|
|
|
|
|
by Junk_Collector
2658 days ago
|
|
If I am using voice traffic, it is very sensitive to routing delay. The total transit time that packet route takes is much more important than the amount of bandwidth I need to throw at it. If I am using an emergency service such as 911, that packet must keep going through regardless of other traffic congestion no matter how many people are streaming Netflix at the moment. Netflix itself is relatively immune to impact from group delay and doesn't need prioritization for reliability but it does require a large amount of bandwidth to be available in order for it to function properly. A packet, is not just a packet that can be simply interchanged if you want maximum performance out of your network. Also you have the issue that sparked the large public prominence of Network Neutrality. Netflix was renegotiating their peering agreements and the ISPs were trying to soak them on the last mile costs. ISPs felt Netflix should bear the cost of network expansion since their service was generating all the traffic that they needed to upgrade to accommodate. Netflix felt the cost should be primarily carried by the ISPs. This has almost nothing to do with you as a consumer, but Netflix successfully branded it that way to garner public support. N.N. complicates adds a lot of uncertainty to what is legal to negotiate in a peering agreement between companies. L3 is probably going to be the big loser here, not Comcast. And finally, because I see a lot of posts on this chain that don't understand what Power Factor is, it probably needs to be pointed out. A refrigerator has an inductive motor in it's compressor that it uses to chill the air. A toaster is a simple resistor. The inductive nature of that motor degrades the power network by transforming some real power into imaginary power (mathematical terms) and basically messing up the phase relationship on the distribution network itself. This ratio of real(useful) and imaginary(useless) power is called Power Factor and the power company measures it and asses a fee based on how much you screw it up. A typical household does not run enough of anything to noticeably impact the networks power factor so most consumers don't know about it but if you were running a hundred refrigerators you would absolutely be charged more than running a thousand toasters. What does that have to do with packet networks? Well, when the traffic (electrical load) impacts the function of the distribution network, the cost is passed on to the user that generates that load. Not all Kilowatts of power consumed are equal. |
|
Netflix was already paying for an uplink connection of sufficient bandwidth. The customers were already paying for a downlink connection of sufficient bandwidth. The assertion that it's necessary to hold one party hostage in order to get more money from the other is absurd.
UPS doesn't demand that Amazon pay for its new trucks, and it certainly doesn't hold Amazon packages hostage until they do so. They're supposed to buy their infrastructure with the profit that they're already generating. And ISPs are seeing record profits right now.