Net Neutrality Is Just a Symptom
corp.sonic.net
corp.sonic.net
Excluding Virgin's cable service, the last-mile infrastructure is wholly owned by BT Wholesale which, despite the name, is "independent" of BT Retail. It's operations are heavily bounded by legislation and politics - it is, in effect, a public corporation. So tied down are they, in fact, that they have a flat price list for almost all customers. Pricing innovation is precisely nil. I've have a copy from two years ago, but I'm under NDA wrt specifics.
(Note that none of the above is a criticism, the idea of letting them free to do what they want is horrifying.)
The result of this is strange to observe from an economists vantage point. On the one hand you can catch a glimpse of what perfect competition might look like in the consumer market as a thousand companies vie to maximise market share with an almost perfectly commoditised good, which in reality means packaging (with TV, phone, mobile, insurance, pretty much anything.)
Aaand on the other hand most of the country is paying £30/month ($50) for 6Mbps down/0.2Mbps up at 50:1 contention and a fair usage policy to boot.
Of course this is much better than straight up nationalisation as the private companies will at least innovate on peripheral services like billing and support, and will use a stronger punch when it comes to motivating BT Wholesale to stay on their toes.
Is this better than the US situation? Most probably. Is this the ideal solution? Most probably not.
I have a 152Mbps connection from Virgin which is around the same price, and am now in a super-exclusive club of Brits who get to moan about not being able so saturate a servers pipe (I want my downloads to read 18MBps every time, dammit!)
The other part of the regulatory environment here that's worth mentioning is the "MAC-code" system which makes it relatively straightforward to switch ISPs. The ease with which customers can vote with their feet is a major part of what keeps (most) retail ISPs competitive.
This may well be, but unbundling itself has had a very large impact on competition and prices. Currently the cheapest broadband plan including line rent is about £10 from plus.net. Not many countries can compete with that and no country without unbundling comes even close.
Since network netrality violations are frowned upon, eyeball ISPs have resorted to playing favorites and extracting payments from over the top / web service providers.
The battlefields are in carrier interconnect points and Internet exchanges, where ISP, transit providers and service providers meet to exchange bits coming and going to broadband users.
The weapons of choice are denial of peering connections and congestion of existing connections.
The shakedown is as simple as it is effective. Want your bits to go through? Pay up.
Some of these peering wars have made it to the press, but most don't being shrouded in NDAs and trade secrets. Here are some examples of the public ones: - Free in France vs. Youtube - Orange vs. Megaupload - Orange in Africa vs. Google - Netflix vs. US ISPs
Public or not, one thing is for sure these peering wars are only about to get worse.
What is needed to avoid these and other potential problems are just basic consumer protection rules: - Truth in advertizing - Minimum performance guarantees - Independent third party verification
The funny part is that we don't need new laws or regulations. We just need to actually enforce the laws we have. And sufficiently penalize infractions to make it economically unfeasible not to comply.
One moment of sadness passes. Let's go and make some money. Gotta pay the bills.
> they include net neutrality, the price of Internet access in America, performance, rural availability and privacy.
In a free-market, rural availability won't improve. Someone posted this the other day: http://www.louisvilleco.gov/SERVICES/CityManagersOffice/Fran.... That page describes the cable franchise negotiations for Louisville, CO, population 4,500. The page notes: "Should another cable provider want to offer cable service in Louisville the City would offer that company the same franchise opportunity that Comcast now has. To date, no other service providers have asked for a franshise."
The only reason phone infrastructure got built in rural America at all is that the Telecom Act's Universal Service Fund diverted billions of dollars a year from long-distance customers to paying to subsidize service for rural customers. What will the free market do when you take that away?
> But in 2004, the FCC took steps to limit competition, turning away from key provisions of the 1996 Telecom Act. They set aside unbundling requirements which serve as a key bridge for competitive carriers. By circumventing Congress this way, the Bush-appointed Chairman of the FCC was able to turn back a competitive tide, creating an intentional duopoly on Internet access in the US.
