Thread: NDN to replace TCP/IP..?

Interesting read on the possible future of network engineering.

The full article can be seen here:
The Mission To Save The Internet By Rewiring It From The Name Up | Motherboard

While most of us have been binge-streaming or strapping computers to our bodies or wrapping our heads around the ins and outs of net neutrality, an international team of academics and some of the world's biggest technology companies have been quietly pondering how to rewrite the basic structure of the internet—for our sakes.

Their idea sounds simple: instead of numbers, use names. Focus not on the locations of things, but on the things themselves.

The proposal, called Named Data Networking, shifts the focus from the numbered locations of data—IP addresses like 174.16.254.1—to the very names of data—something like motherboard/stories/NDN/photo1. Under this system, for example, when your computer makes a packet request for a new Netflix release, you could retrieve the video from the nearest computer that has it, rather than wait to get it from Netflix's heavily-trafficked centralized servers.

"As far as the network is concerned," the project's website says, "the name in an NDN packet can be anything: an endpoint, a chunk of movie or book, a command to turn on some lights, etc.” An internet not of numbers, but, if you will, of things.
What that means, in practice, could be big. An internet focused on the what, not the where, could be a more flexible internet, less likely to get clogged up as a steady stream of new devices join the party. An internet that no longer relies on the aging architecture known as TCP/IP could also be an internet with fewer of the middlemen that currently throttle speeds, gather our data, or control what can and can't be seen.

One person who thinks a lot about the need for switching to NDN is Lixia Zhang, a professor of computer science at UCLA who is known for research into protocols and security that have impacted how we use the internet today. (It's a far cry from her first job: as a teenage tractor driver in northern China during the Cultural Revolution.) Her own interest in NDN—she's now one of the project's leaders—emerged out of what she saw as changing needs for the network: The current internet simply doesn’t fit how we want to use our devices anymore.

Over lunch recently, a colleague at UCLA asked her to clarify that: Why, exactly, does the internet need a new architecture?

"I said, ‘There’s a simple answer.’ I had my phone on the table, and he had his phone on the table. I said, ‘Look, our phones are next to each other, can they talk to each other?’ He said, ‘of course they can.’ I said, ‘directly?’ And he said, ‘no, they cannot.’ And this is not because physically they can not, physically [you can connect] phone-to phone using WiFi, Bluetooth. But the TCP/IP protocol is the roadblock for phones to directly talk to each other. The protocols need to change because it doesn’t enable the communication that we want.”



The plan for NDN is one approach to what's known as information centric networking (ICN), an idea with roots in Ted Nelson's 1980s experiments with hypertext. It wasn't until 2006, however, when projects at Stanford, the University of Texas at Austin, and at the legendary Xerox PARC launched the idea into the networking mainstream. Internet pioneer and PARC researcher Van Jacobson, whose talk at Google in 2006 has been widely cited as the start of the ICN movement, is now a leader of the NDN project alongside Zhang.

Now there are a range of smaller NDN-like initiatives globally, and last September, the ICN's various research questions were the topic of its own conference. Other ambitious projects seeking to address current problems with the internet include software defined networking (SDN) and network function virtualization (NFV).

Zhang said the various approaches aren't competing so much as complementary, tackling related problems. But NDN is the best funded, with $15 million through 2016 under the National Science Foundation, in addition to corporate assistance (more on that below). “I’d rather say that they are near-term solutions, and not really competition. NDN is specific as a new protocol architecture [to] replace TCP/IP."