Is Your Internet Connection as Fast as You Think It Is?
An FCC report says ISPs don’t always deliver the speeds they promise, but they’re getting better at it.
The U.S. Federal Communications Commission released its first comprehensive study of broadband speeds across the United States on Tuesday. The study revealed that many Internet providers still advertise speeds higher than they deliver.
The report, “Measuring Broadband America,” was commissioned as part of the FCC’s efforts to promote improved broadband services across the United States. According to the Internet networking company Akamai, the U.S. ranks 14th in the world in terms of average Internet speeds, behind the Czech Republic, Latvia, and Belgium. Some U.S. ISPs have also been criticized for delivering Internet speeds that are lower than those advertised to users. The new report suggests that most providers now operate within 20 percent of their advertised speeds, even during peak hours; that’s an improvement over the figures recorded in a 2009 report from the FCC.
The report also quantified the effects on home broadband connections of peak-time Internet traffic, which happens between 7 p.m. and 11 p.m. While fiber-optic connections were barely affected, cable and DSL users saw decreases of about 5.5 percent and 7.3 percent in download speeds, respectively.
The report highlights two metrics as most indicative of broadband service quality: throughput, measured in megabits of data per second (Mbps), and latency, measured in milliseconds, which is the time it takes for information to travel across a segment of a network. On faster networks, the effects of latency are proportionately more noticeable.
While other reports, including that of Akamai, have taken a wide view of broadband use around the world, this study offers a more in-depth examination of services in the United States. Shane Greenstein, a professor of management and strategy at Northwestern University’s Kellogg School of Management, says this kind of report is needed. “We did not have a mature electricity industry until everybody agreed on how to measure use of electricity. And we cannot reach a similar state in our broadband industry without a similar agreement. This is a big step towards that, even though we still have a longer conversation in front of us.”
The report is the result of a collaborative study with 13 major ISPs, including Comcast and Verizon; academics and other researchers, including from MIT; and consultants and consumer organizations. SamKnows, an analytics company, was selected to administer the FCC’s broadband performance testing initiative.
According to the report, the FCC examined 6,800 homes and “conducted 13 different tests in each home, multiple times per day, over several months, to produce more than four billion data points from more than 100 million tests of broadband performance.”
The report also quantifies the connection speeds required by consumers for various tasks. For basic Web browsing—“accessing a series of Web pages, but not streaming video or using video chat sites or applications”—a speed of one Mbps is sufficient, says the report, which also found that after 10 Mbps, there’s no significant increase in page download speeds for basic Web browsing.
Sascha Meinrath, director of the New America Foundation’s Open Technology Initiative, points out that some providers seem much better at delivering their promised speeds than others. “This study is indicative of the need for some sort of truth in ISP advertising,” he says.
Meinrath also says that because the FCC did not release this data in advance, his group and other third-party researchers are only now beginning to review the data. He expects thorough analyses to take a few days.
A statement by FCC chairman Julius Genachowski summed up the FCC’s perspective: “I expect broadband providers will look closely at the data we’re releasing today and ensure they’re providing accurate, relevant, and easily understandable information to consumers about their services. Providers should be aware that this survey isn’t intended as a one-time thing.”
Become an MIT Technology Review Insider for in-depth analysis and unparalleled perspective.Subscribe today