The "competitive tide" under unbundling is a fiction. In the early 1990's, you could order DSL from a large number of ISP's . . . all of which used the same copper wire the AT&T monopoly had put in the ground decades before. It was not accompanied by increased investment into the actual wires that ran into peoples' homes. Indeed, it eliminated any incentive to invest in those wires--why pour billions of dollars into infrastructure you'll have to lease out at cost to your competitors?
People mention the "unbundling" requirement as if its a panacea, comparing it to what has been done in Europe, but ignore what makes that policy work in places like the U.K.--the company that owns the wires is a sanctioned monopoly, protected from competition, that is guaranteed return on investment through statutory means.
It's illustrative to look at places in the U.S. where the market for telecom is working pretty well. In Chicago, there is cable via Comcast, and both AT&T and RCN building FTTP networks. In a place like Baltimore, RCN wouldn't be allowed to build its network, because they only serve the wealthier neighborhoods from the Loop northward. When local governments intervene in the market and make such deployments impossible, rational market participants do not respond by building a bunch of unprofitable infrastructure. They do it by not building infrastructure at all.
That's called a utility. Why not do it that way? Then you have a market among retailers. Then you don't need regulation because the one's that dick around with your streams can be selected out of the market.
Someone here probably knows, but I am sure any service provider (LEC, CLEC, Municipality) has to file tariffs in order to get approval to deliver service.
Not in the UK. Anybody can lay their own last mile. It's simply that it'd be expensive to do so, and the company which owns most of the last mile across the country is forced to share it fairly with all ISPs.
Virgin are one company which lays their own last mile, there's another cable company which operates in Scotland which does as well, and there's a few rural community ISPs which lay their own last mile.
This would make for some interesting market dynamics. Does this mean that dominant last-mile owners try as hard as possible to own as equal a share a possible to minimize triggering a "majority must-share" threshold?
They do have a sibling company which is an ISP, but the telecoms regulator ensures that there's no preferential treatment going on.
The trouble is without competition they have exactly the same lack of incentive. Why invest in those wires when you can charge customers the same amount whether you make the internet faster or not, and the customers don't have any better alternative?
There are two valid ways out of this. The first is to somehow create competition at the wire level, but that seems rather implausible. Running new fiber to every American household is so expensive that nobody wants to do it unless they can have a monopoly. Creating competition would require some plurality of companies to do it even though the competition, if effective, would eat all of their margins. If anyone can describe a way to make that happen it would probably be worthy of a Nobel prize.
Which leaves the second alternative, which is writing the incentives into the regulations. Set it up so that the nationwide average ISP speed is regularly measured and the ISPs with above average network performance are allowed to charge more by an amount that justifies making regular upgrades.
Or just require the last mile provider to install fiber everywhere once and then let the competing ISPs who lease it provide their own terminating equipment, since the terminating equipment is the thing that needs to be upgraded but isn't the thing that constitutes a natural monopoly.
> People mention the "unbundling" requirement as if its a panacea, comparing it to what has been done in Europe, but ignore what makes that policy work in places like the U.K.--the company that owns the wires is a sanctioned monopoly, protected from competition, that is guaranteed return on investment through statutory means.
Your criticism seems to be that we aren't doing exactly the same thing they're doing. Is there a reason we can't just do exactly the same thing they're doing?
Did the guy who figured out how to get running water to every American household, also win a Nobel prize?
My criticism is that it's not sensible to implement one part of a two-part solution.
> Is there a reason we can't just do exactly the same thing they're doing?
In principle, I'm fine with the idea. In practice--basically New York is the only municipality I'd trust with infrastructure like that.
Seriously?
By the way, I do have a full solution to both parts of the problem, which this margin is too narrow to contain. :)
I'm in earnest here: I can't exactly divulge the trade secrets of my approach but I do think there's a good solution to it.
And I'm really happy to say that Sonic.net is on the right track. If you're in their service area, try them out. You won't look back.
