Tag Archives: speed test

Cable, fiber systems deliver broadband service at or near advertised speeds, DSL generally doesn’t, FCC report says

by Steve Blum • , , , ,

Fcc 2018 broadband report download

The FCC’s primary broadband metric is now the 80/80 benchmark: the minimum speed that 80% of users experience, 80% of the time during primetime viewing hours. When evaluated against that benchmark, cable modem and fiber-to-the-home systems do a reasonably good job of delivering service at advertised speeds. Among Californian providers, only Comcast fell noticeably short, with actual download speeds hitting around 90% of what they promise.

Telco DSL-based service doesn’t do so well. According to the FCC’s latest field tests, AT&T’s and Frontier Communications’ legacy DSL services – the kind you often find in rural California – deliver speeds that are about 60% of what they promise. AT&T’s advanced DSL systems – the upgraded kind that go into high potential neighborhoods – score around 90%.

On the upload side, cable and fiber providers generally meet or exceed their promised speed levels, but telco copper systems do even worse, again with the exception of AT&T’s upgraded systems.

Fcc 2018 broadband report upload

People use Internet service differently now than they did seven years ago, so the Federal Communications Commission added consistency metrics to its annual report on broadband performance in the U.S. It’s an acknowledgement that a steady stream of data, to support online video viewing, is more important than the occasional bursts of speed that old school web browsing requires…

We found that for most ISPs, actual speeds experienced by subscribers nearly meet or exceed advertised service tier speeds. However, since we started our MBA program, consumers have changed their Internet usage habits. In 2011, consumers mainly browsed the web and downloaded files; thus, we reported average speeds since they were likely to closely mirror user satisfaction. By contrast, by September 2016, the measurement period for this report, many consumers streamed video for entertainment and education. Both the median measured speed and how consistently the service performs are likely to influence the perception and usefulness of Internet access service.

The FCC bundled all of its telecoms and media research – wireline and mobile broadband, and video – into one giant data dump. The report includes a well-deserved shout out to the CalSpeed mobile broadband speed testing program. It’s run by the California Public Utilities Commission and even the FCC considers it a valuable and independent source of information about what mobile carriers (and, soon, wireline ISPs) actually deliver.

Much of the data was held back by the FCC for up to two years. John Brodkin has a good write up on that problem in Ars Technica. Ajit Pai won’t explain why this is the first time the broadband speed and availability analysis was released since he became FCC chair.

In some ways, Pai is remarkably open about FCC deliberations compared to his predecessors. He routinely releases draft decisions three weeks before commissioners vote. In the past, drafts were kept out of the public eye, although lobbyists with sufficiently deep pockets always seemed to know what was coming. But Pai is also cagey about what he releases, holding back this latest round of broadband data, as well as details regarding the millions of apparently bogus emails uploaded to FCC servers during the net neutrality debate.

Commissioners are scheduled to formally adopt the findings at their meeting on Wednesday.

FCC Communications Marketplace Report Collected Appendices, 4 December 2018 (this is the big document with the interesting data)

FCC draft Communications Marketplace Report, 21 November 2018

DSL service hammered in FCC report

by Steve Blum • , , ,

Cable modem service is reliable and fast, and getting faster, but DSL service offered by telcos isn’t. That’s the top line conclusion of the Federal Communications Commission’s 2016 annual report on fixed broadband service across the U.S…

When DSL is used to provide broadband service, the maximum advertised download speeds among the most popular service tiers has increased only slightly since 2011. In contrast, for cable services, the maximum advertised download speeds among the most popular service tiers have increased from 12-30 Mbps in March 2011 to 100-300 Mbps in September 2015.

The increase in cable’s advertised speed is attributed to upgrades from DOCSIS 2 to DOCSIS 3 technology. Cable is also better at delivering advertised speeds…

For subscribers to DSL-based broadband service, the increase in median download speeds has varied among ISPs, with most ISPs showing little or no change. For subscribers to each of the participating cable broadband services, there have been fairly steady and substantial increases in median download speeds…The actual speeds experienced by most ISPs’ subscribers are close to or exceed the advertised speeds. However, DSL broadband ISPs continue to advertise “up-to” speeds that on average exceed the actual speeds experienced by their subscribers.

When consistency of performance is factored in, the difference between DSL and cable modem service is even starker. The chart above shows the difference between advertised speeds and what a user can expect can expect to get consistently (i.e. what 80% of the users tested experienced 80% of the time during peak periods). The actual, consistent speed delivered by most telcos is less than half the advertised speed.

The exceptions are Windstream, which is right around 50% and AT&T’s Uverse service, which delivers advertised speeds consistently about 70% of the time. But AT&T’s legacy DSL is the worst of them all, with actual speed only 35% of advertised. AT&T lobbied the FCC to report those two types of service separately. If AT&T’s overall performance was averaged, it would be in the same range as the rest.

The report also notes falling performance figures for satellite Internet service, which it attributes to increased congestion as more users jump on to limited capacity satellites, and relatively flat results for fiber-based service.

2016 Measuring Broadband America Fixed Broadband Report, a Report on Consumer Fixed Broadband Performance in the United States, Federal Communications Commission, 1 December 2016

Mobile broadband test results speeding back to the FCC

by Steve Blum • , , , , ,

The FCC’s mobile broadband speed test app for Android is a hit. In its first two days, it was downloaded and installed on 30,000 devices. It’s been out now for two weeks, and its getting a 4.4 out of 5 rating on the Google Play store.

Those first two days produced 40,000 reports from all over the country. The FCC says that all 50 states and all the major carriers are represented in the data received so far. It’s proving popular with Californians. 3,500 tests were run in and around Los Angeles and 2,700 in the San Francisco Bay Area, the top two reporting regions in the country.

Source: FCC.

The free app runs in the background, periodically testing cellular and WiFi broadband connections and reporting the information – completely anonymously, the FCC claims – back to a central server. It can also be triggered and viewed manually. The app tests download speed, upload speed, latency and packet loss, then bundles the results together with location, signal strength and device characteristics and ships it all off. By default, it limits itself to using no more than 100 MB of data in a month. Users can change that setting.

So far, the FCC is not releasing either the raw data or compiled results, but at least some of that will come in time. As with the California Public Utilities Commission’s CalSpeed app, (rated 4.8 on Google Play) it should provide useful ground truth about mobile carriers’ coverage claims. So far, CalSpeed test results have shown that actual performance rarely comes close to what’s promised.

If you’re interested in the details, the source code is available on GitHub.