4 ms·

> Why would an ISP bother building out the infrastructure if they can't extract a higher value from those that it most benefits (Netflix)? Netflix is not getti

by zanny 12y ago

> Why would an ISP bother building out the infrastructure if they can't extract a higher value from those that it most benefits (Netflix)? Netflix is not getting the benefit, the end user is. That is the demand side of things. They can and do charge more for higher bandwidth, and Netflix will require more than the national average to reliably stream at 1080p. Netflix only exists to fill that need, and the ISP exists to fill that users bandwidth need. It is not Netflix's direct need. > it must be mutually beneficial for both parties to go along. It is. Netflix makes more profit off working Internet from Comcast than they lose paying the bribe. But we lose, the consumers, because Netflix is getting a fast lane now, and sets a precedent that ISPs can hold ransom their bandwidth from anyone. > I think this fear is overblown, although I could be wrong. It takes no effort on their part to set up their routing to have everyone in a slow lane unless you pay the fee, and they have no motivation not to, because when smaller websites are slow, users stop visiting them, they don't drop their ISP over it. > Why should those consumers be subsidizing those that use broadband heavy services? They should not / aren't. At least in terms of month to month server upkeep. In terms of infrastructure wire, you pay the same taxes that maintain the roads if you drive a hundred miles a year as the guy who drives ten thousand miles a month. Previous models have showed that the overhead of calculating physical infrastructure usage (ie, power) is hard, and unless the per-unit volume is valuable enough it is not worth the bureaucracy. > Why should Netflix not be allowed to help subsidize the cost of providing broadband? Why should this fall solely on the individual? Because they are not subsidizing at all. End user will pay the same for Internet regardless of how much money content providers throw at ISPs because the vast majority are in monopoly positions and can do whatever they want up until the point users just drop service all together. > I don't see how that's a better scenario than currently exists. If it were the case that ISPs were the underdog here, because compelled by legislation to spend their private earnings to build cable that they cannot dictate price on and must lease compulsively, then you would have an argument. But current ISPs have the opposite. They inherited their infrastructure from public works and other telecom companies, where the laying of cable was heavily subsidized by taxpayer money, and in some cases (verizon) they were given money to invest in infrastructure that they then ran away with and faced no consequences. > decrease the barriers it takes to compete. There are only three ways to do this. One, you have "wire lanes" where anyone can run wire across the nation. This means starting an ISP no longer has the insurmountable mountain of regulation surrounding breaking ground, and you can just buy fiber, run it, and hook up each end user that wants it on an individual basis from the street corner. The downside is you end up with tons of redundant infrastructure investments, and wire itself is mad pricey. This means that the industry as a whole is inefficient, due to the redundancy. The good news would be that with laying cable being an actual option, entrenched players would have a motivation to lease what they have at reasonable rates to avoid someone else investing heavily in competing in the infrastructure. But it is also completely infeasible. There is no way to set up New York City or rural Arkansas in a way to just let anyone lay fiber cable. The investment alone in getting to that point vastly eclipses the costs of just laying a public fiber channel owned the state that anyone can hook up to. Option two is compulsory leasing, which is what Title 2 is in effect. If nobody can lay wire and actually compete with entrenched interests, you cannot have price competition on the rental rates of wire in the ground. So you have to use legislation to force it. Option three is either eminent domain of the telecommunications infrastructure or the build out of a publicly maintained fiber network. It could emulate the power companies, where regional switching stations are mandated to maintain the network within their jurisdiction using funds from acting as a private ISP, and the networks just need to be interconnected by mandate. This means that the wire is effectively free to use, and you just have to hook in servers and provide bandwidth. The problem there is that you need junction points where you connect to the backbone, and adding one to an already established network is on the order of complexity as laying wire itself. So it is either that, or wholly public ISPs with taxpayer funded Internet service. The downside to that is that you can look at all public infrastructure in the US and see how well that went given our current organization of government. The roads in many states are decrepit, the power grid is woefully in need of an upgrade, water services all over the place are being sold off privately and are also in deteriorating condition, and at least pedestrian rail is on the brink of death (no idea about commercial rail, but the prevalence of trucks in the PA area tells me it is not doing too well either). Our local, state, and federal government have collectively and decisively shown that given the political leanings and organization of state we in the US have, that none of them can maintain infrastructure, let alone improve it in the way we desperately need our Internet infrastructure improved over time, not just kept in its current state.

