A week after releasing online tools that let consumers determine, and report, the real Internet access speeds they are getting, word comes that more than 150,000 have used the service — with the FCC planning to ultimately use all of that data to drive broadband policy. The new tools are available as Web applications and as mobile apps for iPhone and Android devices. They are essentially front ends to two online test services: Ookla and M-Lab. According to Ookla, the average U.S. download speeds found are 11.5 Mb/s. According to M-Lab, 7.04 Mb/s.
What’s more interesting is what the FCC is planning next, says this Ars Technica report:
To really guide its analysis, the FCC has issued a “request for quotation.” It plans to pay one company to provide a statistically valid, geographically diverse sample of 10,000 broadband connections. The participants in the study will have their lines analyzed for a month or so by small hardware devices that sit inline on the network and record latency, speed, jitter, uptime, and more.
Between the national mapping efforts funded by last year’s stimulus bill, the consumer-facing Web and mobile tools rolled out by the FCC, and the new 10,000 person sample, real broadband data could be in hand within the year.
Connected Planet’s take, Joan Engebretson:
The real question is what the FCC plans to do with this data. The National Broadband Plan, released by the FCC this week, makes a point of defining broadband for the purposes of the new Connect America fund as 4 Mb/s “actual” download speeds — so perhaps the commission sees consumer tools such as these as a means of policing telcos’ performance. The new tools also seem to dovetail well with the commission’s attitude toward competition, which is based largely on the premise that educated consumers will drive a higher level of competition.
That’s our take on this, let us know what you think in the comments below:
The FCC is missing the mark in their so-called \\"speed tests\\". It\\\’s fine to collect the data, but it\\\’s virtually useless for what they intended.
The FCC should have required anyone taking the speed tests on their site to answer three questions first: 1) What speed do you subscribe to from your ISP? 2) What is the fastest speed offered by your ISP? 3) What is the fastest speed offered by ANY ISP serving your location?
So the FCC, intending to find out what speeds are available to the public in any certain location, is really only finding out what speeds that consumers subscribe to — not the AVAILABLE speeds at that location.
Mighty big difference there when determining how much broadband is available across the country.
Mikey, I don’t think FCC is trying to figure out advertised availability so much as trying to determine what people really have in terms of broadband capability. If all they wanted to know where your three questions, they wouldn’t need a speedtest app to get those answers.
Second, I’d bet most consumers don’t know the real answers to those questions so there would be big questions as to the validity of the data received.
So, while I don’t know exactly what the FCC wants to do with the data they are gathering from the test apps, I don’t think asking the questions you suggest would have a lot of value.
Also, related to this, I tested my link (Comcast) using the FCC tests and noticed something that may or may not be interesting. I normally test my connection through DSLReports about once per month. Over the past couple of years I’d been getting results well in excess of 10Mbits/sec with requisitely high upload speeds reported.
I subscribe to 8Mbits/sec and was always leery of the results I was seeing. I would try different tests available from DSLReports and some other major testers and get roughly the same very high numbers.
As my gut feel and some anecdotal tests I run indicate I’m really getting about half of what is being reported and even when my connection seemed ’slow’, the speed tests would indicate in excess of 10Mbits/sec with good latency and ping times. I began to suspect that Comcast might be using traffic shaping to provide speed tests with greater bandwidth thus making my service appear faster than it really is. While of course I can’t prove such a thing, the FCC test results show 4Mbits/sec to 5.5Mbits/sec which jive with my anecdotal testing and my gut feel.
I tested my line through DSLReports today after the FCC tests and got 4 to 5.3 Mbits/sec. It’s ‘interesting’ that my reported speeds from DSLReports have decreased by 50% since the FCC started providing their tests.