Helping the FCC Get Broadband Right

The FCC’s annual inquiry on the state of US broadband is underway and we’re here to help. This process, mandated by federal law, seeks to discover whether advanced networks are being deployed across the nation. If the FCC finds they aren’t – if it finds gaps, for instance – it’s also required to take deregulatory steps to accelerate progress. The fundamental question is “whether advanced telecommunications capability is being deployed to all Americans in a reasonable and timely fashion.”

What Data Should the FCC Examine?

The FCC collects “Form 477 data” from carriers to map deployment, but this data has some limitations. It asks carriers to report on the services they offer by census tract, but census tracts are not households or business addresses. Census tracts (or “blocks”) are also not the same size, either in area or in population.

We’re more likely to find deployment gaps in rural areas, where the problems with census tract granularity are most severe. Rural tracts tend to be twice as large and less than half as populous as urban ones, for example.

Census tracts also have hard boundaries, which wireless systems don’t respect. There may be wireless coverage in areas where network planners don’t expect it to go, and there may not be coverage in all the areas where they do.

So there’s an inherent problem with using a grid based on political boundaries instead of starting with a coverage map and finding out how many people live in the bright colors. The census block approach is wireline-oriented, but we’re in the mobile first world now. So I would ask for coverage maps and them map them on population maps to count served and unserved people.

What About the Rest of the World?

Form 477 is also not helpful for international comparisons because other nations don’t necessarily take such an approach. International coverage comparisons are often drawn from OECD data, but nations tend to report their idiosyncratic measurements instead of taking a fresh look with a consistent approach.

This may not matter given that global data sets such as Akamai’s State of the Internet and Ookla’s Speedtest Global Index are the gold standards for nation-to-nation speed assessment. The nations with a serious commitment to advanced networks also use SamKnows testing, as we do.

So a comparison with all the other SamKnows nations would be meaningful once we apply necessary corrections for densely populated nations with very little rural population. As it turns out, the nations with high scores have very few rural residents as a percent of total population. They also have very high rates of multiple-dwelling unit housing, very cheap to serve.

But the traditional measurements of broadband speed, price, coverage, and use aren’t very useful.

What’s the Benchmark?

The big controversy is about the benchmarks the FCC sets for the definition of “true broadband”. The current threshold – I call it the Wheeler Standard – is 25 Mbps down and 3 Mbps up. If a rural provider offers 10 down and 1 up they can still get a subsidy, but the offering won’t show up in the FCC report as True Broadband.

The Wheeler Standard has no real function but to make it appear that we don’t have a competitive broadband market. A regulatory standard whose value is solely political is an abuse of regulatory power and should be abandoned. But you can’t revise a political standard without a political justification, can you? Apparently we’re stuck in politics if we keep the Wheeler Standard and also if we abandon it.

That’s unacceptable, so we need to think about what the broadband standard should be if Tom Wheeler had never been born. The law says the focus of the inquiry is on advanced networking capability, which suggests Congress is more interested in what we can do with our networks than with arbitrary labels.

Last Year’s Comment

A year ago, I filed comments with the FCC suggesting five priorities:

  1. Develop a coherent methodology
  2. Use clear terminology
  3. Consult public research
  4. Stick to the subject matter
  5. Focus in rural America

The FCC ignored this recommendation, but a lot has happened since. I think last year’s benchmark suggestion was correct:

This report to Congress on the state of broadband in the US will be the twelfth in the series. Regardless of its contents, we will not be able to examine the entire series as a unit to observe trend lines in any coherent way. This is odd because trend lines moving in the right direction are the hallmarks of progress.

This sad state of affairs comes about because every 706(b) report reads as if it were the first ever undertaken. Consequently, my overarching desire is for the Commission to recognize that the 706 report is a continuing obligation that should be discharged in a consistent, coherent, and objective fashion from year to year.

This means creating a methodology that does not require the FCC to create a new magic number for download speed and related metrics every year in order to exclude developments in hard-to-serve communities that are indications of progress. It also means defining “advanced telecommunications capability” in terms of application support rather than as a network-intrinsic capability. And it also means refraining from introducing squishy new network metrics that neither the Commission nor anyone else can measure or evaluate.

So let’s look at the networking requirements of the top applications and web sites and develop a benchmark that enables at least 80-90% of Americans to use all of them. My guess is the 25/3 standard will come off as too high on the download side and too low in the upload side.

It certainly favors cable modem over DSL, fiber, wireless, and satellite, so at least we need to correct that. Three Mbps is too low for any kind of cloud backup or cloud access, and 25 conveniently steps on 24 Mbps DSL offerings. It’s also just 2 Mbps above the average speed of mobile in the US. The 25 Mbps benchmark gives off a mighty stench.

What are the Roles of Wireline, Mobile, and Fixed Wireless Technologies?

The law directs the FCC to consider: “advanced telecommunications capability . . . without regard to any transmission media or technology, as highspeed, switched, broadband telecommunications capability that enables users to originate and receive high-quality voice, data, graphics, and video telecommunications using any technology.”

This leads to great deal of quite pointless discussion every year about whether wired and mobile are substitutes or complements. The real difference between the two is more about marketing than about technology.  Mobile data plans are capped at 20 – 25 GB/month, even if they’re sold as unlimited; wired data plans are capped in the 100s of gigabytes per month, or not at all.

For the time being, wired is the Netflix enabler but mobile is the network we really, truly want all the time. When consumers are strapped for cash and have to give up one of their two ISPs, the wire goes first.

So mobile is the primary connection and wired is an accessory that keeps the bill under control. As we move into 5G and caps grow higher, wired network connections to homes and small businesses will simply fade away except for special purposes.

So the report should enable us to see how fast our progress is toward pervasive, high-speed, mostly uncapped mobile connections. It’s not hard to figure out how to get there once you know where you’re going.

The Role of the Internet

The FCC and Congress have placed too much emphasis on broadband network price and performance progress and too little on the web and the Internet. Our wired networks are way over-built, especially so in consideration of their role as feeders to 5G and Wi-Fi.

All of our performance measurements should include a large dose of web page loading speeds and QoE bandwidth for non-web applications like video chats, voice, and IoT. The Measuring Broadband America program and the International Broadband Data Reports should be merged into the 706 report to provide a comprehensive picture.

And the overall report should emphasize applications at least as much as networks. We want to know if applications and networks are truly goosing each other as the former FCC imagined they do.

If they aren’t, we need to know why. All in all, the 706 report has degenerated into a political exercise in the last decade, more concerned with supplying advocates with talking points than with charting a course for industry development. It will probably take another year or two to get it back on course, but we can start now by clarifying our goals.

Clearing the Clutter

In 2017, “advanced telecommunications capability” means 5G wireless networks. One way or another, the FCC’s report should examine the rate of 5G deployment and the actions it can take to make the rollout as fast and painless as possible.

5G is going to make the edge networks of today – DSL, cable, FTTH and 4G – obsolete and irrelevant. Full deployment requires new chips, devices, and software. It also means permitting for small cell builds with backhaul to the legacy network.

And it means enriching the infrastructure for CDNs and providing QoE for 5G applications. Rather than spinning another round of the old networking battles over thresholds and “neutral networks” that don’t do what we need, the FCC should focus on the things it can do to resolve regulatory obstacles to universal 5G.

If it does this, the legacy wireline stuff will take care of itself.