bko 12y ago
You make a lot of valid points. I guess in general I don't believe that a government can somehow prevent a certain type of commercial transaction and somehow create value. The obvious counterexample that comes to mind, and that you alluded to, is that of ransom. However, I think that's a bit of a stretch and makes light of actual violence. One thing that stuck out to me though: > If it were the case that ISPs were the underdog here, because compelled by legislation to spend their private earnings to build cable that they cannot dictate price on and must lease compulsively, then you would have an argument. I feel that opinions of issues are heavily influenced by the actors involved. Amazon vs Hachette being a key example. No one likes Comcast, nor do I, but I don't think that choosing sides based on the beneficiaries is the right way to go about things. I don't think you can argue that a person or group (corporation) took a benefit in the past and is now beholden to the person that dispensed that benefit. That's especially true if the decision makers and owners (stockholders) are a dynamic group. >The downside is you end up with tons of redundant infrastructure investments, and wire itself is mad pricey I normally discount this argument as it is often used as a justification of centralized control. Besides, one can argue that most commercial activities are redundant; from advertising to the endless variety of consumer goods available to people in the US. I do concede that this likely won't work in a lot of places as the technology currently exists. My hope is that some new technology can come about and change the economics materially. And I think that a light regulatory environment would best help foster this technology and encourage new entrants. Mesh networks anyone? http://www.wired.com/2014/01/its-time-to-take-mesh-networks-seriously-and-not-just-for-the-reasons-you-think/ http://www.wired.com/2014/01/its-time-to-take-mesh-networks-...
- zanny 12y ago
  Dunno who is downvoting you, this is a sound debate. You are not trolling and it is abuse of the downvote button. > I feel that opinions of issues are heavily influenced by the actors involved. It is not so much individual commercial actors than ISP's in general having power. If they were being exploited through regulatory abuse, then they would be in a defensible position. But none of the policy surrounding them harms them significantly more than it helps, in that most of their networks were built on taxpayer dollars and that is a key reason why nobody can compete with them without state funding of infrastructure. > from advertising to the endless variety of consumer goods available to people in the US I agree wholeheartedly advertising is one of the worst industries around, because it creates terrible returns on investment (the real productivity gain of advertising is more informing consumers about goods and services than using psychological manipulation to get people to spend money a certain way). But non-structural goods and services (ie, bread, or housing) has per unit per person utility. Common infrastructure like roads does not behave in the same "here are X units of good and Y units of buyer, and the market will adjust to make sure those numbers stay close to each other regardless of the number of competitors". With networking, power lines, roads, sewage systems, transit systems, etc - they all have the upfront opportunity costs, say the factory, but their marginal costs are nothing compared to them. Additionally, underutilized consumerate goods can just be shelved until sold. With bandwidth and power, unused power gets rerouted to inefficient storage mechanisms or just burns up on the wire. Unused electrical potential on the lines is just overbuilding without need, because maybe your wattage requirements in a region do not match the wires used (though this is rarely a thing, it is more often under-estimated). But there are no units to sell, and lost potential in the now is gone forever. If you build a highway that nobody drives on the investment was wasted, in the same way building a toy factory that never sells a toy is a waste, but for the highway the marginal cost of a driver on it (basically just wear on the road surface) is pennies, and the highway cost billions, whereas the toy factory cost millions and a million toys will sell for ten million. With network wire the same thing applies. Underutilization means wasted potential, where the wire and infrastructure around it was incredibly expensive but the per-unit costs of electricity transporting packets over it is negligible. This means unlike unsold toys, any second the wire is underutilized is wasted productivity and potential. The factory owner does care if hes manufacturing at capacity, but the market is much more fluid for toys and per-unit volume goods than it is for infrastructure. > And I think that a light regulatory environment would best help foster this technology and encourage new entrants. There are a lot of ways a regulation free market could solve our current networking mess. The biggest one would be that without "ownership" of wireless spectrum, you would have many more channels to broadcast on with radio technology, and could have radios that slide along much greater spectrum bands to find a compromise between bandwidth (high frequency) and range (penetration, mostly, at lower frequencies). You can balance usage more appropriately, and most everything from the 1mhz to 100ghz band is usable to some degree intelecommunications - the lower end is extremely low bandwidth, and the higher end has extremely short range, but in between you have a lot of potential value that is completely absent from the market as it is dedicated to private companies or one telecom provider. In such a market it would not be about last mile probably, because the wireless technology would be so good. It makes a lot more sense to have one router with 100gbit bandwidth for a suburb than gigabit ethernet to a hundred homes, because you have the same underutilization problem on a per-consumer basis, and that last mile cable was the most expensive. Besides that, in a regulation free environment, you could use wifi points to bridge property the owners would not let you lay wire on, or you could have communities agree to lay wire themselves. And of course mesh networks, but there is still no solution for a macroscopic mesh network encompassing millions of nodes - you eventually have to gate them and have centralized routing between subnets because of the algorithmic complexity of each added node in a mesh like that. And you can honestly do that today. You just need to wire between all your neighbors, and if they wire between all their neighbors, and you set up a subnet of your own with proper traffic monitoring, you would have a mesh net running under one backbone connection. But you still need that backbone, and you need to connect to it somehow.

