Trust Restored, FCC Broadband Measurement Group Looks to Future

In August, several ISPs spoke of a breakdown in trust, which they traced back to a flawed period of data collection related to problems at some M-Lab nodes in March (CD Aug 6 p5). Now, the state of trust within the group is “pretty good actually,” said David Young, Verizon vice president. “Like any group of people, you go through periods of disagreement and tension and whatnot, but I feel comfortable that we've worked through it, and we've returned to a level of trust that allows us to achieve the common objectives."

"I can’t disagree with that,” said M-Lab engineer Thomas Gideon, technical director at New America Foundation’s Open Technology Institute. On a broader level, the group has released two reports and been “heads down” studying September data, and are “really just focused on getting a good high quality collection,” and “leveraging all the experiences” as they continue to improve their data collection practices, he said. “And I'd rather focus on that continual improvement, regardless of where we may not see eye-to-eye."

Although everyone is pleased with the measuring project, the FCC’s stated goal -- comparing advertised speeds to what’s actually delivered -- isn’t the main concern in rural areas, said Martha Duggan, senior principal-regulatory affairs for the National Rural Electrical Cooperative Association. “Our issue is much more basic: When will we even get broadband service?” Duggan said rural areas need street-level reporting of broadband availability to households, with a much greater ability to “peel back the onion."

The commission would need much bigger samples for more localized reports, said Steve Morris, NCTA vice president. It’s hard to get a valid broadband speed sample for a company that has just a few thousand customers, he said. That the FCC and NTIA’s National Broadband Map provides data on the census-block level is “pretty unbelievable,” he said. NTIA settled on census blocks because it was “the right balancing point between something that would on the one hand be meaningful, but on the other hand not be so burdensome for broadband providers to provide that they would balk and not participate at all,” he said. At the time, an issue was a real-time map at the street level had an increased potential for mistakes, Morris said: Now that NTIA’s a few years into it, if the agency re-asked the question, “maybe companies would be more comfortable” at the street level.

As FCC speed testing continues, there’s a need for transparency not just in the data itself, but with the studies that use that data, said Allan Friedman, research director at the Brookings Institution’s Center for Technology Innovation. Even though the U.S. is not first in broadband, “we are certainly first in broadband data,” he said. “If you have this much data, you can show anything,” he said. “And some people will show anything.” The group will need to make sure the studies that come out are as open and transparent as the data themselves, he said.

Even though the U.S. is not first in broadband speeds, the country is making a lot of progress, Morris said. DOCSIS 3.0, a standard that allows high-speed broadband over cable, is available to 80 percent of Americans, he said. In almost all cases, cable ISPs sell broadband services that reach at least 50 Mbps, and some offer services up to 300 Mbps, he said. “There’s no other country that has cable infrastructure like we do."

Now that there are data, Friedman wants the FCC to revisit the earlier metrics and models to see what was estimated, and how accurate the sampling models were compared to the actual data. Gideon said there’s a need to make data more understandable, and “maybe address more than one question at a time.” Young’s interested in seeing a “more holistic” look at what’s available to consumers, including 4G, he said. “All of the data shows pretty clearly that the state of broadband in America is pretty good and getting better.”