Say hello to the Ubernet

THE decade-long debate in America over “net neutrality”—the assumption that all internet traffic, no matter its origin or purpose, should be treated equally—is inching towards some form of compromise. Internet service providers (ISPs) like Verizon, AT&T, Comcast and Time Warner Cable are adamant that the internet should remain free of regulations that would bar them from limiting or charging bandwidth-hogging users such as Netflix and YouTube. During the prime-time hours of 6pm to 10pm, these two popular websites for streaming video account for half of all internet traffic in America.

On the other side of the acrimonious debate, open-internet activists have bombarded the Federal Communications Commission (FCC) with demands that the ISPs be prevented from doing anything that would erode the practice of treating all data packets travelling over the internet—whether for text, speech or video—the same for all comers. Without such provisions, they fear, two internets might emerge: a fast lane for content providers willing to pay for first-class service; and a slow lane for everyone else. They argue that, unable to afford the internet’s fast lanes, start ups and other innovators might not be able to compete effectively with established web giants—and that could hobble innovation.

There is a good deal of mythology associated with the internet. One mistaken impression is that net neutrality is a hallowed principle espoused by the internet’s founding fathers. That is definitely not the case. Indeed, Vint Cerf, the most famous of the internet’s original developers, recently told the Washington Post that neither he nor any of his colleagues at the time ever said all packets of data should be treated equally. Nor did they say that ISPs could not charge more for greater usage.

Far from being a fundamental principle, the term “net neutrality” was unheard of until it was coined in 2003 by Tim Wu, a media-law professor at Columbia University. For the first three decades of the internet’s existence, all packets of data travelling over it were equal by their very nature. Being predominantly text along with the occasional graphic, none of the packets required special treatment. By default, all were therefore handled in exactly the same way. And thus, in retrospect, was born the myth of net neutrality.

It was only when internet telephony and streaming video arrived on the scene a decade ago that some packets began to need priority over others—to prevent pictures from freezing or voices from stuttering. Ever since, packets have been given the priority (ie, “quality of service”) they require, depending on the application they serve. Without that, Skype and other VoIP services would not exist. Still less would popular video and audio streaming sites such as Netflix, YouTube, Hulu or iTunes.

None of which has stopped net-neutrality advocates from insisting that ISPs should be reclassified as common carriers (ie, phone companies), so the FCC would have the legal authority (which it currently does not have) to ban any form of discrimination concerning internet traffic. Barack Obama has made it clear that he is “unequivocally committed to net neutrality”, but says the final decision rests with the FCC. In the midst of it all, Tom Wheeler, the FCC chairman (and an Obama appointee), has the unenviable task of trying to reconcile the differences.

Earlier this year, a federal court ruled, not for the first time, that the FCC had overstepped its authority with its Open Internet Order of 2010, which sought to strike a balance between the First Amendment rights of internet users and the need to encourage ISPs to invest in speedier connections. The FCC took the court ruling as an invitation to rewrite its net-neutrality rules, so they would conform with existing communications law.

By suggesting a compromise, however, the hapless Mr Wheeler felt the full wrath of net-neutrality’s supporters. The proposal merely suggested that while ISPs would be banned from blocking or slowing access to individual websites, they could strike deals with content providers to improve their services, provided such deals were deemed “reasonable”.

Nothing radical about that. The broadband behemoths had been making such deals with content providers for ages. Late last year, for instance, Netflix found its popular video streaming service had begun to slow down. Whatever the cause, it handed undisclosed sums of money to both Comcast and Verizon for “direct connections” to their networks. Since then, Netflix’s video-streaming speeds have doubled, and customers are happier than ever. But there can be no question that Comcast and Verizon, two of the largest ISPs in America, had Netflix over a barrel.

These were not isolated incidents. Wired magazine recently provided an intriguing glimpse of what resides under the internet’s bonnet nowadays, compared with what existed a decade or so ago. Then, internet traffic was generated by thousands of individual companies. Now, the bulk of it comes from just 30 or so web giants, including Amazon, Facebook, Google, Netflix and Twitter. Because such web firms move so much data, they have had to do deals with the large ISPs to let them bypass the internet’s backbone and connect directly to the last-mile pipes the ISPs use to link users to the internet.

These direct “peering” connections provide some of the fastest lanes on the internet. The web giants have negotiated further arrangements that allow them to install their own routers inside the ISPs’ premises. These so-called “content delivery servers” (CDSs) help stream video faster than ever.

