By Rob Curtis - Deployment Director and Steve Rosenberg - Manager of Infrastructure, Omnibus Broadband Initiative.
In the Plan, we wrote that 14 million Americans in 7 million homes do not have access to broadband service offering actual download speeds of 4 Mbps and actual upload speeds of 1 Mbps. We found closing this gap would cost approximately $24 billion. The Plan recommended closing the gap by fundamentally refocusing the FCC universal service fund.
These figures are based on our analysis of the best available data inside the FCC and available from 3rd party sources. Today, we release The Broadband Availability Gap, a staff technical paper detailing the methodology and model we used in our calculations. This is one of the most extensive, data-driven, detailed, and comprehensive analyses of broadband networks to date. We believe it uses the best possible approach in light of the data currently available. Releasing the detailed documentation of this effort also makes this one of the most transparent network analyses ever undertaken.
While complicated, the model we developed for the Plan essentially does two things.
First, it estimates the areas of the country in which 4/1 Mbps is not likely to be available in the next several years. Our approach uses public and commercial data and relies in part on a statistical model to estimate the availability of broadband in every census block in the country. This analysis focuses on the capabilities of the “last-mile” infrastructure (the access network), not on either subscribership or retail offerings. As we learn more from better FCC data gathering and state broadband data collection funded by NTIA, these estimates will improve.
Second, it estimates the cost of bringing 4/1 Mbps to those unserved areas and the revenues that could be earned by doing so. In doing so, it is conservative and technology-neutral. We only modeled technologies—wireless, cable, satellite, and DSL—that are commercially-deployed today or will be in the near future. We wanted the model to inform practical solutions, not rely on promises of future breakthroughs.
While those steps sound straightforward, this paper shows that the task itself is complex. The financial model includes thousands of inputs, ranging from estimated signal propagation of wireless networks in particular geographies to the density of soil for trenching fiber optic cable in others.
The FCC has multiple servers supporting the model, and doing a run takes as much as 12 hours; the output from one such run produces roughly 4-5 gigabytes of data (and sometimes more). The calculations required to reach the $24 billion gap require 10 model runs. Needless to say, a lot of time, sweat, and yes, on occasion, tears have gone into developing, building and supporting this model.
In developing the Plan, we are committed to openness and transparency. To this end, on May 6 at 3:00 pm, we will host a forum at the FCC to present an overview of our analysis and answer questions. We stand behind our assumptions and conclusions in the technical paper. We look forward to feedback and discussion so that this analysis informs the policy process in the most impactful way.