dslreports logo
 
    All Forums Hot Topics Gallery
spc
view:
topics flat nest 
Comments on news posted 2010-03-12 10:16:49: Just a few days before the FCC is poised to introduce our first ever national broadband plan, the agency has launched a new java-based speed test application at the Broadband.gov website. ..

prev · 1 · 2 · 3 · 4 · 5 · next

Its a Secret
Please speak into the microphone
Premium Member
join:2008-02-23
Da wet coast

Its a Secret

Premium Member

Funny thing...

At the bottom of the first page it says you need java to run the test, but the next page requires flash... So which is it? Java or Flash based?

I think I smell a cow chip.

funchords
Hello
MVM
join:2001-03-11
Yarmouth Port, MA

funchords to Bill Dollar

MVM

to Bill Dollar

Re: The test is inaccurate...

said by Bill Dollar:

The FCC should take a cue from Offcom, and do a SamKnows style testing, which corrects for selection bias as well as all sorts of other intervening factors.
I agree that something like this is necessary to get an accurate picture. Whether we'll get there sooner or later is an open question. In either case, what we did yesterday is a start.
openbox9
Premium Member
join:2004-01-26
71144

openbox9 to Skippy25

Premium Member

to Skippy25

Re: Boost

But after a test, the FCC knows you have a connection and your address...which it most likely didn't before. Besides, what's that definition of broadband again?
openbox9

openbox9 to jus10

Premium Member

to jus10

Re: Dot gov?

Like »arin.net?

xirian
Premium Member
join:2003-01-26
Beacon, NY

xirian to funchords

Premium Member

to funchords

Re: Doesn't finish

it was safari. Of course both tests are wrong, one shows as 25/1, the other 5/5.
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

BHNtechXpert to Its a Secret

Premium Member

to Its a Secret

Re: Funny thing...

said by Its a Secret:

At the bottom of the first page it says you need java to run the test, but the next page requires flash... So which is it? Java or Flash based?

I think I smell a cow chip.
Naw, that smell is just congress

clickwir
join:2001-06-21
Dickson City, PA

clickwir

Member

Cold hard fact

Cold hard fact... this test will be limited by it's own use of Java as the platform for testing.

Sure, it's fairly wide spread. But most people don't use it, if they even have it installed. Those that do have it installed, you have to worry about the 987340958724535 different versions and if it even works.
chuckkk
join:2001-11-10
Warner Robins, GA

chuckkk to BHNtechXpert

Member

to BHNtechXpert

Re: Govt Seed test

It looks like there is a slow internal link at »miranda.ctd.anl.gov/ that is causing the limit.
Someone must have added a slow router.
Core0000
Premium Member
join:2008-05-04
Somerset, KY

Core0000

Premium Member

What the heck is Jitter?

Anyways, I took the test the first time, used OKLA.. and got
6479 kbps Download
509 kbps up

The jitter and Latency test did not work..

Second time time it was going to use a different service to do the test, but it did not work..

Took the test a third time with OKLA again.. and got
7377 kbps Down
723 kbps up
Lat 58ms
Jitter 20ms

Anyways, Newwave must have some kind of boost thing or something or the sites some how way off. I subscribe to 6/1 services from newwave.

paradigmfl
join:2005-07-16

2 edits

paradigmfl to BHNtechXpert

Member

to BHNtechXpert

Re: Broadband.gov speed test = Fail!

Having address data allows them to determine right down to the local neighborhood who is getting access and who is not. There has been controversy in the past that poorer areas or otherwise more rural areas were not getting broadband access. This is an issue to the government because it has and will be distributing stimulus money partially to provide broadband for the underserved.

I imagine many people will simply use a nearby address or such (not sure that this strictly legal but whatever). I agree that they should not be requiring address information (only requesting after a polite explanation) however I see where it could be considered helpful towards the stated goals at broadband.gov.

I also very much like the idea of the government partnering with M-lab (who is also researching network neutrality and providing tools for users to test this) and gaining information by which they can better monitor ISPs. I think it can potentially be a huge win for American consumers.