Edit: I'll add Verizon FiOS is fine.
Regulatory regimes are systems. Leaving out key provisions can result in a "solution" worse than the problem. It's like leaving a support off a bridge design--you don't get partial credit. If you impose unbundling without creating a compensating incentive structure, you'll just drive investment out of the industry and end up with stagnation.
Also, the choice isn't between Comcast/Verizon and Ofcom. It's between Comcast/Verizon and the City of Baltimore (or San Francisco, or <other city run by crazies here>).
But non of this works if you have one company or the state build up a network and only then you start competition. Some people ignore the diffrence between free markets and privatisation/monopoly deregulation. This is the problem we are facing in tons of places. Free market get a bad wrap because privatisation often produces bad results.
Solving this problem now is really hard. One solution is that you just have the government build all the wires and rent them out low level to companys. Another is to deregulate completly and hope after some dark times market pulls threw. Another is to start doing fine grained regulation of every aspect.
I am very sceptical about government regulation that involve things lime you are suggesting, regulatory capture and the revolving door are a real problem if your regulation is to technical.
I dont have the solution, economiclly speaking all of these are very suboptimal.
That is not a valid statement. If a cable company has to put more money building the network in rural areas than what it can recover from it, it makes a much greater sense for that cable company to partner with someone like Google or Netflix or Facebook, share the cost and then give preferential bandwidth to these providers. That way it is a win win situation.
The best example to illustrate the point is that of India. A desperately poor and rural country. In 2001, Indian government came up with the law that any telecom company that wants to provide telecom service in any city/town must also provide coverage in the rest of the "district". Consequence ? Other than handful large cities telecom companies failed to to expand. They were not interested in bidding for any telecom licenses.
When government abolished that clause, it turned out to be a telecom revolution in India.
The fact is that there are costs and benefits to living in a rural area. Rent is cheaper. Land is abundant. Infrastructure is more expensive. Similarly for city folk, land is expensive, rent is high, but economies of scale make fast internet feasible.
There is no compelling reason that people in cities should be subsidizing rural people, apart from rural congresscritters attempting to buy votes. And if there were such a reason, lets do it honestly: raise taxes in cities and funnel the money directly to rural areas.
If no private company want to lay down high speed internet you have to reavalute if you prefer to live rually or if you want top high speed internet.
ATT (SBC at the time) was pouring billions into infrastructure that they were required to lease out at tariffed rates to competitors. They weren't running new copper, but they were actively building out remote terminals, and rerouting the existing copper to shorter paths. Example press release from 1999 http://www.internetnews.com/xSP/article.php/220301
<y understanding is the wholesale price is to be set 'at cost, plus a reasonable profit,' not simply at cost. However, regulatory action was pretty week here: in the 2003-2004 time frame, SBC/ATT was routinely charging more for wholesale access than for retail customers, even though providing wholesale access is less expensive (simplified billing, collection, support, reduced/zero marketing, no IP transit, etc).
Incidentally, in the past month, ATT has started allowing reselling of U-Verse (although it's at Layer 3, instead of Layer 2 with DSL, so IP transport is provided by ATT instead of the reseller); so, they must have found it to be economically viable to wholesale, in the absence of regulatory compulsion. (I don't find a wholesale Layer 3 product to be as compelling as a wholesale Layer 2 product though; ATT is fine for last mile, but I don't want their peering and transit)
In a free market, rural availability will be priced appropriately and then everyone can make a decision where they want to live without expecting a subsidy from people who made a different choice.
Cities are greener, and more economically efficient. I see no problem with making living in the city slightly cheaper, and living in the countryside slightly more expensive.
It is a civilizational norm that people settle along commerce routes, rivers, coasts, highways. Now, it's starting to make sense to adjust for the flow of information.
The networks becomes more valuable in a non-linear manner vis a vis coverage. The network is worth more to the government and to the users when it nears 100pc coverage. This is why rural telco is subsidized.
