Re:Google "uses" 21 times more bandwidth than it pays for
by chaostheory on Friday December 05, 2008 @11:00AM
(#387119)
I still don't understand how this article got over 20 up mod points, let alone 2. Did the people upvoting it even read it?
Re:Google "uses" 21 times more bandwidth than it pays for
by olefoo on Friday December 05, 2008 @11:44AM
(#387207)
I don't know, but since it was posted right before midnight and a number of funky things are happening with voting on this thread... the possibility that we are being played does need a look.
Lets say google used 100% of all bandwidth, they would need to pay for 334mil/.165 = 2024.24. Now to calculate the overage: Other peoples share = 44000-2024 = 41976 divide by google payment (2024.24) = 20.7 - magic - the overage number didn't move! Still 21.
Now lets try the reverse.
Say google used 1% of trafic: 334/16.5 = 20.25 - other people will then pay 44000-20.25 = 43980 divide by google cost = 2172, adjust for 1% usage (* .01) = 21.72.
My head asplode - is seems no matter how much google uses they always seem to have an overage of 20 to 22 times as much!
Could it be, no, but could it be that home users pay 21 times too much for bandwidth??
No couldn't be, then this firm "funded by broadband companies" would have to tell it's people to drop prices by 22 times.
Seriously this has to be the stupidest study ever funded. I wonder if the authors of the study put their names on it. They completely ignore the fact that the recipient of the traffic should also pay for it, so lets see how much of that 44bil is home users, adjust for 16% and then we'll see.
Edit: As I thought, the authors did not put their name on it. And to make it stupider the 44bil is consumers bill only! Doesn't include costs paid by any other company!
They are seriously comparing the costs of home users vs the cost of google, and ignoring every other site on the internet! It's like they think the only thing home users do is browse google and nothing else.
At the very least adjust that 44bil by 16% to at least account for that. That would turn the overage into 3 times as much - which would imply that wholesale prices for bandwidth are 3 times cheaper than retail.
Actually ignore that last conclusion - the study is riddled with so many errors I would not want to draw any conclusion whatsoever from it.
Take a look at this little gem:
...[some math]... trucks pay 4 cents/mile, cars pay 1 cent/mile. So "trucks/businesses pay over four times more for their usage of the U.S. Interstate Highway system than cars/consumers do."
Is he brain dead? I mean if you want to compare adjust for the weight of the vehicle, the volume of it, the number of people you could fit in it, something.
But no - just straight compare.
(PS. Damage to roads goes by 4th power of weight. So 3000lb car/2 axles=5.1 * 10^12 of damage, and truck is 34000 per axle=1.3 * 10^18 of damage - so trucks do 254,901 times as much damage to roads as cars do, but they pay just 4 times as much.)
And he testified before congress! Is there some way to tell them he's an idiot?
You have to look at the graph on page 11 - he extrapolates data in such a way that he assumes that by 2012 google will have a 130% share of the video-to-pc market. Yes you read that number right, google will have 30% more share than actually exists. Not sure how they'll manage that.
PS: Someone please register for his site (he - blog author - is the author of the "study") and post this as a comment - I give you copyright permission. Let's see if he'll have the guts to post it.