When Netflix agreed to pay Comcast to stream its movies to the Internet service provider's customers without infuriating delays, it essentially yielded to monopolistic extortion. Comcast, which has just agreed to acquire Time Warner Cable, will soon be the biggest ISP in the U.S., and stuttering video transmission to its clients is not an option for Netflix. The outcome, however, is good for consumers: They will not be charged for what they are not buying.

The Internet is a system of business relationships that the Netflix-Comcast deal is helping to revolutionize. In the original business model, customers pay for access to data "pipes" that connect them to the rest of the world. They pay their home internet service provider, such as Comcast, which, in turn, pays so called Tier 1 networks such as AT & T and Verizon for transit. (Of course, end customers can also work directly with Tier 1 providers).This model provides pure "net neutrality": It does not matter what kind of traffic is carried up and down the pipes.

ISPs undermined their business model in the late 1990's – early 2000s when they started charging flat fees for their services. They couldn't imagine how fast Internet traffic would grow. According to Cisco Systems, in 1997 the Internet carried just 100 gigabytes per hour, but by 2007 that increased to 2,000 gigabytes per second. In 2010, four large European telecoms commissioned AT Kearney to assess the viability of their Internet business models. The consultants' found that "the increased costs of handling rapidly growing traffic are not matched by additional revenues or a key segment of the value chain: those who operate the networks."

One can sometimes hear claims that broadband providers enjoy 97 percent profit margins. They would only be true if telecom infrastructure were free. As it is, ISP's have some of the lowers returns on invested capital in the tech industry. Comcast's average ROIC over the past five years is just 4.5 percent.

AT Kearney suggested that the ISPs either go back to price differentiation, charging their clients according to traffic or service quality, or start charging content providers for seamless access to clients.

The dilemma is clear: Someone has to cough up funding for the ISPs to maintain and improve infrastructure, and that someone can only be the end user or the content provider.

According to Cisco Systems, consumer Internet video traffic made up 57 percent of all consumer traffic in 2012 and will reach a 69 percent share by 2017. Video dominates Internet traffic today because it is so "heavy" and in such high demand. With its 30 million subscribers, Netflix generates much of that traffic. YouTube is another traffic hog. From time to time, users suspect their ISPs of discriminating against such services, causing customers to look at spinning wheels to cut their traffic. Last year, the French telecommunications regulator even examined how a local provider, Free, routed YouTube traffic, but found no evidence of discriminatory practices.

ISPs have the technical capacity to handle YouTube and Netflix traffic, but that's because they paid for it. They could shovel off the costs onto the consumers by charging per megabit, as mobile operators do. In the U.S., they could even get away with it. According to the Federal Communications Commission's December, 2012 data, the most recent available, only 34 percent U.S. households were located in areas where three or more providers offered speeds of at least 6 megabits per second, the recommended speed for seamless Netflix. 64 percent were confined to one or two providers.

If the ISPs did that, consumers would be faced with a choice of several data plans and, as usual, unsure which one they needed. Many would end up paying for a more expensive plan even though they do not subscribe to Netflix or, in fact, often watch any kind of video on the Web.

Alternatively, the ISPs could charge their clients for premium access to Netflix and whatever other services they deemed to be big bandwidth hogs. That would force customers to think in advance about what content they were going to use, and few people give a lot of thought to that. Again, customers would often end up paying for unneeded services.

Or else the ISPs could get Netflix and other content providers to pay for the burden they place on the infrastructure. That way, Netflix can transfer the increased cost to its own customers, and people would only pay more if they wanted a specific service.

From a consumer's point of view, the third option is clearly preferable, though Comcast did use its near-monopolistic market power to break down Netflix's resistance.

If the physical Internet's business model shifts toward collecting payments from bandwidth-hogging content providers, they will have to negotiate deals with lots of ISPs in every country of operation. That will give some some ISPs a competitive advantage over the others. It may also hurt content companies unable to make such deals, giving established players an extra edge over new competition and raising entry barriers. That, however, could finally force a technological breakthrough in video compression: both the upstarts and giants like Netflix hoping to cut their costs will want to invest in making video "lighter".

Whether we want it or not, Internet infrastructure is run for profit as a business. There is no such thing as a 100 percent fair business model. Some of them, however, have more advantages than disadvantages for the parties involved, and the Netflix-Comcast deal helps establish one of these.

To contact the author on this story:
Leonid Bershidsky at lbershidsky@bloomberg.net