The cost of <not> having 100 pc coverage would be an externality borne by the state. And so the economic ROI of teh incremenatal coverage has nothing to do with free market economics at that stage. Optimizing the cost benefit is strictly a function of fiat decisionmaking.
How so? The cost would be primarily borne by people who choose to live in the sparsely populated areas. Often because they want to get away from others so it's not even really a cost.
When it comes to just mere connectivity (for emergency communication and the like), satellite coverage is more than enough and cheaper to provide in unpopulated regions. Plus, Motorola already sunk a ton of money (in Iridium). Now you want to do it again?
Also, a perspective where the government owns the network(?), and maximizes its value is pretty bizarre to begin with. Maybe it would make sense under some sort of hereditary monarchy. Otherwise, all the incentives are completely out of whack. The government will just go with whatever is beneficial to the backers of the coalition currently in power, like the fed and fcc already seem to be doing.
So the telco has to evalute, what the investment cost and how many subscriptions they can sell. If they set it to high, not enought people will buy it. So the market is esentially against how high people value your product.
Also how is the cost cost covered by the state? It is not generally true that the internet is better with more people, specially people who dont value internet highly.
Also your argument can be made for anything, for example, if everybody loved product X then marginal cost of production goes down and that is bettwr for everybody. So would you want the state to enlarge the market for product X, just because it would be better for everybody?
Also its a question of cost, for example, i prefer if there are more people in the internet, however if you ask me how much im willing to pay for otber people getting internet then my answer is pretty much 0.
You are completly ignoring the cost aspect of the cost benefit calculation.
False. In a free market, non-profits, co-ops and other parties with lesser ROI requirements than publicly listed companies can and may build and/or provide network infrastructure and/or service.
For concrete examples look at local fiber network and WISP build-outs in Europe, particularly in Scandinavia.
> The "competitive tide" under unbundling is a fiction.
False. Service level competition with unbundling is very real and creates a quantifiable consumer surplus even without intermodal competition.
For concrete examples see any European country with unbundling.
> It was not accompanied by increased investment into the actual wires that ran into peoples' homes.
False. A more accurate statement would be that the unbundling requirements were not in force long enough for competing ISPs to start building their own access networks.
Infrastructure build-outs require scale. Unbundling requirements enable competing ISPs to build up scale before starting on the next investment ladder step, i.e. their own infrastructure.
Concrete example: In France, one explicit reason for Free to start on their fiber network buildout in Paris was to stop paying the unbundling fees to the incumbent and instead use the money to fund their network buildout and improve their margins.
> Indeed, it eliminated any incentive to invest in those wires--why pour billions of dollars into infrastructure you'll have to lease out at cost to your competitors?
False. Unbundling and other infrastructure charges are not at cost. Regulatory schemes for tariffs are: costs plus, retail minus or in some egregious cases retail plus. All these schemes include (sometimes a substantial) profit.
So, why pour billions of dollars into infrastructure you'll have to lease out to your competitors? The answer is, as always, money or more exactly profits.
Once a network buildout has been completed it becomes a sunk cost with the additional downside of maintenance cost. The one single metric determining if this sunk cost turns a profit is take rate, i.e. how many subscribers you can get and subsequently how much it will cost to service each subscriber.
Leasing your infrastructure drives up the take rate. It's basically free money, analogous to airlines selling empty seats to third parties.
> People mention the "unbundling" requirement as if its a panacea, comparing it to what has been done in Europe, but ignore what makes that policy work in places like the U.K.--the company that owns the wires is a sanctioned monopoly, protected from competition, that is guaranteed return on investment through statutory means.
False. Wireline infrastructure is not a monopoly in Europe and certainly not in the UK. Anybody can get a license to deploy wireline infrastructure in the UK and it's not like BT has 100% coverage either. For example KCOM has a de facto monopoly in Hull where BT has no infrastructure.
But it seems apparent to me that what we really need is more publicly owned networks, at the very least on the local level. Let ISPs and other service providers (phone, television, etc) compete for the consumer over open fiber networks in the same way cab companies compete over publicly owned roads. Fund the networks with taxes on network service providers, and provide federal subsidies for rural networks.
