Charging for peak-load usage, or congestion pricing, might take care of these contention issues. If the cost to deliver a bit to a particular house during key evening download hours were higher, users would probably change their behavior. But that would involve tinkering with all subscribers’ bills, not just the hogs’, and—to be cynical—might not discourage people from subscribing to the online video services that could cause them to exceed the network operator's cap. Providers are not interested in this solution.
Because the United States has given up on rate regulation for high-speed Internet access services, and the reporting requirements that go with it (number of subscribers, revenues, costs, service outages, quality of service), regulators have no reliable data about how pricing is computed.24 No regulator seems to know what it costs to deliver the extra gigabyte the operators want subscribers to pay for. The actual cost of delivering bits over the last mile is probably pretty low; according to Netflix, an Internet service provider's cost to “deliver a marginal gigabyte, which is about an hour of viewing, from one of our regional interchange points over their last mile wired network to the consumer is less than a penny, and falling, so there is no reason that pay-per-gigabyte is economically necessary.”25 But overage charges (per-gigabyte charges imposed once the user has exceeded the network operator's cap) can be two dollars per gigabyte or more; Canadian ISPs have been known to charge five.26
In the absence of concrete information, regulators are stuck: carriers claim they need to charge overages, and the government—needing to encourage the building of infrastructure by these private parties—has no choice but to agree. And given the concentration in the marketplace for network operators, users have no choice but to pay.
According to the FCC, network operators in the past have routinely advertised “up to” speeds that are twice as high as the speeds subscribers actually experience; part of the reason for this phenomenon may be the prevalence of shared (contended-for) connections. Cable operators, in particular, routinely oversell their services.27 If everyone is downloading a movie between 6 P.M. and 9 P.M., those shared networks burden the bitrates that everyone gets. The same amounts of bits go through the pipes, but they go more slowly. Between six and nine, you're battling your neighbor for bandwidth.
As a result, online video distributors face the prospect of being squeezed out: users won't sign up for independent online video if they believe they will end up paying more for Internet connectivity as a result.
To see how this might play out, consider our frozen neighbor to the north, where usage-based billing has been a major consumer issue. In 2010, Bell Canada convinced the Canadian telecommunications regulator, the Canadian Radio-television and Telecommunications Commission (CRTC), to approve a rate structure that it could impose on buyers of its wholesale data services—entities that planned to resell Internet access to their retail customers. The wholesale offering would have a number of options, ranging from a “Lite” rate of up to two gigabytes per month, with a $1.87 surcharge for every gigabyte over the cap, to a “Basic” plan of up to sixty gigabytes per month with a $1.12 overage fee.28 Those are pretty meager usage rates before the surcharges kick in: by streaming video you could use up the Basic monthly allotment within six hours.29
Bell Canada argued that it had to impose caps to deal with fast-rising usage of video that had caused a 25 percent uptick in the volume of data carried over its networks.30 But the real targets were Netflix and other online video providers. Netflix had launched services in Canada in 2010. Rather than build out its networks to allow consumers to watch video more readily, Bell and the other Canadian network providers had decided to enforce scarcity.31
Bell's technology, though, seemed incapable of accurately measuring how much subscribers were using, and the resulting overcharges and undercharges caused a furor. More than five hundred thousand people signed an online petition to the Canadian government demanding an end to usage-based billing.32 Both Michael Geist, Canada Research Chair of Internet and E-commerce Law at the University of Ottawa, and Cory Doctorow, a Canadian-British futurist and author, pointed out that Bell had conceded that the rates it was charging for overages had nothing to do with the actual costs of providing services. The fees, instead, were designed to constrain users’ behavior by making it unattractive for them to do things that required a lot of data. “In other words,” Doctorow wrote, Bell had “set out to limit the growth of networked based business and new kinds of services, and to prevent Canadians experimentation that enables them to use the Internet to its fullest.”33
Government officials called for an end to the practice and asked the Canadian regulator to reverse a ruling that would have forced retail ISPs using wholesale services from the large Canadian incumbents to adopt usage-based billing.34 In March 2011, Bell Canada backed down from applying mandatory caps to independent retail ISPs buying its wholesale services, but it continued to charge hefty overages to its own subscribers who exceeded its bandwidth caps.35
Netflix quickly sent its own political signal, declaring that it would automatically compress its Canadian online video services into a third as much data in order to avoid the Canadian caps. Thirty hours of streaming film or television typically uses 31 GB of data; under this default compression setting, only 9 GB would be coming across users’ wires. But this resulted in measurably reduced video quality.36
According to Geist, usage-based billing only helps the large Canadian telecom companies squeeze out their small competitors and squelch innovation. “The effect extends far beyond consumers paying more for Internet access,” he has said. “There is a real negative effect on the Canadian digital economy, harming innovation and keeping new business models out of the country. Canada is not competitive when compared to most other countries and the strict bandwidth caps make us less attractive for new businesses and stifle innovative services.”37 Jesse Brown of the news Web site Macleans.ca wrote, “Yes, that's innovation in Canada: new players can indeed compete, by grossly degrading their product to a level beneath anything they'd dare offer to Americans. True, it may be a publicity stunt on Netflix's part, a calculated move to embarrass Canada into getting with the times. If so, it's a brilliant one.”38 Netflix, for its part, argued that Bell Canada's wholesale usage-based billing rates provided margins in excess of 99 percent.39 In late 2011, Netflix's arguments carried the day: the Canadian regulator required large providers like Bell Canada to sell access to independent ISPs at set rates based on actual costs plus a reasonable profit, and prohibited the incumbent from limiting the plans that the independents could sell to their customers.40
American network providers would like to charge for consumption of bandwidth based on rates that they choose, but following a disastrous attempt by Time Warner Cable to impose a cap of 75 GB in 2009, the American companies have been cannier about implementing usage-based billing. Protests over Time Warner Cable's proposed cap forced the company to reverse its policy, prompting Senator Charles Schumer (D-N.Y.) to issue a proud press release: “By responding to public outrage and opposition from community and elected officials, Time Warner Cable made the right decision today.”41 The message to the rest of the providers: be extremely careful about how you set your limits.
