Sunday, February 8, 2015

What Everyone Gets Wrong in the Debate Over Net Neutrality

What Everyone Gets Wrong in the Debate Over Net Neutrality

Even Sunday night HBO watchers are worried the Federal Communications Commission will soon put an end to net neutrality.
Earlier this month, on the HBO comedy news show “Last Week Tonight,” host John Oliver went on a 13-minute rant against the new set of internet rules proposed by the FCC. He warned that the rules would lead to a world where internet service providers like Comcast and Verizon can sell special treatment to web companies like Google and Netflix, charging extra fees to deliver their online videos and other content at fast speeds, and he urged viewers to bombard the FCC website with protests, saying the rules would end up hurting smaller web outfits that can’t afford to pay the fees. The next day, the FCC site buckled under the traffic and went offline.
It was just part of a sweeping effort to squash the proposed rules. When the rules first leaked out in May, protesters camped out in front of the FCC’s Washington offices. Big tech companies such as Google, Amazon, and Netflix signed a letterasking the government communications agency to bar internet providers from discriminating “both technically and financially against internet companies.” And last week, two big name Democrats on Capitol Hill unveiled a bill that seeks to undermine the new rules. Nearly everyone, it seems, wants to prevent the FCC from allowing some companies to have internet “fast lanes” while others toil at slower speeds.
The only trouble is that, here in the year 2014, complaints about a fast-lane don’t make much sense. Today, privileged companies—including Google, Facebook, and Netflix—already benefit from what are essentially internet fast lanes, and this has been the case for years. Such web giants—and others—now have direct connections to big ISPs like Comcast and Verizon, and they run dedicated computer servers deep inside these ISPs. In technical lingo, these are known as “peering connections” and “content delivery servers,” and they’re a vital part of the way the internet works.
“Fast lane is how the internet is built today,” says Craig Labovitz, who, as the CEO of DeepField Networks, an outfit whose sole mission is to track how companies build internet infrastructure, probably knows more about the design of the modern internet than anyone else. And many other internet experts agree with him. “The net neutrality debate has got many facets to it, and most of the points of the debate are artificial, distracting, and based on an incorrect mental model on how the internet works,” says Dave Taht, a developer of open-source networking software.
The concepts driving today’s net neutrality debate caught on because the internet used to operate differently—and because they were easy for consumers to understand. In many respects, these concepts were vitally important to the evolution of the internet over the past decades. But in today’s world, they don’t address the real issue with the country’s ISPs, and if we spend too much time worried about fast lanes, we could hurt the net’s progress rather than help it.
Even Tim Wu, the man who coined the term net neutrality, will tell you that the fast lane idea isn’t what it seems. “The fast lane is not a literal truth,” he says. “But it’s a sense that you should have a fair shot.” On the modern internet, as Wu indicates, the real issue is that such a small number of internet service providers now control the pipes that reach out to U.S. consumers—and that number is getting even smaller, with Comcast looking to acquire Time Warner, one of its biggest rivals. The real issue is that the Comcasts and Verizons are becoming too big and too powerful. Because every web company has no choice but to go through these ISPs, the Comcasts and the Verizons may eventually have too much freedom to decide how much companies must pay for fast speeds.
Click image to enlarge.  WIRED
We shouldn’t waste so much breath on the idea of keeping the network completely neutral. It isn’t neutral now. What we should really be doing is looking for ways we can increase competition among ISPs—ways we can prevent the Comcasts and the AT&Ts from gaining so much power that they can completely control the market for internet bandwidth. Sure, we don’t want ISPs blocking certain types of traffic. And we don’t want them delivering their own stuff at 10 gigabits per second and everyone else’s stuff at 1 gigabit. But competition is also the best way to stop these types of extreme behavior.
Though the network will never be neutral, we can find ways of promoting a vibrant market for fast internet speeds that’s open to everyone. At the end of his rant, John Oliver actually comes pretty close to the real issue. Advocates, he says, “should not be talking about protecting net neutrality. They shouldn’t even use that phrase. They should call it preventing cable company f***ery, because that is what it is.”