"The Federal Trade Commission's Bureau of Competition enforces the nation's antitrust laws, which form the foundation of our free market economy. The antitrust laws promote the interests of consumers; they support unfettered markets and result in lower prices and more choices."
Unfettered markets being the operative words.
The reality is that antitrust law has been used more often to destroy competition that underpriced you rather than taking out evil monopolys. Many free market economist are against antitrust regulation.
Bitstream access is susceptible to the lowest common denominator of access technology and is at the mercy of the infrastructure owners upgrade cycle or lack of it.
As an option to unbundling it is fine.
That said, I think it's just good policy to have the same for startups. A minimum level of protection (bankruptcy protection, as well as a level playing field) for new entrants. I think that on the level of Netflix, they should indeed pay for special agreements. After all -- come on -- Netflix does special peering agreements with Tier 1 ISPs and their traffic accounts for 34% of all traffic to consumers. So basically, I think the general principle is:
REDISTRIBUTE THE ECONOMICS TO ENSURE THE MINIMUM LEVEL THAT MAINTAINS INNOVATION, FREEDOM, BASIC EXPECTATIONS, BUT BEYOND THAT, LET THE MARKET BE FREE (BARRING REGULATIONS TO MINIMIZE AGAINST NEGATIVE EXTERNALITIES).
PS: If you downvote, I'd love for you to actually reply with a reasoned comment.
Given the current situation, the simple solution for peering is to force providers to connect at cost (which is thousands of dollars... not millions like what netflix paid).
I think that, in general, having the government guarantee something and then regulate it should be done only for the "minimum layer", it should have a limit. Above that, there should be a free market where only negative externalities are regulated.
I think this is a bad idea.
Firstly: This sets a hard upper limit on how much a startup can scale until the ISPs start to get to play favorites. Success thus becomes less about market fit and innovation and more about politics and back-room deals.
Secondly: Why should fairness and a level playing field be guaranteed only at small scale? Why should gatekeepers and oligopolies be given more power at scale?
What's nice about free things is that it's a safety net, and also lowers the barrier for new entrants and trying new ideas. But it's only free up to a point. In MY opinion, you shouldn't be sending 34% of the internet's traffic and expect the same treatment as a person who is hosting a small website. At some point, you're setting up SPECIAL agreements for crying out loud, with tier 1 providers, Google fiber, etc. Your cables are in their physical centers, and that's favoritism, that's physical presence which doesn't scale to millions of companies even theoretically. So yeah, at some point the market participants should be free to negotiate their own agreements and prices, or refuse a level of service that exceeds what a "free" tier would get.
I am talking, of course, about the provider side. Consumers already pay for their service, but I personally think that the UN is right about internet access being a "human right" these days everyone should get an unconditional basic income which is enough to pay for, among other things, a basic level of internet service.
According to reports they do.
> They connect to... was it Cogent?
Cogent is one of their IP transit providers, but they have a lot of others as well.
> So, Comcast customers pay Comcast for the connection, and Cogent pays Comcast to deliver the data, but Comcast also wants Netflix to pay them to deliver their data.
I do not believe it's a matter of also. Comcast wants either Cogent or one of Netflix' IP transit providers to pay them or Netflix to pay them directly for any direct connects.
Both do not have to pay. Comcast is fine with either or paying, as long as somebody is paying them.
It's wholly another question if it is reasonable for Comcast to get any payment at all...
http://www.theguardian.com/science/2012/nov/19/better-angels...
A level playingfield? That does not even mean anything.
I know why people want it, and I think generally that the basic idea is not bad. However the idea that we can controll all markets to achive this goal is just not relistic. This kind of thing was tried in "indian socialism" were they would have permits for every market to controll how many firms operated in each market.
All you need is effective banking and low regulatory burden. A safty net is of course nice, but Im not sure I would put it into "low barrier to entry" category.