Restructuring the Internet

Mario Felix | April 3, 2011

Professor Edmund Yeh: Associate Professor of Electrical Engineering, and Computer Science & Statistics in the Yale School of Engineering & Applied Science. Yeh is part of a multidisciplinary team of scientists that is currently developing its prototype for Named Data Networking (NDN). Photo Courtesy of Edmund Yeh.

When the Internet launched 40 years ago, its specific purpose was to share information among the computers of scientists in the Defense Department. It was built upon a specific architecture called TCP/IP. Essentially, this protocol allows for one machine to send a data packet with a source and destination address to another machine. Once a connection is established between the machines, the data packet will flow from the source to the destination IP address.

However, the number and types of machines that can send and receive information have increased dramatically in the past four decades. The way the Internet is used has also transformed. Websites such as YouTube, Twitter, and Facebook have proven that Internet services have become more content-centric. “It is not just about one machine having a conversation with another machine anymore. It is all about obtaining information,” says Professor Edmund Yeh, Associate Professor of Electrical Engineering and Computer Science & Statistics in the Yale School of Engineering & Applied Science.

Yeh is part of a multidisciplinary and multi-institutional team of scientists that was recently awarded $7.9 million over three years by the National Science Foundation as part of the organization’s Future Internet Architecture Program. Yeh’s team, led by Lixia Zhang of University of California-Los Angeles, was one of four teams that received this award for their idea of a new type of Internet architecture. The team’s innovative idea, called Named Data Networking (NDN), focuses more on data than the route it travels, which is how the present system functions. This would upgrade the Internet to operate in a very content-driven world.

The current architecture contains many problems that NDN would be able to address. One problem is congestion. For example, under the current architecture, a popular YouTube video with a recorded five million views means millions of different machines accessing the same source to download the video, thus creating congestion and delay. Under NDN, millions of machines can still watch the same video but with less congestion. “Imagine the first person who watched the video. Under NDN, the data would be cached on the pathway from the original source to the person’s machine. Then anyone who lives near that pathway can access and download that data without ever connecting to the source. This would relieve congestion on the source website server and therefore decrease delay,” explains Yeh.

Another problem that NDN would be able to address would be that of security and privacy. Under the current system, security is sought by securing the pathway between point A and point B. “This is not an effective system,” explains Yeh, “because there are multiple domains between these two points that we do not necessarily control. NDN can fix this problem by securing the data itself rather than securing the pathways.” According to Yeh, many of the technologies that could be used to authenticate data, such as digital signatures, already exist today. Because NDN would not utilize source or destination IP addresses, no one will be able to see who is accessing or sending data, therefore increasing privacy.

Currently, Yeh’s team is working to create a functional prototype of this system. Yeh will oversee the development of new routing, caching, and forwarding algorithms for this new Internet architecture. With this innovative type of architecture in place, Internet users will find the Internet much faster and safer.