Search:  

 
 
   News
newer
story category Open Source Bandwidth BS Detection
Creating a neighborhood watch for UK broadband consumers....
04:03PM Wednesday May 07 2008 by Karl Bode
tags: business · bandwidth · world · networking
Tired of users being offered half their promised speed, half-truths about traffic shaping, and constant whining by UK ISPs when customers actually use their connection, UK industry-watcher Sam Crawford developed a plan. He wants to send out 200 free open source firmware updated Linksys routers and create a reliable dataset of user disappointment. Using hardware instead of software allowed more reliable 24/7 monitoring, according to the Register. From the project FAQ:

We're deploying small hardware-based monitoring units to volunteers all around the country. Once connected to the volunteer's home network the units will perform a series of tests at regular intervals during the day, every day of the year. The results of these tests are fed up to our reporting engine and combined with the results of others on the same ISP to form a national view of how that ISP is performing. . . The units are checking latency, packet loss, DNS query times and failures, web page loading times, as well as the obligatory suite of speed tests. Additional tests are also in development too. The goal is to provide a statistically sound measurement of overall broadband quality.
UK consumer watchdog "Which?" recently reported that users in the UK often see a third of the speed stated in advertisements. The group found that buyers of an "up to 8 Mbps" tier actually achieved 2.7Mbps on average, and some didn't see speeds any faster than 0.09Mbps. Nobody surveyed saw speeds faster than 6.7Mbps.

Here in the States, AT&T conducted a similar test for marketing purposes, but instead of routers, actually purchased 150 Comcast connections. In a presentation, AT&T then claimed they found that peak downstream speeds were between 3-4Mbps, while average downstream speeds for the users ranged between 300kbps and 400kbps, significantly less than the advertised rate of six to eight megabits per second.

Related:
  1. Connected Nation's $134 Billion Fish Tale?
  2. 16.4 Tbps Over 1,500 Miles
  3. Bell Canada Redefines 'Satisfaction' and 'Fairness'
  4. Getting Around Bell Canada's Traffic Shaping
  5. Remember How The Net Neutrality Fight Began
  6. Bell Canada: Throttling Aids Innovation
  7. Monday Evening Links
  8. Friday Morning Links
Forums » Open Source Bandwidth BS Detection
view: topics flat text 
Post a:
SilverSurfer

join:2007-08-19

No Surprise Here

Of course these ISPs will consider no such suggestion as a free, open source router that will accurately gauge anything...such a device will blow the collective, bullshit whining/excuses/traffic shaping excuses completely out of the water.

justbits
More fiber than ATT can handle
Premium
join:2003-01-08
Chicago, IL
·AT&T Midwest
·AT&T Yahoo

Arms race?

I can see this becoming an arms race.
Each ISP will try to figure out which of their customers has one of these devices, then they'll prioritize the packets for those customers so their benchmarks are better.

Also, who's to say that Sam Crawford's real plan isn't just to extort money from ISPs to report falsified data that makes their network seem to perform better compared to others?
espaeth
Misanthrope
Premium
join:2001-04-21
Minneapolis, MN

The mark of any good network..

30-40% monitoring overhead.

It would be fun to compile stats on what percentage of BBR reader's Internet traffic is speed tests.

wruckman
Ruckman.net

join:2007-10-25
Northwood, OH
·RoadRunner Cable

BS Detected

I think it is a good idea. I have no idea why people allow companies to promise one thing and then not deliver it. I want what is promised and to be told exactly what to expect. Such as, bandwidth shaping and so on. It should all be disclosed fully up front. Half truths are half lies but still lies alike.

SSX4life
Premium
join:2004-02-13
>_ lolz

Cut the bullsh!t... it's about time

ISP's complain about users sucking up too much bandwidth

while...

Users complain about not getting the speeds advertised

It's a vicious cycle.

Imagine if Ford / GM / Honda / Toyota / etc. all posted that their cars get 30 mpg highway but you only ever got 20, can you think of the lawsuits? Same thing happens with ISP's the only difference is they hide behind all the legal bullsh!t in their contracts.

WTB dumb pipes.

k thx bye.
--
»www.google.com is your best friend... please use it before asking your question.
EPS

join:2008-02-13
Hingham, MA

Re: Cut the bullsh!t... it's about time

Hm, clearly you don't know how mpg figures work if you think those are actual, in-use figures...

SSX4life
Premium
join:2004-02-13
>_ lolz

Re: Cut the bullsh!t... it's about time

said by EPS See Profile :

Hm, clearly you don't know how mpg figures work if you think those are actual, in-use figures...
I come darn close to the EPA mpg rating on all my cars. The point being is that Ford / GM / Honda don't cripple your car or f*ck with the car after you buy it and drive it down the road. You sell something you don't screw with it, impose regulations on how the user can drive the car, where they can drive it, or who travels with them. You see ISP's doing this all the time.

This is my point and ethics would say that I'm also on equal ground here.

DEE02TA

@ti.com
Hmm, that used to be the case. However, starting in 2008, they are much more realistic. That is why all the MPG ratings went down in 2008. They are now pretty close.
EPS

join:2008-02-13
Hingham, MA

Re: Cut the bullsh!t... it's about time

Ah, my mistake then- the car I drive most often is a few years old and so probably has the inflated figures, good to see they're fixing that system.

