A Bear's Face on Mars Blake Lively's New Role Recognizing a Stroke Data Privacy Day Easy Chocolate Cake Recipe Peacock Discount Dead Space Remake Mental Health Exercises
Want CNET to notify you of price drops and the latest stories?
No, thank you

The strange resurrection of Net neutrality

At the annual meeting of the Congressional Internet Caucus, the only issue that appeared likely to appear soon on lawmakers' agenda was an old one -- Net neutrality.

WASHINGTON, D.C.--At this week's State of the Net conference, an annual event of the bipartisan Congressional Internet Caucus, members of Congress, staffers, and technology policy junkies gathered once again to explore the government's Internet-related priorities for the new year.

A few themes emerged, including possible legislation over cybersecurity, a rewrite of the 1996 Communications Act, reforming federal electronic-surveillance laws, and the continuing threat of both national governments and the United Nations trying to wrest control of Internet governance from engineering-driven groups.

The general consensus, however, was that for at least the next several months, the fiscal cliff, debt ceilings, and budget sequestrations were likely to keep Washington fully occupied, leaving little time for legislators to tinker, for better or worse, with the Internet.

One topic that could get renewed Congressional attention, however, is Net neutrality. In a case pending for nearly two years at the U.S. Court of Appeals for the D.C. Circuit, Verizon and MetroPCS are challenging the legality of the Federal Communications Commission's 2010 "Open Internet" order, more commonly known as the Net neutrality rules, which prohibits ISPs from blocking lawful Web content. Should the FCC lose its case, the fight could return to Congress.

So far, Congress has resisted a decade's worth of calls to pass legislation establishing a role for the FCC in governing Internet access or content. But in a keynote address on Tuesday, U.S. Rep. Anna Eshoo (D-Calif.) argued that a key House subcommittee should focus on user protections.

"First and foremost," Eshoo said, "this means preserving the basic 'rules of the road' that the FCC adopted to ensure a free and open Internet. Should the court overturn the FCC's rules, I will be prepared to introduce legislation clarifying the Commission's authority to ensure a free and open Internet, while preventing the use of Internet 'fast lanes' or other discriminatory tools."

Eshoo, whose district includes Silicon Valley, also said she hoped the subcommittee would take up issues related to Internet video, including the impact on companies like Netflix of tiered pricing arrangements that could "deny consumers the freedom and flexibility to stream video content whenever and wherever they want."

Netflix's "jujitsu"
Eshoo's comments may prove unintentionally ironic. Just a few weeks ago, at the Consumer Electronics Show, Netflix announced that Cablevision Systems had agreed to join Open Connect, a proprietary content delivery network (CDN) that Netflix has been deploying largely in Europe for the last year.

CDNs place cached servers at strategic locations along an ISP's network, speeding up the delivery of high-bandwidth content, such as high-definition video, which requires minimal latency to maintain its quality. Given that Netflix traffic now accounts for as much as 30 percent of total downstream Internet activity at peak times, the company has been understandably focused on ensuring it can serve more customers with higher quality.

But at the most basic level, CDNs are nothing if not "fast lanes" on the Internet, delivering information from content providers who pay a premium price to use them.

And at the same time, Netflix also announced that exclusive new "Super HD" and 3D content would only be available to Netflix customers whose ISPs signed on to Open Connect.

Time Warner Cable quickly complained that by withholding these new services from its customers until the company qualified for Open Connect, Netflix was using its "fast lane" to discriminate against some Internet users -- indeed, against some of Netflix's own customers.

"We believe it is wrong for Netflix to withhold any content formats from our subscribers and the subscribers of many other ISPs," the company said in a statement to Multichannel News. "Time Warner Cable's network is more than capable of delivering this content to Netflix subscribers today."

Other critics pounced on Netflix as having committed "Net neutrality jujitsu." Using a loophole in the FCC's rules -- which restrict only the business practices of ISPs, not content providers -- Netflix was perpetuating precisely the kind of consumer harm the rules were intended to prohibit.

To paraphrase Eshoo, it appeared the company was itself denying customers "the freedom and flexibility to stream video content whenever and wherever they want."

A genuine Net neutrality problem had at last emerged, or so it seemed. And Netflix wasn't the company needing the FCC's protection after all. It was the company committing the violation.

Whatever is happening, it's not a violation
The flood of commentary already generated by Time Warner Cable's objections, if nothing else, highlights once again how unhelpful the squishy term "Net neutrality" really is in discussions of Internet policy.

Does it refer to a largely theoretical view of network architecture, in which every packet is treated identically regardless of its origin or the nature of its content -- data, voice, or video? Is it a set of guarantees for consumers that they will not be constrained in any way from accessing the lawful content of their choice?

Or, as implemented in the FCC's rules (which never actually use the term "Net neutrality") is it a set of protections for content providers to ensure they are not unfairly disadvantaged by ISPs that may also offer competing content, including television programming and on-demand movies?

Let's clear away some of the debris. For one thing, it's clear that however Netflix is deploying its Open Connect network, the company is not committing any actual violation of the FCC's rules. For one thing, the rules apply only to ISPs. Even if, as it now appears, some content providers are creating their own networks and have become powerful enough to dictate terms to ISPs rather than the other way around, the FCC does not prohibit discrimination by content providers and likely lacks the authority to do so even if the agency wanted to.

More to the point, even though CDNs create "fast lanes" that speed the delivery of content for those providers who pay premium prices, they are not prohibited no matter who offers them. In its final order (PDF), the FCC explicitly exempted CDNs from the Open Internet rules, noting that, "the record does not demonstrate that the use of CDNs has any material adverse effect on broadband end users' experience of traffic that is not delivered via a CDN."

