I have no trouble believing that both Comcast and Netflix are in the wrong.
None of the ISPs can actually deliver what they advertise unless you are spending hundreds (or thousands) of dollars per month for a business connection with an SLA.
Sure they do, but if traffic goes up substantially then equipment needs to be upgraded or added to handle it.
When ISPs send and receive the same amount of traffic from each other they usually don't charge each other for peering (they share the costs equally). When the relationship is no longer balanced one ISP usually starts charging the other ISP for transit.
Akamai was colocating caching servers with the ISPs and most likely installing new lines to carry the traffic and passing this cost on to Netflix.
Netflix didn't like that because it's hard to make a profit when your customers are costing you more in bandwidth fees than you collect in subscriptions so they tried to get L3 to dump all the traffic onto Comcast's network without paying for transit.
In a fair market the customers using more bandwidth would pay more than the customers using less bandwidth.
There's more than one way to get there. Netflix users could end up paying more for their broadband service than the non-Netflix users. Alternately, the Netflix subscription itself would cost enough to account for the higher bandwidth costs.
Personally I'd like to see metered bandwidth so the everybody pays for exactly what they use but the way ISPs manage peering agreements gets the job done as well.
Don't forget the role of Akamai. The reason that Netflix switched from them to L3 is because Akamai was charging them the true cost of moving that many bits across the country.
Netflix's business model couldn't handle that price to they tried to get L3 to cheat for them.
The problem is that Comcast has already been compensated for services that they've never had to deliver. Now that they have to deliver more, they want to be paid more. They're the ones who choose their business model, they're the ones who should suffer from it.
Comcast was delivering the traffic and Akamai was paying them to do it. Now that Netflix switched providers L3 is going to end up paying the same fees that Akamai was already paying.
Netflix was sending all its traffic through Akamai and Akamai was paying Comcast for the unbalanced traffic. Netflix switched to L3 to try to get a better price so now Comcast is just imposing the same terms on L3 that they were previously imposing on Akamai.
His point is that local bandwidth is cheap but long-haul bandwidth is expensive and the equipment necessary to stream the kind of bandwidth Netflix needs to a significant portion of their customers simply can not be purchased and maintained for the current price of a residential broadband connection.
Since the traffic can not be carried at the current price it won't be, because no amount of complaining or regulating will make the impossible happen. One way or another somebody is going to pay the true cost of moving the bits or else they aren't going to get moved.
I can't directly confirm his numbers but the guy ran a major ISP for several years and has no reason to lie about it now.
I generally respect Karl Denninger's viewpoint on these issues since he was one of the people actually involved in building out the internet.
It's not about content, it's about volume and flows, and who pays for the infrastructure build necessary to handle them.
What amounts to poaching other people's resources works well right up until you drive that other party into the wall and force them to spend a crapload of money for which they receive nothing in return. That is, they don't receive any renumeration for the additional expense - but you do!
This is the base problem with all overcommitted services where the business model is predicated on fractional use of maximum possible resource consumption. When that model is violated costs go up dramatically. This is ok provided the person who has the cost also gets the revenue that is occasioned by the violation of the original model.
But in the case at hand, Netflix and similar get the revenue, but Comcast gets the cost.
I saw this one coming a mile away. If L3 manages to get the FCC involved and Comcast is prohibited from doing this they will be forced instead to either cap-and-charge customers or dramatically raise their prices, which will also blow back on the content folks like Netflix.
Suddenly that $8 "video any time" subscription becomes not $8, but $28 as Comcast adds another $20 to your monthly cable internet bill.
And there goes the pricing model that everyone loves so much about Netflix!
I've got a weimaraner and a lab/weimaraner mix and they are both crazy smart. Keeping the two of them mentally stimulated enough to prevent boredom is almost a full time job.
Protip: the last thing you want in your house is a bored weimaraner.
In amount of actual deaths per amount of effort, I suspect bombing a plane is still the most efficient.
The metric a rational terrorist would use is amount of terror per unit of effort.
Is the idea of being blown up while waiting in line to go through security (or for that matter, standing in line anywhere) more or less terrifying that the idea of being blown up in mid-air?
The fact that none of this is happening in the US despite how easy it would be to do and how impossible it is to stop proves that actual terrorists are extremely rare.
The wikipedia page talks about the "well dressed man" and congressional testimony revealed that various TLAs knew about him and intentionally chose not to revoke his passport or put him on the no-fly list.
The State Department didn't revoke the visa of foiled terrorism suspect Umar Farouk Abdulmutallab because federal counterterrorism officials had begged off revocation, a top State Department official revealed Wednesday.
Patrick F. Kennedy, an undersecretary for management at the State Department, said Abdulmutallab's visa wasn't taken away because intelligence officials asked his agency not to deny a visa to the suspected terrorist over concerns that a denial would've foiled a larger investigation into al-Qaida threats against the United States.