In regards to speedtests themselves there has been issues with ISPs artificially inflating speeds (yet other normal sites get the normal lower speeds) to known speedtests in order to deceive consumers. Perhaps eventually with help from M-lab the government can develop a more accurate test which can catch such things and provide a better metric for American consumers.
paradigmfl

paradigmfl to pnh102

Member

to pnh102

Re: Pointless

said by pnh102:

So now ISPs will simply optimize connections to that site so they can show higher speeds.

Meanwhile, things like unfees continue unabated.
There are ways to counter this though. If you read up on M-labs you will see that they are doing a lot of work on network neutrality and have considered such issues. While they definitely have a ways to go it's good to see a well funded and government connected organization begin to research and provide tools to the public.
paradigmfl

1 recommendation

paradigmfl to AnonPerson

Member

to AnonPerson

Re: Boost

said by AnonPerson:

Comcast and Time warner Boost are going to screw up the results quite a bit. Lets see if they find a way to overcome this.
Perhaps check out nano agent »www.gtnoise.net/nano/ and if you have a linux box install it and contribute data. The project seems ultimately goaled towards detecting such network neutrality violations and deceptions. Hopefully in the future they get a windows client going so more can help and test.

With proper research and implementation ISPs will have have a very difficult time getting around such tools and will not be able to escape detection.
paradigmfl

2 edits

paradigmfl to Camaro

Member

to Camaro

Re: Hey at least

said by Camaro:

I may get knocked for this but at least they are trying,the last administration wouldn't even tried a deal like this for consumers,yes it may be inaccurate and this that but it is a tool nonetheless that hopefully they will build a database for all isp's,but i agree that broadband min speed should be a hell of lot higher the kb range is pathetic.
Agreed. This is just the initial implementation and they are probably getting hammered far more than they expected. I'm sure that it will get better over time. I find the idea behind this to be great for consumers.

Btw, I also note the test claims to be "Beta" (at the bottom "(beta)" appears) which means it isn't officially released and is still in testing.

compuguybna
join:2009-06-17
Nashville, TN

compuguybna to clickwir

Member

to clickwir

Re: Cold hard fact

WOW! there have been that many revisions of Java?
What were the fixes in all of those?
said by clickwir:

Cold hard fact... this test will be limited by it's own use of Java as the platform for testing.
Those that do have it installed, you have to worry about the 987340958724535 different versions and if it even works.
compuguybna

compuguybna to Core0000

Member

to Core0000

Re: What the heck is Jitter?

By the way folks, OOKLA is the host of such speed tests as SPEAKEASY, SPEEDTEST.NET and a few others.

The test is usually very inaccurate, and a little on the HIGH side.

I can't believe the FTC is using OOKLA to host their speedtest...

linicx
Caveat Emptor
Premium Member
join:2002-12-03
United State

linicx to Shamayim

Premium Member

to Shamayim

Re: Dot gov?

Amen! The FCC does not need to harvest or share personal information.

graycorgi
Premium Member
join:2004-02-23

graycorgi to FFH5

Premium Member

to FFH5

Re: The test is inaccurate...

FCC's test says I get 2200kbps, which is impossible since my line barely can eck out 1500kbps. It's inaccurate, at least for me.
tmc8080
join:2004-04-24
Brooklyn, NY

tmc8080

Member

verizon recommends speedtest.net

but I get better results here:

»speedtest.tds.net/

Really, the results you care about are the actual use to the servers you connect to for web & other data xfers be it p2p or otherwise.. 25megabits nets you about 3.2 megabytes of data per sec-- pretty sweet if your connection remains rock solid.. and it's tough to get that consistent from many senders of data. I won't go into specifics.. but even most cablemodems top out @ 15 megabits upload, so unless the other end is rock solid on the upstream you can get the "upto" performance mixed in.. BTW, the sad fact is I only have a 400gb hard drive so, until I upgrade to a 2tb drive.. FIOS will be 'under utilized'.. hehe

Hopefully if we consumers find a DEMAND(aka use) for the service.. that in theory should bring more consumers onboard & bring prices down.. in reality-- Comcast, AT&T, Verizon, Cablevision and Time Warner are greedy companies who don't usually have the consumer's best interests in mind... so it will be a tug of war to get there.