by zanny 12y ago

bko 12y ago

You make a lot of valid points. I guess in general I don't believe that a government can somehow prevent a certain type of commercial transaction and somehow create value. The obvious counterexample that comes to mind, and that you alluded to, is that of ransom. However, I think that's a bit of a stretch and makes light of actual violence. One thing that stuck out to me though: > If it were the case that ISPs were the underdog here, because compelled by legislation to spend their private earnings to build cable that they cannot dictate price on and must lease compulsively, then you would have an argument. I feel that opinions of issues are heavily influenced by the actors involved. Amazon vs Hachette being a key example. No one likes Comcast, nor do I, but I don't think that choosing sides based on the beneficiaries is the right way to go about things. I don't think you can argue that a person or group (corporation) took a benefit in the past and is now beholden to the person that dispensed that benefit. That's especially true if the decision makers and owners (stockholders) are a dynamic group. >The downside is you end up with tons of redundant infrastructure investments, and wire itself is mad pricey I normally discount this argument as it is often used as a justification of centralized control. Besides, one can argue that most commercial activities are redundant; from advertising to the endless variety of consumer goods available to people in the US. I do concede that this likely won't work in a lot of places as the technology currently exists. My hope is that some new technology can come about and change the economics materially. And I think that a light regulatory environment would best help foster this technology and encourage new entrants. Mesh networks anyone? http://www.wired.com/2014/01/its-time-to-take-mesh-networks-seriously-and-not-just-for-the-reasons-you-think/ http://www.wired.com/2014/01/its-time-to-take-mesh-networks-...

zanny 12y ago
Dunno who is downvoting you, this is a sound debate. You are not trolling and it is abuse of the downvote button. > I feel that opinions of issues are heavily influenced by the actors involved. It is not so much individual commercial actors than ISP's in general having power. If they were being exploited through regulatory abuse, then they would be in a defensible position. But none of the policy surrounding them harms them significantly more than it helps, in that most of their networks were built on taxpayer dollars and that is a key reason why nobody can compete with them without state funding of infrastructure. > from advertising to the endless variety of consumer goods available to people in the US I agree wholeheartedly advertising is one of the worst industries around, because it creates terrible returns on investment (the real productivity gain of advertising is more informing consumers about goods and services than using psychological manipulation to get people to spend money a certain way). But non-structural goods and services (ie, bread, or housing) has per unit per person utility. Common infrastructure like roads does not behave in the same "here are X units of good and Y units of buyer, and the market will adjust to make sure those numbers stay close to each other regardless of the number of competitors". With networking, power lines, roads, sewage systems, transit systems, etc - they all have the upfront opportunity costs, say the factory, but their marginal costs are nothing compared to them. Additionally, underutilized consumerate goods can just be shelved until sold. With bandwidth and power, unused power gets rerouted to inefficient storage mechanisms or just burns up on the wire. Unused electrical potential on the lines is just overbuilding without need, because maybe your wattage requirements in a region do not match the wires used (though this is rarely a thing, it is more often under-estimated). But there are no units to sell, and lost potential in the now is gone forever. If you build a highway that nobody drives on the investment was wasted, in the same way building a toy factory that never sells a toy is a waste, but for the highway the marginal cost of a driver on it (basically just wear on the road surface) is pennies, and the highway cost billions, whereas the toy factory cost millions and a million toys will sell for ten million. With network wire the same thing applies. Underutilization means wasted potential, where the wire and infrastructure around it was incredibly expensive but the per-unit costs of electricity transporting packets over it is negligible. This means unlike unsold toys, any second the wire is underutilized is wasted productivity and potential. The factory owner does care if hes manufacturing at capacity, but the market is much more fluid for toys and per-unit volume goods than it is for infrastructure. > And I think that a light regulatory environment would best help foster this technology and encourage new entrants. There are a lot of ways a regulation free market could solve our current networking mess. The biggest one would be that without "ownership" of wireless spectrum, you would have many more channels to broadcast on with radio technology, and could have radios that slide along much greater spectrum bands to find a compromise between bandwidth (high frequency) and range (penetration, mostly, at lower frequencies). You can balance usage more appropriately, and most everything from the 1mhz to 100ghz band is usable to some degree intelecommunications - the lower end is extremely low bandwidth, and the higher end has extremely short range, but in between you have a lot of potential value that is completely absent from the market as it is dedicated to private companies or one telecom provider. In such a market it would not be about last mile probably, because the wireless technology would be so good. It makes a lot more sense to have one router with 100gbit bandwidth for a suburb than gigabit ethernet to a hundred homes, because you have the same underutilization problem on a per-consumer basis, and that last mile cable was the most expensive. Besides that, in a regulation free environment, you could use wifi points to bridge property the owners would not let you lay wire on, or you could have communities agree to lay wire themselves. And of course mesh networks, but there is still no solution for a macroscopic mesh network encompassing millions of nodes - you eventually have to gate them and have centralized routing between subnets because of the algorithmic complexity of each added node in a mesh like that. And you can honestly do that today. You just need to wire between all your neighbors, and if they wire between all their neighbors, and you set up a subnet of your own with proper traffic monitoring, you would have a mesh net running under one backbone connection. But you still need that backbone, and you need to connect to it somehow.