What Is Networking?

What are the challenges in adopting a content-centric networking model for the Internet?

• From Wikipedia: Content-centric networking (also content-based networking, data-oriented networking[1] or named data networking[2]) is an alternative approach to the architecture of computer networks. Its founding principle is that a communication network should allow a user to focus on the data he or she needs, rather than having to reference a specific, physical location where that data is to be retrieved from. This stems from the fact that the vast majority of current Internet usage (a "high 90% level of traffic"[3]) consists of data being disseminated from a source to a number of users. The contemporary Internet architecture revolves around a host-based conversation model, created in the 1970s to allow geographically distributed users to use a few big, immobile computers.[4] The content-centric networking seeks to adapt the network architecture to current network usage patterns.

• Answer:

  Great question. And before I answer (it's going to be a long, long answer), let's say that building a new network model from the scratch would be impossible in practice. But gradual evolution towards a new model from the current one is not only possible, but it's what happening right now. l always refer people to a paper written by Andrew Odlyzko in the turn of the millenium (http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/833/742), aptly named "Content Is Not King". Odlyzko argues that in the historical timescale, connectivity - or people-to-people communication business - was always much bigger than the content business. There are plenty of numbers, since the advent of newspapers, mail, telegraph, telephone, radio and TV, to back it. It's not a technical issue, it's a business and human/social one. Also, it's a very well written paper so I'll not waste your time (or mine) trying to explain it thoroughly. You should read it yourself, it really deserves it. Now, if you look at the data traffic pattern today, it may seem as if the Internet is trending towards a "content centric" model. And that's partly true. But the content that most people want to read is not generated in a broadcast model. Instead, people are actively producing and contributing a immense amount of content, using a few channels such as Twitter, Facebook and YouTube. So, content generation is distributed (and that does not have anything to do with the network architecture), but content distribution is getting more and more concentrated on a few players. What we are seeing is an artifact of the way content is being distributed today. People aren't relying on direct person-to-person connections the same way they relied on connection-oriented networks such as the telephone. They now produce and post content, and this content is acessed by other people, anywhere and anytime. It's an asynchronous content model - one that was not predicted by Odlyzko but that has a direct impact on content production and consumption, data traffic, and network architecture. I prefer to call these services "hubs of content", or even "content bazaars", and not content providers, due to the role they have in the current Internet. I would like also to point out a huge problem that the end-to-end communications model faces today, that is the transition from IPv4 to IPv6. It's not clear if  this transition will be as sucessful as its proponents believe. And the fact that there's a transition going on gives the opportunity to think about other alternatives. I believe that there's now a window of opportunity for an alternate model for a global network, one that is based on content (identified by its URL) and identity. It's a incremental change that is already taking place, under own our noses. We now rely on URLs for almost anything we do, and our digital identity is converging on a few big players (Google, Facebook, Twitter, and others such as Microsoft and Yahoo). So we don't need to build a new Internet, because the necessary changes are already happening. In this new network model, we will see some changes. CDNs will be come more and more pervasive. Network connectivity will be richer (more networks connecting directly), with hubs of traffic getting closer and closer to the users. The need for a single IP addressing space is going to be less and less noticed, as CDN nodes and protocol gateways start to be installed at strategic points of the Internet. End to end connectivity will be replaced by a mediated connectivity architecture, where two arbitrary people can talk directly using a third party as a relay node. And while new protocols may be developed to allow higher performance over "fat pipes", these protocols will probably be developed on top of the existing ones (IP and TCP) and not over a completely new protocol. There's no need for that. Note that most of the problems with NAT go away if you can assure the user identity and also address content directly. And the transition towards a "content network" is already taking place, but that's not the content network media business keeps talking about. It's about ourselves, our content, being summarized and distributed in an efficient model. This is what is happening, right now.