Structure of Dark Nets
One usually thinks of the Internet and especially online social networks like Facebook as networks that connect people to people. A different, less known part of the Internet, commonly called the Dark Net, is where users can operate in complete anonymity. It has a tainted reputation as being an underground marketplace forum for people to buy narcotics or access illegal material like child porn, but others also use it for political advocacy, whistle blowing, or just browsing the regular Internet.
A Dark Net is a private network with internet-like routing between peers. Tor is the most well known Dark Net, and was initially developed by the U.S. Naval Research Laboratory to protect the communications of the U.S. Navy, however, since all the traffic going through the network was from the U.S. Navy, they were not very anonymous, so they opened the system to other groups of people. Tor is now publically accessible, allowing them to hide in a large group of anonymous Tor users.
It is difficult to remain anonymous while retaining the ability to communicate with a large number of users. Every website one visits receives an ISP-issued IP, which gives it an approximate location. In contrast, a Dark Net allows one to use an internal, non-globally routed IP address. Participants in the Dark Net, excluding direct peers, will never be able to see your real IP address.
There are two main components to starting a Dark Net. First of all, to join a network’s routing infrastructure, one needs to peer with one or more existing participants on the network. You can do this either through physical connections (Ethernet links, wireless connections) or a virtual Ethernet tunnel using VPN software. This is called Peering.
The second component is Routing. The author of the linked article analyzes a BGP routing scheme, and a sample network map would look like the following.
The network is strategically set up so that only neighboring nodes know each other’s real IP addresses. For an entity (such as the police) to track down a user’s IP, it would have to go through a chain of peers and break each one to find out whom they’re connected with. This is extremely expensive on large Dark Nets.
Compared to a regular network graph, a graph of a Dark Net has much fewer edges, as it is in the interest of the user to limit peering to only very trusted connections, as that would increase its anonymity and make it less easy for its address to be found. From the sample graph, you can see that if the central link between 787 and 3090 break, then the network separates into two components. The breaking up of a network like this into multiple components seems much more likely in a Dark Net due to the nature of its peering and its decentralized structure, compared to a global friendship network that has just one large component.
While some networks’ main goal is to increase the number of links between nodes, the Dark Net works to function on the least number of links necessary.
http://security.stackexchange.com/questions/29366/what-are-darknets-and-how-can-they-be-used-to-provide-security-and-anonymity-in