But it is not clear that U.S. regulators will follow Canada's lead in this area. A major development was the FCC's agreement in late 2010 to let network providers use their discretion in charging additional fees for exceeding caps that the providers themselves would establish. Julius Genachowski, the FCC chairman, had warm words of support for usage-based pricing just weeks before the Comcast-NBCU merger was approved: “Our work has also demonstrated the importance of business innovation to promote network investment and efficient use of networks, including measures to match price to cost such as usage-based pricing.”42 The rumor in Washington was that the FCC had promised to praise usage-based pricing in order to garner AT&T's support for its “open Internet” rules in December.43 Comcast was then required as part of the merger n
ot to treat “affiliated” broadband network traffic differently from unaffiliated traffic when it was implementing caps, tiers, metering, “or other usage-based pricing.”44 But Comcast could get around this limitation by calling its own Video on Demand services something other than “broadband”—again, “specialized services” are not subject to neutrality obligations, and the FCC's authority to say anything about neutrality in the first place is tenuous.
Following the approval of the merger, in March 2011 AT&T announced caps for both its DSL and U-verse fiber-to-the-neighborhood services. The caps would kick in at 150 GB per month for the DSL service, and at 250 GB per month for U-verse. Users exceeding the caps would be charged ten dollars for each additional 50 GB of use. AT&T already had in place a dramatically low cap for its wireless services: new mobile consumers can no longer access unlimited data and have to choose between a 0.2-GB-per-month plan and a 2-GB-per-month plan.45
Given Americans’ appetite for television, the AT&T caps seemed to be designed to discourage the substitution of online video for traditional television; anyone watching more than two high-definition movies a day would be subject to the wired cap.
Most of the other high-speed Internet wired access service providers were expected to follow suit. Comcast has had a cap of 250 GB in place since 2008, terminating users who consume more data than that per month, and plans to raise this to 30 GB and charge for additional data transfer.46 And the iPad mobile plans were clearly focused on limitations: 2 GB for a flat fee, and then ten dollars per gigabyte after that.47 In short, American distributors have learned from their own early mistakes and, aided by the permissive statements of regulators, have brought usage-based limits to the market without consumers even noticing, let alone protesting. Usage caps have allowed American carriers to impose scarcity (so that average revenues can continue to increase), while other countries have focused instead on providing abundant bandwidth for new ideas and new ways of making a living.
Raising the cost of Netflix's access to programming, as described in Chapter 5, was one way the cable complex could keep Netflix in a box. Another would be to make it more expensive for Netflix's users to stream its content, and imposing caps would do the job: high-definition streaming movies on Netflix, at five gigabytes each, would almost certainly become a luxury if usage-based pricing became the norm in America. Usage-based pricing would be a useful tool for cable distributors, giving them an opportunity to leverage control over their giant IP-enabled pipe to discriminate against competing offerings, using their position as the only conduit for content companies to reach their subscribers.
Usage-based billing would not apply to the portion of the pipe that the cable distributors could label specialized services, such as their own proprietary video or gaming services—data using the Internet Protocol that the network operator could prioritize—but Netflix would be subject to the caps imposed by the network operators if it remained a pure online video distributor. Comcast could use its control over access to the last mile to collect a toll that would not be collected from its own content services.
Consider how powerful Comcast's position is: it can label TV Everywhere a specialized service or say that it has to prioritize its own video for reasons of “reasonable network management.” The FCC has said it will consider on a case-by-case basis whether “specialized” video services offered by cable distributors are an end run around its light-touch common-carriage (net neutrality) rules; under these circumstances, what investor would take a risk on a new online video-distribution company?48 David Cohen, Comcast's executive vice president, has argued in the past both that without reasonable network management “these networks collapse” and that it is very difficult to put into words just what reasonable network management means.49 Welcome to the land of uncertainty, where new online businesses go to die.
