Oh wait, never mind. Here’s why the latest panic over the Internet’s coming collapse doesn’t add up.
Just for fun, try to guess the year in which the following warnings about the Internet’s impending meltdown were issued:
No. 1: “Over the coming six to 12 months, computer users around the planet are likely to experience the Internet equivalent of the Great Blackout, or at least frequent brownouts, as our information infrastructure staggers and struggles under the heavy onslaught of new users and new demands.”
No. 2: “Internet access infrastructure, specifically in North America, will likely cease to be adequate for supporting demand within the next three to five years.”
No. 3: “Will Netflix Destroy the Internet? American broadband capacity might not be able to keep up.”
The answers: 1997, 2007 and this week—and that’s just a small sampling from the past 20 years. Such predictions of the Internet’s breakdown are always premised on the arrival of a scary new device or application that will send lots of digital bits over the Net. Back in 1995, when Internet sage Bob Metcalfe tried to explain why he foresaw “the Internet’s catastrophic collapse,” he cited a wave of new “Internet appliances,” in particular the dangerous Sony Playstation, which for the first time had Internet access!
These days PS1’s don’t seem so scary. Now the boogeyman is Netflix (NFLX), which runs 77,000 servers with big hard drives that it has placed in every nook and cranny of the Internet. When a college student downloads “Dexter Season 1” from Netflix there’s a good chance the show is already stored on campus on an Akamai box.
“That video is growing rapidly and going to be huge is true,” says Akamai co-founder Tom Leighton. “But there’s tons of capacity out at the edges of the network….plenty of capacity in the last mile to your house.” That capacity, he says, combined with smart delivery of Netflix content from nearby servers, means the Internet can handle Netflix just fine. If all that traffic had to travel closer to the center of the Internet then many larger peering points would be overwhelmed, Leighton adds. (There’s reason to trust Leighton’s numbers on both counts: he’s also a professor of applied mathematics at MIT.)
Luckily, most last-mile Internet providers happen to be much better at providing the big downloads a company like Netflix needs. (As for upload speeds, that’s another matter.) Cable systems that use Motorola (MOT) gear, for instance, can easily add more downstream capacity by “channel-bonding.”
Another reason to doubt the Netflix-is-flooding-the-Internet meme: Netflix. According to the company’s most recent quarterly report, the booming demand in downloads forced it to increase its spending on delivering that content by a not-so-whopping $3.6 million more than it spent the previous quarter.
Let’s round up to $4 million and still—who cares? We live in a world of cheap bits, especially for companies like Netflix that buy them in bulk. Even if the company really does make up one-fifth of Internet traffic during prime time, it can clearly afford it.
