers that describe the broadband services they offer in each census block. As it turns out, self-reporting by carriers is a big problem when it comes to the accuracy of the Map. In tomorrow’s blog I will show a real life example of how this affects new deployment of rural broadband.
Broadband service providers don’t generally track their network by census blocks, so part of the problem is that census block don’t match the physical way that broadband networks are deployed in a rural area. Anybody who lives in rural America understands how utilities work there. In every small town there is a very definite line where utilities like City water and cable TV stop. Those utilities get to the edge of the area where people live and they stop. That doesn’t match up well with Census blocks that tend to extend outward from many small towns to include rural areas. Rural census blocks are not going to conveniently stop where the utilities stop.
There are three widely used rural broadband technologies – cable modem, DSL and fixed wireless. Let’s look briefly at how each of these match with the broadband mapping effort. Cable is the easiest because every cable network has a discrete boundary. There is some customer at the end of every cable route and the next house down the road cannot get cable. So it is not too likely that the cable companies are claiming to serve census blocks where they have no customers.
DSL and fixed wireless are a lot trickier. Both of these technologies share the characteristic that the bandwidth available with the technology drops quickly with distance. For example, DSL can transmit over a few miles of copper from the last DSLAM in the network. The household right next to that DSLAM can get the full speed offered by the specific brand of DSL while the last house at the end of the DSL signal gets only a small fraction of the speed, often with speeds that are not really any better than dial-up.
The same thing happens with fixed wireless. A WISP will install a transmitter on a tower or tall structure and the customers close to that tower will get decent broadband, and those transmitters tend to be installed in small towns where people live. But wireless broadband speeds drop rapidly with distance from the transmitter and if you go more than a few miles from any tower there is barely any bandwidth.
Both telcos and WISPs input their coverage areas into the National Broadband Map database. And in doing so, it appears that they claim broadband anywhere where they can provide service of any kind. But for DSL and fixed wireless, that service-of-any-kind area is much larger than the area where they can deliver actual broadband. Remember that broadband is currently defined as the ability to deliver 4 Mbps download. Because of the nature of their technologies, a lot of the people who can buy something from them will get a product that is slower than 4 Mbps, and at the outer ends of their network speeds are far slower than that.
I don’t necessarily want to say that the carriers inputting into the system are lying, because in a lot of cases customers can call and order broadband and a technician will show up and install a DSL modem or a wireless antenna. But if that customer is too far away from the network hub, then the product that gets delivered to them is not broadband. It is something slower than the FCC definition of broadband, but it is probably better than dial-up. But customers with slow connections can’t use the Internet to watch Netflix or do a lot of the basic things that require actual broadband. And as each year goes by, and as more and more video is built into everything we do on the Internet there are more and more web sites and services that out of reach for such customers.
But unfortunately, there are also areas where it appears that the carriers have declared that they offer broadband where there isn’t any. If you were to draw something like a 5-mile circle around every rural DSLAM and every WISP transmitter you will see the sort of broadband coverage that many rural carriers are claiming. But the reality is that broadband can only be delivered for 2-3 miles, which means that the actual broadband coverage area is maybe only a fourth of what is shown on the Map. If you go door-to-door and talk to people outside of rural towns you will find a very different story than what is shown on the National Broadband Map. Unfortunately, the Chairman’s numbers are distorted by these weaknesses and distortions underlying the Map. There are a lot more rural Americans without broadband than are counted in the Map and rural America has far fewer broadband options than what the Chairman’s charts claim.
Tomorrow, a real life example.