Indeed, CDNs, along with paid peering arrangements, VPNs, mobile traffic backhaul, and nearly a dozen other forms of non-neutral traffic management were all excluded from the final rules (PDF), leading many Net neutrality advocates to complain that the Open Internet order had failed to deliver on the promise of a truly neutral Internet.

But network management techniques that prioritize some packets over others have been part of the Internet for over a decade. Like the Netflix Open Connect network, many of these techniques have been driven by innovations in engineering aimed at making the network more efficient in delivering the most requested content or content, such as voice and video, which has lower tolerance for latency.

The FCC was right to exempt these practices from its rules. Except to some legal academics and self-styled consumer advocates, they are uncontroversial features of the network that enhance the consumer experience.

Bigger concerns and the bigger picture
But even if Netflix isn't violating the FCC's rules, there are worrisome signs that the Open Connect content delivery network isn't quite as open as its name implies. For one thing, the network is open only to Netflix-hosted content. It therefore excludes competing video services from companies including Amazon, Hulu, and YouTube, as well as programming offered by the ISPs themselves.

And while the company describes the network as being "free" to participating ISPs, the company's "guidelines" include requirements that ISPs connect to the same peering locations used by the Netflix network and invest in connections of at least 10Gbps.

Though the specific details of TWC's dispute with Netflix aren't public, the objections seem to be focused on the fact that CDNs ordinarily pay the ISPs to co-locate servers that cache the content of their customers. While Netflix is offering its Open Connect network for "free," other CDNs pay for access to the ISP's facilities, operation of co-located equipment, and for the unbalanced peered traffic flow.

As attorney Jonathan E. Nuechterlein of Wilmer Hale pointed out last week at an American Enterprise Institute conference on Internet competition, in principle there's nothing unusual about such arrangements. Indeed, "paid peering" is the norm for connections between backbone providers when traffic flow is significantly imbalanced. In effect, CDNs represent a kind of paid peering for content that is either frequently used or which requires minimal latency -- or, in the case of video, both.

Settlement-free peering, paid peering, and co-location are all common and uncontroversial arrangements made between major Internet backbones and large content providers around the world. (Most peering agreements are done on a handshake basis.) The "open" Internet works well because of, not despite, arrangements and engineering that optimize network management.

Aside from Net neutrality, potential legal problems could arise, however, if Netflix uses its competitive leverage to demand participating ISPs house and operate its equipment without charge.

Beside other video providers that cannot host their content on the Netflix Open Connect network, for example, traditional CDNs without Netflix's growing market power could also be disadvantaged. These include general purpose CDNs such as Akami and Limelight, as well as start-ups including CloudFlare, which bundles advanced malware protection into hosting and caching services. It also includes Level 3, which was, prior to the launch of Open Connect, the company that Netflix paid for CDN services.

In addition to the leverage Netflix now has from its 30 million customers, the company has improved its market position by signing exclusive content deals, most recently with Disney and its affiliates, including Pixar and Marvel.

But is that kind of market power a problem, and, in particular, a problem for the FCC? Netflix, at least, seemed to think so in 2010, when it filed its comments in the Open Internet proceeding.

The company wrote:

[T]he Commission should not take too narrow a view of the type of discrimination by network operators that should be cognizable under the new rules. There is substantial discrimination and consumer harm if a network operator uses its ownership affiliation with a program content provider, or even its bulk buying leverage with a video content provider, to deny attractive programming to a competing online video service.

That, of course, is precisely what Netflix is now doing with its exclusive licensing and "bulk buying leverage." The Open Connect network not only denies programming to customers of ISPs that don't agree to the company's "guidelines." Moreover, it denies "attractive programming" to "competing online video services," including those of other over-the-top video providers.

The choice for lawmakers
Fortunately for Internet users and, it now seems, Netflix itself, the FCC wisely rejected the company's calls for broader and more detailed rules. Such regulations could have foreclosed the company's own future innovations, including the Open Connect content delivery network.

This is precisely why many argued during the FCC's Open Internet proceeding that the growing complexity of the Internet ecosystem, coupled with rapid changes in technology, made it unlikely the FCC could ever craft rules that wouldn't become quickly obsolete if not counterproductive -- and, given a dynamic competitive landscape in which content providers were gaining market power, unnecessary.

The bottom line is clear. People consume more high definition video all the time. We already need all the content delivery networks, IP networking technologies, and fiber optic cabling we can get to meet exploding demand for better quality, faster speeds, and more choices. For the Internet to continue expanding in both scope and quality, we'll need a wide variety of innovation -- from ISPs, content providers, and regulators.

But let's not kid ourselves. These are already fast lanes on the Internet, and Netflix its right to be deploying technologies to expand them as fast as possible. Critics who reflexively point to such developments as violating the spirit if not the letter of the FCC's Open Internet rules dangerously oversimplify a complex technical and business architecture. They aren't helping anyone, and certainly not consumers.

As U.S. Rep. Steve Scalise (R-La.) said in the closing keynote at State of the Net, "Our response to this newly converged marketplace should not be to level the playing field by extending legacy regulations to new entrants and technologies, but rather to let the Internet work and repeal the archaic and obsolete rules now on the books."

Scalise was talking about regulations that are unintentionally constraining the ability of legacy phone companies to replace their switched networks with new IP-based technologies. But, it seems, the FCC's more recent efforts to "preserve" the open Internet may have aged even more quickly.

Even if the FCC's Net neutrality rules survive the impending court challenge, they may prove to have little to offer Internet users -- little, that is, except for unfortunate and unintended consequences.