PARC's plan to stop the Internet from crashing
March 7, 2011, 4:35am PST | Length: 00:05:39
"Content distribution has reached a scale that simply doesn't work," says Van Jacobson. The scientist and research fellow at Xerox’s Palo Alto Research Center talks to ZDNet's Sumi Das about Content-Centric Networking (CCN) a new technology he's developed that could make content distribution on the Net more efficient.
>> Sumi Das: As more and more bandwidth heavy applications and Web sites crowd the Internet, web congestion becomes an increasingly difficult challenge. But scientists at Silicon Valley's storied research lab PARC are working on the problem, developing new technologies to rearchitect how the Internet runs. With me is Van Jacobson, a research fellow at PARC. Van, thanks for joining me today.
>> Van Johnson: Thank you for having me.
>> Rebecca Jarvis: Before we get into the issue of web congestion, we wanted to talk a little bit about you. You're a scientist and one of the key contributors to developing the original protocols and laying the foundation for the Internet backbone. So tell us a little bit about your background. It sounds like there's a lot that you've done in the past.
>> Van Jacobson: I was working at Lawrence Berkeley National Lab, making control systems for big particle accelerators, and that took me into computer networking in the early '70s. And it was more interesting than accelerators, so I kept working on it, and for about 25 years I ran the Network Research Group at LBL. Then I left there and spent a few years at Cisco as their chief scientist; left Cisco to do a start-up called Packet Design with Judy Estrin, and then in 2006 went to PARC and started working on Content Centric Networking.
>> Sumi Das: So that's the technological breakthrough -- Content Centric -- do I have that right?
>> Van Jacobson: Yes.
>> Sumi Das: Content Centric Networking, CCN, as you sometimes call it.
>> Van Jacobson: Exactly.
>> Sumi Das: And this is going to help keep the Internet from getting overloaded. Explain how it works.
>> Van Jacobson: If you look at how people use the Internet, they're asking for some piece of content that they want, a video that they want to watch, a song that they want to download from iTunes, tweets that are coming out. But the Internet was originally designed before we had this notion of data that people would want, and it was designed for computers to make phone calls to other computers. And that's a really inefficient way of distributing content. If you think of having to get your TV programs by making a phone call to the TV station, that kind of works out where you're making the phone call, but at the TV station they can be getting hundreds of thousands of calls, all asking for the same stuff. And the goal of Content Centric Networking is to get out of this phone call world and instead ask the network for what you want, not say who you want to talk to, but say what content you want. And that lets it be far more efficient to doing distribution.
>> Sumi Das: So what's the danger of proceeding with sort of the status quo that we have?
>> Van Jacobson: The problem is the content distribution has reached a scale where what we have simply doesn't work. And an easy example of that is if you remember the Susan Boyle phenomenon of a couple of years ago. Over a period of four days, a hundred forty million copies of that video were played by YouTube to people watching it all over the world. And if you multiply the bandwidth of that video about a megabit times a hundred forty million viewers, there's no technology, no communications technology on the planet that can deal with that kind of load. YouTube can handle it, because they fooled the Internet. They're not really a place that you make a phone call. They're actually big distributed content source. They're spread out all over the planet, and so they don't get the kind of traffic concentration that would prevent something from working. But basically Google's the only place that can do that, and--
>> Sumi Das: If you're not Google, tough luck?
>> Van Jacobson: Exactly, and I think that's probably not a world we want to end up with, that there's a lot of good ideas that people have for content distribution, things like Twitter and Facebook, that are easy to think about, are easy to roll out at a small scale, but making them Internet scale is really hard because you have to invest all your effort in fooling the Internet. And we'd like to make a Net that works the way that people think it works, that you ask for what you want and you ask the Net directly for what you want, and it can give it to.
>> Sumi Das: You shouldn't have to be a billion-dollar company to be able to distribute information on the Net, in other words?
>> Van Jacobson: Exactly.
>> Sumi Das: So what will it take to make CCN a reality? You're going to be facing some challenges, I would imagine?
>> Van Jacobson: We spent the last five years at PARC doing an OpenSource prototype of the protocols. They're available at a site called ccnx.org. Anybody can download and play with them. And the biggest challenge is rolling it out, getting people to try it in different environments, show that it works, get the rough edges up. But we're using it to solve problems now with companies like Samsung trying to make the mobile environment work better.
>> Sumi Das: Well, I think all the consumers of web information and viral video fanatics are pulling for you, so thank you so much for explaining this to us.
>> Van Jacobson: Thank you for having me.
>> Sumi Das: Thanks for watching.
==== Transcribed by Automatic Sync Technologies ====