Network News

X My Profile
View More Activity

FCC responds to questions about broadband speed tests

The Federal Communications Commission responded Friday evening to questions about the broadband speed tests it launched Thursday, saying that data collected from the effort will be supplemented by other information in the agency’s analysis of connection speeds.

The point of collecting data on broadband speeds and service quality is so that the FCC knows — down to the level of a home address — how well different parts of the United States are being served by Internet providers.

About 80,000 people have tried the tests — using an iPhone application and a Web-based test — in the first day. Some users expressed concern about widely varying results, saying that because the tests were voluntary, the agency wouldn’t get a full portrait of broadband access and speeds in any geographic area.

An FCC lawyer, Jordan Usdan, responded in a statement Friday that information from the tests would be combined with other data, such as a mapping data, and used in the agency’s final broadband proposal, which is scheduled to be presented to Congress next week.

“There is a real value add here, giving the FCC another granular data layer of broadband service availability,” Usdan said. “Yes, software-based tools can provide individuals with inconsistent performance results, some of which are out of the control of the ISP. ... However, this crowd-sourced speed data will be useful: Given a large sample size, the FCC can analyze performance trends over time and on a comparative basis across large geographies.”

Public interest groups say that data collected by the FCC should include the names of ISPs providing the service so the agency can know how many providers operate in different areas. And information on transmission speeds should also be accompanied by data on prices for service.

By Cecilia Kang  |  March 12, 2010; 10:11 PM ET
Categories:  Broadband , FCC  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Cable companies lose court challenge of FCC on channel access rules
Next: China flexes its muscles, balks at Google in online censorship battle: reports


The FCC has not yet responded to questions about the usefulness of its data, given that there is no guidance telling users that they should ensure no other applications or users on their networks should be consuming bandwidth during the test.

Cecilia, if you were to run the FCC's test from the Washington Post network, while other users of the network are doing research, uploading articles to the blog, downloading YouTube videos, watching Congressional hearings, and listening to Pandora, you will only measure the amount of bandwidth available after all the other users are accommodated.

If someone in my home or my office is watching streaming video and someone else is downloading application or OS updates, the bandwidth available to me for a speed test is pretty limited, and the FCC will get a very inaccurate image of my net connection if I use its speed tester.

Another issue is that the FCC's test doesn't disclose (at least to the public) what network location is being used to assess one's network performance.

This matters hugely. I live and work in the Washington area. If I do a network test using I can select the server; likewise I can select an NDT Web100 server. If I use a server in the DC area that has a fat pipe, I will get fabulous up- and down-stream figures as well as very low ping and jitter. This accurately reflects the service my ISP provides.

If, on the other hand, I use a server outside the area, even in a highly connected area such as San Francisco or Silicon Valley, my ping is much higher (simply because of the increased distance), jitter may be higher, and up- and down-stream speeds will vary greatly.

I could also select a server that is not connected to the Net by a big pipe, or is in a foreign country far away, or in a rural area with limited bandwidth. The poor scores resulting from such a test would be falsely attributable to the ISP I use, because the FCC's test records my IP address (and thus it knows how I connect to the Net) but it doesn't (as far as I know) record any characteristics of the server used for the test.

Posted by: MichaelDSullivan | March 13, 2010 1:32 AM | Report abuse

URL for test

Posted by: lostinthemiddle | March 13, 2010 11:23 AM | Report abuse

I did the test and somehow magically got double the upload & download speeds than I'd ever seen before. That was a rather unexpected result which I am unable to replicate anyplace else.

Posted by: Nymous | March 13, 2010 12:23 PM | Report abuse

Cecilia Kang, the advertising for whose blog is provided by Google, conveniently fails to mention in this article that one of the two tests mentioned in the article was developed by a group funded by Google and staffed by Google employees and lobbyists. She also fails to mention that they access Google servers -- thus biasing the results in favor of ISPs which have better connectivity to Google.

Posted by: LBrettGlass | March 14, 2010 1:43 PM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2010 The Washington Post Company