You could be excused for doubting that the Internet can keep up with the rapidly increasing demands on it. In addition to all the e-mails, songs, Skype calls, and Web pages flitting about, the amount of video online is mushrooming. By 2014, video will account for roughly half of consumer Internet traffic, up from just 12 percent in 2006, according to Cisco Systems. If the Internet’s ability to handle all that traffic can’t keep scaling, the resulting congestion will slow all kinds of applications for everyone.
But Paul Sagan, CEO of Akamai Technologies, says that people underestimate how much capacity there is. He’s in a position to know, because Akamai ushers along 15 to 30 percent of Web traffic. The company, which was founded by an MIT mathematician and graduate students in 1998, has ingenious methods for speeding the delivery of Web content. One such trick is to make sure popular pieces of content don’t get trapped in bottlenecks; Akamai’s technology makes highly sought files available from 75,000 servers it has stashed inside the networks of Internet service providers. From there, the files have an easier time going the “last mile” to end users. Sagan described his view of the Internet’s future to Technology Review’s deputy editor, Brian Bergstein.
TR: What’s a good way to describe just how much Internet traffic has exploded?
Paul Sagan: This year, in one week we delivered as much data as we did in all of 2005.
TR: Why do you think there’s still room for so much more?
PS: If you think of the capacity as the sum of the last miles, rather than wherever the narrowest point is at any given time, capacity is much greater. If you said, “I’m going to use one data center to deliver [online video streams of] the World Cup”—several terabits a second of simultaneous traffic for over a million users—you simply can’t get that much capacity out the door, let alone through all the networks that make up the Internet, to the edges. We were able to do it because we were delivering it from the edges, where people connect to.
TR: You’re making copies of videos and other Web content and distributing them from strategic points, on the fly.
PS: Or routes that are picked on the fly, to route around problematic conditions in real time. You could use Boston [as an analogy]. How do you want to cross the Charles to, say, go to Fenway from Cambridge? There are a lot of bridges you can take. The Internet protocol, though, would probably always tell you to take the Mass. Ave. bridge, or the BU Bridge, which is under construction right now and is the wrong answer. But it would just keep trying. The Internet can’t ever figure that out—it doesn’t. And we do.
TR: But there must be an increase in demand that the architecture couldn’t handle.
TR: Couldn’t that be caused by millions of hours of high-definition video?
PS: I think what would happen, if it gets ahead of itself, is that people would have a less than good experience, and they’d stop watching some of it, and it would drop down, and the buildout would catch up. People are putting huge investments in data center capacity. And they’re putting huge investments in the last-mile networks. Why? There’s huge money to be made in both. If you’re in the middle, there’s not a lot of money to be made, and so as we’ve seen Internet traffic skyrocket for 15 years, the middle always lags what’s going on at the first and the last mile, and that’s not going to change. [But] our view is, if you can skip the middle, the Internet will continue to scale along.
TR: We’ve heard dire claims about the state of the Internet for a long time.
PS: Predictions that the Internet is going to crash are not new. They’ve been around as long as we’ve been here, and even back to dial-up. And there are many good business opportunities that people have found by investing to make sure that that doesn’t happen. And I think it’s going to continue to happen. I don’t think the Internet is just going to crash and burn because there’s too much video.
TR: And it’s not all or nothing with video, right? Services like yours can adjust the quality of a video stream to make it take up less bandwidth if needed.
PS: People in fact have that experience today and they don’t even know. They’re watching on a big screen, the bandwidth is varying a little bit in the background, and we’re delivering an uninterrupted picture. There’s almost the equivalent of a power brownout in the background, but it’s not affecting their experience.
TR: Isn’t the problem of getting bits across the last mile much thornier for wireless?
PS: There is a challenge in wireless, which is: as new as that infrastructure is, it was built for voice, not for data. These networks don’t even speak IP as the protocol of the tower. So they’re taking Internet traffic and then converting it to proprietary protocols. That’s changing with 4G and probably what will follow that. And as that happens, it gets easier for Akamai to extend farther. There will be new things we’ll do in wireless over the next five years.
Geoffrey Hinton tells us why he’s now scared of the tech he helped build
“I have suddenly switched my views on whether these things are going to be more intelligent than us.”
ChatGPT is going to change education, not destroy it
The narrative around cheating students doesn’t tell the whole story. Meet the teachers who think generative AI could actually make learning better.
Meet the people who use Notion to plan their whole lives
The workplace tool’s appeal extends far beyond organizing work projects. Many users find it’s just as useful for managing their free time.
Learning to code isn’t enough
Historically, learn-to-code efforts have provided opportunities for the few, but new efforts are aiming to be inclusive.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.