A five-year-old app gets bought for $19 billion (WhatsApp). A Web document sharing firm gets valued at $10 billion (Dropbox). And cable TV firms are losing customers in their main video business. The Internet is booming, challenging established industries all around. Yet many greeted thenews that Netflix would connect its video servers directly to Comcast with apocalyptic shrieks of Internet doom.
The fears are misplaced. The Internet is changing, as it was always meant to. That’s the genius of its design. But no, Netflix’s decision to send video directly to Comcast, instead of routing it through other network intermediaries first, sets no precedent. Other firms like Google, Amazon and Microsoft have been linking directly to broadband networks for years. So have content delivery networks such as Akamai, Limelight and EdgeCast, which store videos, banner ads and other content closer to end users.
The grouchy insistence on U.S. Internet decline is a curious, and serious, matter, because the policies that govern our increasingly important digital economy, and all the things that touch it, will determine whether the U.S. has any chance to solve its growth, employment, and budget challenges. If we get technology policy wrong, all our troubles become far more severe.
The success of the U.S. Internet economy can be seen in apps and operating systems (Apple and Google), the cloud (Amazon, Microsoft, Salesforce.com), tablets and smartphones and in the networks that power all these tools. The U.S. has more 4G LTE mobile users than the rest of the world combined. Despite much confusion on the topic, U.S. broadband is among the most widespread and very fastest in the world. And America uses its networks more intensively than anyone but South Korea. The U.S. generates and consumes two to three times more traffic (per capita and per Internet user) than Western Europe.
These achievements are largely the result of a light-touch regulatory model going back two decades and more. Large companies were allowed to make enormous capital investments in fiber optics, cell towers, spectrum, software and data centers, and millions of entrepreneurs had the freedom to build the services, apps and tools that entertain us and improve productivity in (and disrupt) most other industries.
The pessimistic charge that the U.S. is failing, however, could send policy in a very different direction. The Netflix-Comcast news led to calls for applying the regulatory regime known as “net neutrality” (twice struck down by the courts) to not just last-mile broadband networks but the whole Internet ecosystem. Others called on the Federal Communications Commission to “reclassify” the Internet as a “Title II” telephone service. Do we really want regulate the Internet like an old monopoly phone network?
In a new report, called “How the Net Works,” we offer a condensed history of the Internet, hoping a primer on interconnection would help the quality of the debate. The reality is that smaller networks and content providers have always paid for access to larger networks. From the beginning, consumers, enterprises, and websites paid Internet service providers (ISPs), who paid larger regional “Tier 2” ISPs, who paid even larger continental or global “Tier 1” ISPs.
The rise of Web video did alter the equation because delivering smooth YouTube streams, for example, required less latency, or delay, between content firm and end user. Beginning 15 years ago, content delivery networks arrived to store oft-used rich content closer to end users, and nearly 10 years ago Google, with its purchase of YouTube, began building its own global data center. Content delivery networks gave a huge boost to the performance of the Net — for a price.
Web video also changed the nature of data traffic. Video requires far more bandwidth than email or webpages, and movies and video clips are mostly one-way. A Netflix movie, for example, is nearly all downstream traffic. When two networks of similar size and capability exchange roughly equal amounts of traffic, they might engage in “settlement free peering” — trading traffic at no cost. But Netflix-Comcast traffic is mostly downstream, so Netflix is willing to pay Comcast to deliver its bits in a more efficient and direct way.
Not exactly the end of the Internet. In
fact, if we get policy right for the next 20 years, like we did the last 20,
this looks more like the beginning.