"Revocation action would've disclosed what they were doing," Kennedy said in testimony before the House Committee on Homeland Security. Allowing Adbulmutallab to keep the visa increased chances federal investigators would be able to get closer to apprehending the terror network he is accused of working with, "rather than simply knocking out one solider in that effort."
Slashdot Mirror
I have no trouble believing that both Comcast and Netflix are in the wrong.
None of the ISPs can actually deliver what they advertise unless you are spending hundreds (or thousands) of dollars per month for a business connection with an SLA.
What exactly did he do to earn the title "most despised internet citizen"?
They were paying for their own bandwidth via Akamai but it turns out that they can't actually afford it. That's why they switched to L3.
Sure they do, but if traffic goes up substantially then equipment needs to be upgraded or added to handle it.
When ISPs send and receive the same amount of traffic from each other they usually don't charge each other for peering (they share the costs equally). When the relationship is no longer balanced one ISP usually starts charging the other ISP for transit.
Akamai was colocating caching servers with the ISPs and most likely installing new lines to carry the traffic and passing this cost on to Netflix.
Netflix didn't like that because it's hard to make a profit when your customers are costing you more in bandwidth fees than you collect in subscriptions so they tried to get L3 to dump all the traffic onto Comcast's network without paying for transit.
In a fair market the customers using more bandwidth would pay more than the customers using less bandwidth.
There's more than one way to get there. Netflix users could end up paying more for their broadband service than the non-Netflix users. Alternately, the Netflix subscription itself would cost enough to account for the higher bandwidth costs.
Personally I'd like to see metered bandwidth so the everybody pays for exactly what they use but the way ISPs manage peering agreements gets the job done as well.
Don't forget the role of Akamai. The reason that Netflix switched from them to L3 is because Akamai was charging them the true cost of moving that many bits across the country.
Netflix's business model couldn't handle that price to they tried to get L3 to cheat for them.
Yes, they were paying Akamai who was charging them the true cost of moving that many bits across the country.
They didn't like how much that cost so they got L3 to take the traffic for less money.
The only way L3 could beat Akamai's price was to cheat by abusing their peering agreement with Comcast.
Comcast declared shenanigans and L3 is trying to cover their tracks by riling up the net neutrality crowd.
Pretty much
Comcast was delivering the traffic and Akamai was paying them to do it. Now that Netflix switched providers L3 is going to end up paying the same fees that Akamai was already paying.
ISPs have always taken into consideration whether or not the traffic is equal in both directions when negotiating peering agreements.
Netflix was sending all its traffic through Akamai and Akamai was paying Comcast for the unbalanced traffic. Netflix switched to L3 to try to get a better price so now Comcast is just imposing the same terms on L3 that they were previously imposing on Akamai.
If the other comments on this article are any indication it's going to work.
I do believe this was one of the suggestions mentioned in the article I linked to.
Of course this shifts some of the costs on to the customers who don't use Netflix but maybe they won't complain as loudly as the Nexflix users.
His point is that local bandwidth is cheap but long-haul bandwidth is expensive and the equipment necessary to stream the kind of bandwidth Netflix needs to a significant portion of their customers simply can not be purchased and maintained for the current price of a residential broadband connection.
Since the traffic can not be carried at the current price it won't be, because no amount of complaining or regulating will make the impossible happen. One way or another somebody is going to pay the true cost of moving the bits or else they aren't going to get moved.
I can't directly confirm his numbers but the guy ran a major ISP for several years and has no reason to lie about it now.
I generally respect Karl Denninger's viewpoint on these issues since he was one of the people actually involved in building out the internet.
Yes, but three lefts do.
I've got a weimaraner and a lab/weimaraner mix and they are both crazy smart. Keeping the two of them mentally stimulated enough to prevent boredom is almost a full time job.
Protip: the last thing you want in your house is a bored weimaraner.
It's very dependent on the breed. I have a friend whose dog does this.
That just means it took longer to fail than the other competitors.
They left out the most viable competitor.
The reason you are surprised is that terrorists are far more rare than you've been lead to believe.
Of course the government is doing its very best to manufacture domestic terrorists so at some point you must assume that they will be successful.
The metric a rational terrorist would use is amount of terror per unit of effort.
Is the idea of being blown up while waiting in line to go through security (or for that matter, standing in line anywhere) more or less terrifying that the idea of being blown up in mid-air?
The fact that none of this is happening in the US despite how easy it would be to do and how impossible it is to stop proves that actual terrorists are extremely rare.
The wikipedia page talks about the "well dressed man" and congressional testimony revealed that various TLAs knew about him and intentionally chose not to revoke his passport or put him on the no-fly list.
Where did you come up with that number?
Here's the data from the US DOT for miles driven and death rates.
As far as the 1 per 500,000 miles I am using 6 million car accidents per year in the US. If you have better data I'd be glad to see it.
Exactly - near misses happen all the time but rarely result in an accident.