AMDUSER
Premium Member
join:2003-05-28
Earth,
ARRIS CM8200
ARRIS SB6183

AMDUSER to funchords

Premium Member

to funchords

Re: The test is inaccurate...

said by funchords:

said by AMDUSER:

I tested my cable internet on it, and it said I had about 30 Meg (even though I only have 7 Meg Road Runner, but with Powerboost(tm).

If they want accurate, they are going to have to do something to mitigate the Powerboost to get the rated service speed.
If it measured 30 meg, you want it to dismiss it?

The goal here is to see whether you're getting what you paid for.

If it says 3 Mbps and you're paying for 7 Mbps, then you're not.
If it says 7 Mbps or 30 Mbps and you're paying for 7 Mbps, then you are.

Powerboost is real speed. So some credit should be given to that fact. But it is really temporarily and not assured, and I agree that something should also measure that. But what we don't want measurement tools to do is to discard actual performance.
What I would like is for the speedtest to test for actual internet access speed.. as 30 Meg is not available locally.
[At least not without paying for a business class data circuit [ie fractional T3 / OC3 for $$$$.]
BarneyBadAss
Badasses Fight For Freedom
Premium Member
join:2004-05-07
00001

BarneyBadAss to knightmb

Premium Member

to knightmb

Re: Boost

Um... er... could these folks be involved with this??

Barack
Nancy
Harry
Stenny

Could this lead anyone to a conclusion about how

1) accurate this is likely to be?
2) how useful this will turn out
3) how much more tax we'll be charged for this "new" tool that's already been developed by private companies and the G'ment wants to re-engineer it ?
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

BHNtechXpert to compuguybna

Premium Member

to compuguybna

Re: What the heck is Jitter?

said by compuguybna:

By the way folks, OOKLA is the host of such speed tests as SPEAKEASY, SPEEDTEST.NET and a few others.

The test is usually very inaccurate, and a little on the HIGH side.

I can't believe the FTC is using OOKLA to host their speedtest...
The Ookla test is probably fine for connections 20mbps and under. Beyond that the http based test they use fails to deliver consistent results at higher speeds. A socket based test would be preferred.

houkouonchi
join:2002-07-22
Ontario, CA

houkouonchi to Ericthorn

Member

to Ericthorn

Re: The test is inaccurate...

Well it doesn't seem to do well on fast connections (it got the speed of my home connection pretty accurately).



I get much better results to various servers on speedtest.net:








FFH5
Premium Member
join:2002-03-03
Tavistock NJ

FFH5

Premium Member

said by houkouonchi:


»www.la.bbb.org/Business- ··· 13131294
hottboiinnc4
ME
join:2003-10-15
Cleveland, OH

hottboiinnc4

Member

LMAO! the BBB. LMAO!

mr sean
Professional Infidel

join:2001-04-03
N. Absentia

mr sean

Eat Better!

said by Karl Bode:

...At a variety of broadband-centric websites (which we hear are filled with the worst sort of ruffians)
Yes, but ruffians are a good source of dietary fiber.

john131971
join:2003-05-05
Louisville, KY

john131971

Member

Rofl

Not sure if I should laugh or cry about all this, wonders about all the millions spent for a speed test. Another example of mismanagement and bloatware.

CW
@t-mobile.com

CW to iansltx

Anon

to iansltx

Re: They should partner with someone on this for servers

The Ookla and M-Lab uses different test algorithm to calculate throughput. M-Lab reflects user experience throughput=total bits (bytes)/time used
Ookla checks the instant throughput at rate of 30 samples per second, and average the upper half samples, which reflects 70-80 percentile of peak throughput
prev · 1 · 2 · 3 · 4 · 5 · next