Seandhi
Seeing From a New Level
Premium
join:2003-04-19
Humble, TX

said by EPS See Profile :

Hm, clearly you don't know how mpg figures work if you think those are actual, in-use figures...
I don't know why, but I actually fair 2 MPG better than the EPA rating... Go figure.
--
You're an enlightened cat, and I dig that.

dvd536
as Mr. Pink as they come
Premium
join:2001-04-27
Phoenix, AZ

said by SSX4life See Profile :

Imagine if Ford / GM / Honda / Toyota / etc. all posted that their cars get 30 mpg highway but you only ever got 20, can you think of the lawsuits?
When was the last time you got anywhere near what was on the sticker? i think those are based on optimal conditions, going downhill with a 35mph wind behind you.
--
When I gez aju zavateh na nalechoo more new yonooz tonigh molinigh - Ken Lee

beerbum
Premium
join:2000-05-06
Here!
·Comcast

speeds

I hope they won't be using sites with heavy traffic and or low bandwidth.. otherwise the results will not be valid..

attempting to base download speeds using a site or sites with say.. only a T1 connection or better yet hosted on a persons home computer with residential dsl/cable service -- pointless since the source would be bandwidth limited..

en102
Canadian, eh?

join:2001-01-26
Valencia, CA
·RoadRunner Cable
·DSL EXTREME
·DSL EXTREME

Spin doctors...

Like to put a spin on all output

said by Karl :
In a presentation, AT&T then claimed they found that peak downstream speeds were between 3-4Mbps, while average downstream speeds for the users ranged between 300kbps and 400kbps, significantly less than the advertised rate of six to eight megabits per second.
'Peak downstream = 3-4Mbps' can happen with 'Average dowstream speeds of 300-400kbps'

Eg. Streaming audio and video typically do not run at 6Mbps, but at their fixed rate. Also, running many apps (i.e. streaming + download + BT + botnet/spam/virus + etc.) may not show the whole picture. Also, DSL has ATM overhead, and you'll be luck to hit ~85% of peak rates.
Eg.

Speed Test #50478653 by dslreports.com
Run: 2008-05-07 18:33:30 EST
Download: 2568 (Kbps)
Upload: 441 (Kbps)
In kilobytes per second: 313.5 down 53.8 up
Tested by server: 2 java
User: 297537 @ dslreports.com
User's DNS: dslextreme.com
Compared to the average of 138 tests from dslextreme.com:
* download is 51% better, upload is 23% better

2568 ~= 3008 * 0.84

--
Canada = Hollywood North

Vahid

@dslextreme.com

Re: Spin doctors...

I don't know about you but using dsle 6 mb I get a constant peak of 5.4 mbps It's steady as well when i use cable i can't get anywhere near that steady rate even though it's at 10mbps. Sad that they want to crap up their service ever futher by imposing caps on their already crappy service.

en102
Canadian, eh?

join:2001-01-26
Valencia, CA

Re: Spin doctors...

I'm past 10,000' so I can only get the 3008/512kbps package, and do get a stable line which will get me ~2600kbps throughput - all the time.
--
Canada = Hollywood North

jonnyz
Premium
join:2003-03-20
Canfield, OH
clubs:
·Armstrong Zoom In..

Speed Test #53887774 by dslreports.com
Run: 2008-07-07 22:01:02 EST
Download: 4741 (Kbps)
Upload: 474 (Kbps)
In kilobytes per second: 578.8 down 57.8 up
Tested by server: 190 java
User: 788114 @ dslreports.com
User's DNS: zoominternet.net

My line is sold as 5000/512

4741/5000 = 94.82%
474/512 = 92.58%

I'm glad my independent ISP actually cares about advertised speeds
--
Join the RC5 team.
iansltx

join:2007-02-19
Fredericksburg, TX
·magicjack.com
·Comcast
·Qwest.net
·Sprint Mobile Broa..

Overhead by butt

The point is, barring botnets etc. on your computer, you should get the speeds advertised. No matter how much overhead is being eaten up by ATM. An eight megabit connection is an eight mesabit connection, and no matter what transport medium, you should be able to get that eight megabits whenever you want.

Hopefully these open-source routers will get things done with regard to mainline ISPs working ad advertised. C'mon guys, upgrade your networks and people will stop whining!
dynodb
Premium,VIP
join:2004-04-21
Minneapolis, MN

Re: Overhead by butt

You know the car makers who advertise a 300 horsepower engine? Guess what- after drivetrain losses you'll be lucky to see 255 horsepower actually delivered to the wheels. Same thing with broadband.

That's just the way bandwidth has traditionally been sold, be it the 56k connection to a cash machine or an OC-3 feeding an ISP- by speed of the physical connection, not observed throughput.

One could offer a higher sync rate to compensate for overhead, but overhead percentage varies with packet size; it's not a constant and thus any attempt to do so would be an educated guess.

Furthermore, all providers- cable, DSL, and the ISPs oversubscribe- that's why a dedicated 1.5/1.5 Internet T1 costs $400 and a 1.5/1 DSL connection costs $35.00. Unusually high traffic on a node? Less bandwidth available.

Perhaps the time will come when bandwidth is cheap enough that speeds can be guaranteed, but that time hasn't come.
Forums » Open Source Bandwidth BS Detection


Friday, 05-Sep 19:44:04 Terms of Use | Privacy Policy | Hosting by www.nac.net - DSL,Hosting & Co-lo | feedback | contact
over 9 years online! © 1999-2008 dslreports.com.republican-creole