The Internet Works Fine for Those Who Have It

As we move into the third month of sheltering in place, it’s becoming evident that the isolation is taking a toll on the stability and lucidity of some of our neighbors. One recent news item chastised the FCC for claiming that all Americans are connected to broadband networks, for example.

The claim was bizarre because the FCC has never said any such thing; in reality, the agency’s recent appendix to the 2019 Broadband Deployment  Report simply declared that the nation continues to make reasonable and timely progress toward extending advanced networks everywhere that people live.

While hyperbole is commonplace in communications policy, this example was egregious because it was made by an FCC commissioner who had presumably read the report in question before voting to disapprove. In fact, the FCC says the country’s digital divide has narrowed but not yet closed: 5.6% of Americans cannot subscribe to wired broadband today.

Professor Meinrath’s Shocking Improbability

Another critic suffering from a lack of perspective, Penn State affiliate professor Sashcha Meinrath, beseeches all who will listen to declare the Internet utterly broken. In an op-ed in The Hill he claims: “increased home internet use is killing our connection speeds across the country.”

Meinrath is a long-time overseer of a former Google project known as Measurement Lab (M-Lab) that makes broadband performance data available to researchers, so he should know better. Authoritative performance benchmarks such as SamKnows and Speedtest have come to starkly different conclusions.

SamKnows is a for-fee service supported by national regulators around the world, including our FCC, so it must be accurate. Its system runs on specialized equipment installed in consumer households that runs a variety of sophisticated measurements.

Speedtest is the go-to resource for consumers troubleshooting balky connections. It runs 10 million tests each day on a network of ten thousand test servers located in thousands of locations. Speedtest is owned by Ookla, a division of the Ziff-Davis publishing company with no ax to grind.

Imperceptible Problems

The SamKnows Critical Services Report finds a nearly imperceptible slowdown in US broadband: “Our results show that almost all states have seen a slight drop in download speeds. The largest drop we observed was 3.9%, but the majority saw around a 1% decline.”

Speedtest made a similar finding: As of May 4, average wired broadband speed in the US was down 2%, and mobile speed was up 2% compared to the week of March 2nd.

M-Lab is a rudimentary tool incapable of assessing the high speeds of today’s networks. It was built by Google in 2009 as a free service to the research community, part of a public relations blitz aimed at convincing regulators tasked with examining proposed mergers with direct competitors such as Yahoo and DoubleClick that Google was one of the good guys. For the past decade, M-Lab has languished as broadband speeds have taken off.

While most speed tests fill the broadband pipe with data to assess capacity under realistic conditions, in the U. S. M-Lab simply measures the speed at which a single data stream traverses the path between the user’s computer, the user’s home network, the ISPs, and one of nine measurement locations.  By contrast, browsers use 8 – 16 data streams in parallel, the scenario that both SamKnows and Speedtest duplicate. SamKnows also provides regulators with single stream performance data, but it does not pretend this measurement says anything about the nature of the broadband pipe as a whole.

Coverage Rather Than Speed

Perhaps because he wants us to disbelieve our own experience, Meinrath combines the claim that the Internet is crumbling with a plea for more spending on rural broadband. Neither the FCC, the ISPs, nor the public interest objects to greater subsidies for upgraded rural networks.

Similarly, none of these players objects to increased subsidies to enable low income Americans to pay for computers, routers, and broadband services. Some carriers even subsidize such purchases out of their own pockets: the Comcast Internet Essentials program stands out.

But neither the FCC nor the ISPs dictate the budgets for new networks and new Internet users: these funds follow highly political Congressional mandates. Taxpayer spending in the wake of the pandemic is going to be heavily scrutinized, so hard choices will need to be made between performance upgrades and user subsidies.

It behooves us to be realistic, to assess facts honestly, and to avoid rushing to solve non-existent problems while real needs remain unaddressed. The Internet works fine for those of us who have it, but it doesn’t work at all for the rest of us.

We must fix that problem with sober, data-driven analysis, leaving the hysteria at the door.