“Peering” is a term you should know if you care about the future of the Internet. Although you might guess the word is related to NSA surveillance, it’s actually related to issues like net neutrality, and has implications for the proposed merger of the country’s two largest cable companies: Comcast and Time Warner Cable.
Most people don’t think much about how content is delivered to their computer or on their television screen via a device like Apple TV or a Roku box. Normally, you click a link or enter a web address into your browser, and the requested content usually appears almost instantly. Sometimes you get a perpetual “spinning circle of doom” or the dreaded hourglass, indicating that you may experience delays in receiving the content you requested. Or occasionally you get an error page instead of reaching the webpage you were trying to visit. For people trying to access Netflix, you’re probably familiar with the screen that says there’s a network problem.
Network neutrality protections were designed to address some of the reasons you might have trouble reaching a certain page. The FCC’s Open Internet Order, which was mostly overturned in January by the D.C. Circuit Court of Appeals, contained two rules for wired broadband providers: A “no-blocking” rule that prohibits Internet service providers (ISPs) from blocking content from their subscribers, and a prohibition against discriminating against certain content or types of content using other means. But where does peering fit into this discussion?
For people trying to access Netflix, you’re probably familiar with the screen that says there’s a network problem.
First, let’s dive a bit deeper into the depths of the Internet. Rather than one monolithic network, the Internet is better understood as a bunch of interconnected networks, with different companies owning and controlling different parts of the pipes. For the Internet to work as we’ve come to expect, those companies all have to figure out a way to interconnect and carry each other’s traffic. Without interconnection, a Time Warner Cable subscriber could only access the content that existed on the Time Warner Cable network, for example, and not content that is hosted on someone else’s network. When you click a link, you expect it to take you to the content you are trying to reach—you aren’t necessarily concerned with (or even aware of) where that content resides.
Historically, interconnection happened in an uncomplicated way, because everyone benefited when end-users could access content on all parts of the various networks, and the flow of traffic was mostly balanced. The resulting arrangement was called “settlement-free peering”, and companies would simply carry each other’s traffic without exchanging money. In this scenario, there was relative parity and fairly robust competition between so-called “backbone” providers (companies like Level 3, Cogent, etc.), and negotiations were mostly straightforward.
However, in recent years we’ve seen the rise of content delivery that creates a heavier capacity stream in one direction (usually in the direction of the user, and usually in the form of streaming video), which has made the network economics more complex. This shift led to debates about peering, with different network operators disagreeing about how content should be carried, whether certain traffic should be paid for, and who should pay. As the debates played out, generally content providers like Netflix would negotiate with backbone network providers operating as Content Delivery Networks (CDNs) to host their content.
The ISPs (companies like Comcast, Time Warner, and Verizon) argued that because of the so-called one-sided strain on the network, CDNs should have to pay a premium for delivery to the ISPs’ end user. This is a change from settlement-free peering, because in this case, someone’s got to pay. In addition, there’s relatively robust competition in the backbone, but a lack of competition among ISPs at the edge of the network, meaning there the market forces that drove interconnection practices at the backbone don’t necessarily exist at the edges of the network. For example, even if there are multiple ISPs available in a given area, home Internet users typically only subscribe to one of them at a time. Because of that fact, ISPs have what we call a “terminating access monopoly” to their subscribers, which means there is only one pipe to reach the user. That gives ISPs a large degree of gatekeeper power to control and charge for access to customers in their homes. Thus, some have argued that the terminating access monopoly allows ISPs to charge the CDNs a fee that has no basis in market realities, all while also charging end users for expensive monthly Internet subscription.
This all changed further last week. Netflix and Comcast announced that the two companies have come to a new—indeed, unprecedented—peering arrangement. Sources indicate that the arrangement involves payment from Netflix to Comcast for the direct carriage of Netflix content to Comcast subscribers. Verizon is expected to follow suit any moment with a similar agreement with Netflix.
Because the world of peering arrangements is shrouded in a cloud of non-disclosure agreements, except for the occasional public debate when things go totally off the rails, it’s extremely difficult to assess the reasonableness of the rates charged for this agreement or for any others. And given the degree to which content providers rely on those tenuous agreements to get their content to users, there is an inherent disincentive for them to cry foul when problems arise. But peering issues nonetheless have real implications for all of us as Internet users, and implications for content developers as well.
Because the world of peering arrangements is shrouded in a cloud of non-disclosure agreements, except for the occasional public debate when things go totally off the rails, it’s extremely difficult to assess the reasonableness of the rates charged for this agreement or for any others.
For content creators, the opaque world of peering arrangements means that it’s impossible to account for future peering fees, particularly now that content companies have to negotiate directly with ISPs, rather than through a CDN. Large, game-changing media delivery systems like Netflix didn’t happen overnight, and the next big “Netflix” or Netflix-competitor might be just an idea at this point. Knowing that if they get too big, a company will suddenly have an additional, potentially huge expense to pay could be just the type of hurdle that would deter them from developing innovative tools or services.
These problems inevitably carry over to end-users as well. Netflix was the disruptive force that facilitated a world of “cord-cutting”—where users ditched their cable service for online streaming. But if new peering arrangements prevent companies from developing products to compete with Netflix, customers of Netflix might see their prices continue to rise or miss out on new, disruptive services entirely.
Moreover, new challenges may arise in light of the proposed merger of Comcast and Time Warner Cable. Comcast already owns a huge amount of media content from its acquisition of NBC Universal, and stands to gain more if it acquires Time Warner. Moreover, Comcast has also been willing to experiment with caps on subscriber data usage, while Time Warner Cable historically has not. An increase in the use of data caps might give a combined Comcast/Time Warner Cable the ability to indirectly control what content subscribers access online—by exempting certain content from caps but not others. Comcast is still bound by the Open Internet Rules as a condition of its merger with NBC Universal, but many have argued that data caps exemptions and peering arrangements all fall outside the bounds of those Rules. (Whether or not that is the case is fodder for a separate post.)
These issues underscore precisely why the need for the non-discrimination protections didn’t disappear when the Open Internet Rules were overturned. To the contrary, its become more important now than ever, because the need for protection doesn’t merely exist at the very edges of the Internet. As the Internet has evolved, the points at which problems occur have shifted into the world of peering and interconnection. It’s therefore imperative that the FCC exercises the clearest authority possible not only to revive the Open Internet Rules but also to deal with the Internet as an ecosystem and address the various problems that can result within that ecosystem.