The Great Rewiring

The net neutrality debate is based on a mental model of the internet that hasn’t been accurate for more than a decade. We tend to think of the internet as a massive public network that everyone connects to in exactly the same way. We envision data traveling from Google and Yahoo and Uber and every other online company into a massive internet backbone, before moving to a vast array of ISPs that then shuttle it into our homes. That could be a neutral network, but it’s not today’s internet. It couldn’t be. Too much of the traffic is now coming from just a handful of companies.
Craig Labowitz made this point last month, when he testified before a Congressional committee on the proposed Comcast-Time Warner merger. Ten years ago, internet traffic was “broadly distributed across thousands of companies,” Labovitz said in his prepared statement to the committee. But by 2009, half of all internet traffic originated in less than 150 large content and content-distribution companies, and today, half of the internet’s traffic comes fromjust 30 outfits, including Google, Facebook, and Netflix.
Because these companies are moving so much traffic on their own, they’ve been forced to make special arrangements with the country’s internet service providers that can facilitate the delivery of their sites and applications. Basically, they’re bypassing the internet backbone, plugging straight into the ISPs. Today, a typical webpage request can involve dozens of back-and-forth communications between the browser and the web server, and even though internet packets move at the speed of light, all of that chatter can noticeably slow things down. But by getting inside the ISPs, the big web companies can significantly cut back on the delay. Over the last six years, they’ve essentially rewired the internet.
Google was the first. As it expanded its online operation to a network of private data centers across the globe, the web giant also set up routers inside many of the same data centers used by big-name ISPs so that traffic could move more directly from Google’s data centers to web surfers. This type of direct connection is called “peering.” Plus, the company set up servers inside many ISPs so that it could more quickly deliver popular YouTube videos, webpages, and images. This is called a “content delivery network,” or CDN (see glossary, right).
“Transit network providers” such as Level 3 already provide direct peering connections that anyone can use. And companies such as Akamai and Cloudflare have long operated CDNs that are available to anyone. But Google made such arrangements just for its own stuff, and others are following suit. Netflix and Facebook have built their own CDNs, and according to reports, Apple is building one too.

The Google Edge

Does this give companies like Google and Netflix a potential advantage over the next internet startup? Sure it does. But this isn’t necessarily a bad thing. In fact, this rewiring has been great for consumers. It has allowed millions to enjoy House of Cards, YouTube, and Kai the hatchet-wielding hitchhiker. It’s the reason why the latest version of high-definition video, Ultra HD 4K, is available for streaming over the internet and not on some new disk format.
Plus, although Google does have an edge over others, not every company needs that edge. Most companies don’t generate enough traffic to warrant a dedicated peering connection or CDN. And if the next internet startup does get big enough, it too can arrange for a Google-like setup. Building the extra infrastructure is expensive, but making the right arrangements with a Comcast or a Verizon is pretty cheap—at least for now.
Traditionally, ISPs have not charged for interconnection points. They’re happy to have Google or Netflix or Akamai or Level 3 servers or routers in their data centers because they speed up service for their customers and reduce the amount of traffic that has to flow out of their network. You could look at these arrangements as fast lanes—but because ISPs have treated their networks as an open marketplace and delivered real value to consumers and businesses, they’re not the kind of thing that the FCC should be discouraging.
The problem today isn’t the fast lanes. The problem is whether the ISPs will grow so large that they have undue control over the market for fast speeds—whether they can independently decide who gets access to what connection at what price. “The question is which kinds of fast lanes are problematic and which kinds are not,” says Marvin Ammori, a lawyer and net neutrality advocate.

Muscle Gets Flexed

Ammori is worried that, recently, internet service providers—Comcast, in particular—have been started to flex their muscle. Transit provider Level 3 says that, as it sends traffic into their networks, the big U.S. ISPs are letting some of its routers overload with data. And amidst these bottlenecks, Comcast is exploring ways of selling its own CDN services that can help companies increase delivery speeds.
For Ammori and others, this seems like a shake-down that lets the service providers get paid at at both ends—by their home subscribers and by the web companies that deliver stuff to these subscribers. Ammori worries that the ISPs will start throwing their weight around unfairly. “We don’t want AT&T and others to impose a tax and to treat those who pay a tax better than others,” he says. The strange thing is that even some of the biggest “net neutrality” advocates downplay the importance of these peering deals, saying they have nothing to do with net neutrality. But this is largely an argument of semantics. The point is that, whatever terms you use to describe it, the situation could lead to an unfair playing field.
One way to prevent this is through greater competition among ISPs. If consumers and web companies have many ISPs to choose from, no one ISP can control who gets what. Ammori doesn’t see more competition among ISPs as a panacea, but he thinks it would help. And another network activist, Seth Johnson, believes competition can change things, but that it will take government action to make that happen. “Competition and regulation are not at odds in every case. Particularly in telecom,” he says. In any event, competition is a bigger issue than net neutrality. The internet has evolved, but the debate must evolve along with it.
If Comcast’s last-mile of cable connection was available to all competitors under the same terms that gave dial-up service providers access to all copper telephone networks back in the 1990s, we would have more ISPs in more geographical areas. Consumers could simply switch providers whenever Netflix or YouTube started to get choppy. And that would give Netflix and YouTube more leverage in their deals with the ISPs. At the moment, this option—where ISPs are treated as “common carriers”—is on the table, but it seems like a remote possibility. Maybe it shouldn’t be. Instead of railing against fast lanes, we should be pushing Washington to explore ideas like this that could actually promote competition among ISPs. “In the present situation,” Johnson says, “the debate is misdirected.”

No comments:

Post a Comment