Like it or not, such “paid prioritisation” deals are a fact of life. No matter what the FCC may ultimately decide, nothing can roll them back. Nor would users, who stream their nightly viewing material from Netflix and the like, want the FCC to do so. By allocating the appropriate quality of service to the data packets that need it most, paid prioritisation has delivered the goods. And far from degrading the internet, direct peering and CDSs, which make such high-speed lanes possible, have helped reduce traffic congestion on the network’s backbone. That has made life easier for more humble users, too.

The FCC’s task now is to craft a set of rules that accepts the internet’s de facto tiered structure; protects the freedom-of-information rights of individual users; introduces greater competition between ISPs to make it harder for them to twist the arms of content providers; and ensures that those who use the web lightly are not forced to subsidise the heavy bandwidth hogs. It would be better if such heady goals could be achieved without incurring the cumbersome machinery of bureaucratic regulation. Reclassifying broadband as a telecommunications service under the Telecommunications Act of 1996 could have all manner of unintended consequences. And far from encouraging innovation, it could just as easily do the opposite.

Whatever the outcome, the issues that lie ahead are going to make today’s squabbles over net neutrality seem trivial. For one thing, the internet is evolving rapidly from a platform for surfing, searching and shopping to become predominantly a medium for entertainment—in particular, a personalised alternative to cable and satellite television.

The past week’s announcement by HBO, American pay-TV’s premier network (which was followed the next day by a similar announcement from CBS), that it intends to make its prized content available over the internet, instead of solely via a set-top box, gives American viewers the most compelling reason yet to ditch their costly cable contracts—and end the nonsense of having to subscribe to hundreds of unwanted television channels to get just the handful they wish to watch.

So far, only 7.6m homes in America—out of the 90m that receive television by cable or satellite—have cut the cord. With other networks now expected to follow HBO and CBS, internet TV would seem to have reached a tipping point. The trickle of cable-cutters could now easily turn into a flood. The demand for bandwidth would then go through the roof—especially as the bulk of the video being streamed would be in high-definition, and soon ultra-high-definition format. Whatever reforms the FCC may muster, the internet is going to be stretched to the limit.

Add the fact that the world is moving inexorably towards ubiquitous connectivity, where internet services become as pervasive as electricity. In a landmark study earlier this year by the Pew Research Centre, a non-partisan think-tank based in Washington, DC, most of the 2,500 or so information experts canvassed expected to see a global, immersive and ambient computing environment emerge over the coming decade. If they are right, people will no longer turn on a device and fire up a browser to log on to the internet; they will be permanently connected, possibly without even realising it. Whether mobile, wearable or embedded, sensors will tie people together through an “internet of things” that will allow them to tap into smart information in the cloud to answer their every query, even anticipate their needs.

This is not science fiction; much of the enabling technology has started to trickle out of the laboratory. The question, then, is whether the internet can survive in anything like its present form? The answer is clearly no.

Five years ago, the Internet Society suspected that that might be the case. At the time, it asked itself two questions. Will the world embrace, or resist, the open-internet model? And what model will be more successful: one that relied on command and control; or one embodying the distributed and decentralised features of today?

This was not a poll for possible answers, but an exercise in scenario planning—a methodology widely used in business for examining alternative futures. The sole aim was to reveal plausible versions of how, for better or worse, the internet might evolve over the next eight to ten years. The society’s two questions defined four possible scenarios, none of which was favoured as being more likely (such is the nature of scenario planning). But the scenario the society fervently hoped would prevail was the decentralised, open-internet model of today.

Five years on, that is looking less and less likely. Indeed, it now seems inevitable that the internet of tomorrow will rely on more top-down command and control than the bottom-up freedom of yesterday. In other words, more tiers of paid prioritisation will have to be added inside the ISPs’ routing centres to handle the coming flood of traffic, as the internet becomes a serious alternative to cable television. Users are unlikely to notice such internal differences, but the external ones will be clear for all to see.

More than likely, people going online could find themselves spending most of their time within the confines of one or two mega-sites. Instead of visiting a multitude of different websites for different things, users could be confronted with a series of “walled gardens” built around app stores and proprietary services that offer everything from streaming video to holidays and household goods. As such, they will satisfy the visitors' every need and whim, save one—the ease of venturing far and wide in the scary wilderness beyond the garden walls. Welcome to the Ubernet.