What's more, Comcast already has the hammer in place. The company installed meters for all its subscribers that can be triggered if it decides to switch to finer-grained usage-based billing for high-speed Internet access. And although it swears that it has no immediate plans to start billing separately for each service used over its Internet access connection, Netflix is not missing the signals. Comcast can implement more stringent usage-based pricing—which might dissuade more people from signing up for Netflix—at any time. At the same time, it can expand the portion of its pipes used for Internet-like interactive services—“specialized services”—over which it has unquestioned control.50 Comcast will have the power to ensure that users reach only online sites with which Comcast has some form of relationship.
Netflix is thus engaged in a game of chicken with Comcast. It is banking on the fact that users who take a while to reach Comcast's caps will stay loyal to Netflix until this happens, and that by the time the caps began to bite, users will have started protesting. Not that this will happen immediately: Netflix's own 2011 report showed that the average user streamed Netflix at just over 2 Mbps—much slower than the 50 Mbps or 150 Mbps of which Comcast is capable.51 When it came to the last mile, Americans are just starting to buy the very highest speeds in large numbers.52 Netflix's hope must be that in time—when the inevitable showdown comes—Comcast will need Netflix more than Netflix needs Comcast, and that Netflix's loyal subscriber base will give the company protection from Comcast's high charges.
But the threat remains. Cable's response in this game of chicken means that investors in Netflix (or future purchasers of the entire company) may be discouraged by the deeply contingent nature of Netflix's plans. Usage-based billing poses risks to all kinds of new online businesses. As Stacey Higginbotham of GigaOM puts it: “In the broadband arena where there is little competition among providers and a tendency to avoid investment in networks because of pressure from Wall Street and … a lack of competition, usage-based pricing could lead to expensive broadband and stifle burgeoning technologies such as online video and HD video conferencing.”53
In the end, the FCC's “specialized services” category creates a business development opportunity for operators like Comcast. Comcast can say to a company like Facebook—the ESPN of the Internet—that its traffic won't be subject to a cap or the limits of best-efforts transmission because it will be treated as part of a specialized service. Facebook, in turn, will have to share revenue with Comcast or pay a premium in exchange for this treatment. (Or, perhaps, because Comcast needs Facebook, Facebook will be able to get this categorization for free.) A cable-distribution company that “rates” Facebook sessions in this way could effectively avoid the threat of the Internet. Kids these days don't use e-mail or phone calls, but they do use Facebook, which provides them with an AOL-like walled garden of media content and intimate interactions with friends. To avoid being hit by high overage charges, customers might choose to avoid buying Internet access altogether or limit themselves to packages that allow for limited Web browsing. After all, they'll still have Facebook.
Even if usage-based billing did not slow the arrival of competitive online video services right away, Netflix faces other squeezes from last-mile network providers. Because other networks have no other means of reaching Comcast subscribers, Comcast can charge any networks Netflix signs up that try to send traffic to Comcast's gateways—let's call them “connecting networks.” And that would raise Netflix's costs.
Traditionally, the big pipes crisscrossing the country “peered” (or connected) with one another at interconnection points and swapped traffic for free. The major actors were AT&T, Sprint, and Verizon, and while they sold wholesale traffic carriage to smaller companies, they traded traffic at no cost among themselves. They also used a technique called hot-potato routing, in which data traffic was handed over to other carriers at the point closest to the point of origination.54 This may sound technical, but it's not: if you are a pipe provider with network interconnection points in Los Angeles, New York, and Miami and a peer pipe, and your customers in Pasadena are sending out traffic that is destined for that peer's network subscribers, hot-pot
ato routing would have you handing over that traffic in Los Angeles.
In that traditional environment, if you started to hot potato a lot more traffic to the other network, and the other network was carrying your traffic from Los Angeles to its destination in New York City, the other network might start arguing that you should pay if you were sending, say, twice as much traffic to them as they were sending to you. Your relationship with that network might switch from a free exchange of traffic to a paid exchange.
The problem for these traditional free peering relationships is that, unlike the days when most network usage consisted of telephone conversations, consumers today are no longer sending and receiving the same amount of traffic. Instead, with so much more video online, they are receiving a lot more traffic than they are sending. Comcast can argue that the traffic with connecting networks is out of balance—they are getting a lot of traffic, but they are not sending as much—and so the connecting networks should pay Comcast.
But connecting networks—pure-play, no-retail pipe providers like the company Level 3—argue that the hot-potato aspect of the traditional equation no longer applies. Rather than drop off traffic to Comcast at the point where it originates, they are portaging traffic as close to the relevant consumers as possible—taking it to the New York metro area, in our example, instead of handing it off in Los Angeles—and so the connecting networks should not have to pay Comcast.55 In fact, Comcast should pay the connecting networks for bringing so much content that Comcast's subscribers want almost the entire way. And by the way, the connecting networks argue, Comcast's outgoing traffic has no way of reaching subscribers of Verizon and Qwest without going through them—another reason Comcast should pay them, rather than the other way around.
Captive Audience: The Telecom Industry and Monopoly Power in the New Gilded Age Page 22
