dslreports logo
site
spacer

spacer
 
   
spc
story category
FCC: One Million Speedtests And Counting
Isn't using real-world data to inform government decisions fun?
by Karl Bode 02:02PM Monday Jul 19 2010
According to the FCC, they've now had one million people use their new speedtest. As we mentioned when the FCC launched the site back in March, this is really the first time the FCC has used honest-to-goodness real data to make policy decisions, which is frankly a little terrifying if you stop to think about it too long. Granted this is only the first step of the FCC's new speed data collection efforts and these kinds of tests don't give a full enough picture -- so the agency has also partnered with UK data collection firm SamKnows to get more detailed data using in-home customized routers. You can help by visiting the FCC's TestMyISP website.

view:
topics flat nest 
xenophon

join:2007-09-17

Honest-to-goodness data?

quote:
first time the FCC has used honest-to-goodness real data to make policy decision
Speedtest sites do not test some ISPs very well at all. I can range from 1.5-3Mbps on one site and 10-15M on another - same time, same location, same PC. Using a multi-streaming file test is the only method I found that truly and consistently tests performance.

»Best way to test network performance

On my EVO Android phone, the FCC test may only show 1.5-2Mbps while others show 7Mbps.

Someone needs to come up with a multi-threaded, multi-site speedtest method.

fifty nine

join:2002-09-25
Sussex, NJ
kudos:2

Re: Honest-to-goodness data?

I have to agree here. The speedtest does not show my true speed because it's limited by the 802.11g wifi on my ipod.
cahiatt
Premium
join:2001-03-21
Smyrna, GA

Speedtests

I see the same issue with different computers testing at the same site on the same connection. On one computer I can get 35meg down, the other maybe 8 meg down.

All these people with crappy wireless network connections are skewing the stats quite a bit.

pnh102
Reptiles Are Cuddly And Pretty
Premium
join:2002-05-02
Mount Airy, MD

Pointless

There's 2 big reasons why these tests won't mean squat for any sort of broaband policy:

First, you cannot confirm that ISPs are not gaming their routing to these test servers so that any connection automatically gets higher priority over other connections. We've seen ISPs do this in the past with speed test websites, what is to say they are not doing it now, especially when they want to look "good?"

Second, how does this data collection tell the FCC where broadband service isn't available? Isn't this sort of the whole point of the FCC needing more reliable data?

Speed testing isn't needed. Residential broadband speeds have been increasing ever since service has existed in the US. If what really matters is to find out where broadband isn't available, then perhaps the FCC should stop paying foreign companies with our tax money and instead hire a few unemployed US citizens to go through publicly accessible (i.e., non-secret!) broadband availability websites (I've posted them many times here), enter known US addresses, and find out which locations come up as unservicable. That process alone would give the FCC a better idea of where improvement is needed.
--
"Net Neutrality" zealots - the people you can thank for your capped Internet service.
Ecwfrk

join:2001-03-02
Fort Smith, AR

Re: Pointless

Exactly. I have 25mbps internet. My parents live in a small rural community Cox and AT&T have decided to not service while other neighborhoods a quarter mile away have DSL and Cable. They have to use Satellite which is absolutely atrocious with constant outages, speed dips and anemic daily caps. And they are lucky to be able to afford that. Most the people in that area can't afford the equipment and install fees of Satellite and are stuck with dialup that rarely exceeds 33.6k. Those are the places that need Broadband access.

pnh102
Reptiles Are Cuddly And Pretty
Premium
join:2002-05-02
Mount Airy, MD

Re: Pointless

Your direct observation provided more information about broadband availability in the US than all of the money the FCC has spent on this speed testing BS.
--
"Net Neutrality" zealots - the people you can thank for your capped Internet service.

fifty nine

join:2002-09-25
Sussex, NJ
kudos:2
Here's a novel idea, how about the census conducting a broadband survey? Kill two birds with one stone.

pnh102
Reptiles Are Cuddly And Pretty
Premium
join:2002-05-02
Mount Airy, MD

Re: Pointless

said by fifty nine:

Here's a novel idea, how about the census conducting a broadband survey? Kill two birds with one stone.
The only problem with that is most people are uncomfortable with letting total strangers into their houses to verify broadband connectivity.

The other problem that I think merits mentioning is that what we think of as broadband is changing. Most of us think of broadband as strictly a wired service connected to one or more desktop or laptop computers.

What about smartphone users with unlimited data?

What about MiFi type device users? Should they count as broadband too?

Does having access to 3G or 4G wireless service count as broadband? If it does, does that change what the broadband availability map of the US might look like? etc.
--
"Net Neutrality" zealots - the people you can thank for your capped Internet service.

tiger72
SexaT duorP
Premium
join:2001-03-28
Saint Louis, MO
kudos:1

Thank you SpeedBoost for skewing the results.

I wonder if the FCC takes temporary speed boosting technology into account?

Knowing the FCC, probably not.

jester121
Premium
join:2003-08-09
Lake Zurich, IL

1 recommendation

Re: Thank you SpeedBoost for skewing the results.

Of course not -- yet many people are thrilled at their involvement, on the misguided assumption that corporations are inherently bad and government agencies are inherently good.

Assuming a government agency will look out for your best interest and make intelligent decisions in matters of technology flies in the face of decades of evidence to the contrary.

Unit649
I B U, Who U B?
Premium
join:2000-01-22
Stockton, CA
Probably not, but seeing that most "speedboost" stuff is only a burst and usually doesn't last more than a few minutes, its likely not going to skew the results too much.

I have a feeling what the FCC is looking at is average speed anyway. Just because I have a cable connection CAPABLE of doing x k/sec down, it doesn't mean I'm using that, in fact, I know I have to use a lot less, lest I totally violate my caps.

I think they are looking more at saturation claims made by some providers as an excuse, and they are blaming the users for using the service to its full potential at times instead of making sure the network can handle the actual loads they claim.

I'm doubting anyone who torrents will be requesting a FCC router as well, so its likely that will be skewing the results as well-the people who get them will be doing basic surfing and will have a lot less usage than a power user-which will affect the results also.
wcnghj

join:2008-05-01

Speedboost

Speedboost is affecting these results, they need to make the download size larger..

HappyBunny9
Hi. Cram It.
Premium
join:2001-06-23
Long Beach, CA
kudos:1

What a shocker.

Fail. Test does not work in either Safari or Chrome, on Mac OSX.