������ý

The internet is amazingly robust, but like any complex network is still prone to the occasional failure. A new analysis using network theory explains why the dark net – the hidden underbelly of the regular internet, invisible to search engines – is less vulnerable to attacks. The lessons learned could help inform the design of more robust communications networks in the future.

The regular internet’s design is deliberately decentralised, which makes it very stable under normal circumstances. Think of each site or server as a node, connected to numerous nodes around it, which in turn connect to even more nodes, and so on. Take out a node or two here or there and the network continues to function just fine. But this structure also makes it more vulnerable to a coordinated attack: take out many nodes at once, as happens during a distributed denial of service (DDoS) attack, and the result can be catastrophic failure that cascades through the entire network.

The dark net is much less vulnerable to such directed attacks, thanks to its unique structure. and at Rovira i Virgili University in Tarragona, Spain, used data from the Internet Research Lab at the University of California, Los Angeles, to build their own model of the dark net. They ran simulations to see how it would react to three failure scenarios: random node failures, targeted attacks on specific nodes, and cascading failures throughout the network.

They found that an attack on the dark net would need to hit four times as many nodes to cause a cascading failure as on the regular internet. This stems from its use of “onion routing”, a technique for relaying information that hides data in many layers of encryption. Rather than connecting a user’s computer directly to a host server, onion routing bounces the information through various intermediary nodes before delivering it to the desired location. This stops an attack from spreading so widely.

Powerful connections

Another reason for the dark net’s resilience is its lack of something called the “rich-club effect”. In the regular internet, powerful nodes connect more readily with other powerful nodes, creating what at Carnegie Mellon University in Pittsburgh, Pennsylvania, terms a “smoky back room” of “network elites”. An attack on one such node can trigger the failure of others, which can in turn lead to cascading failure across the network. The dark net doesn’t have this high level of connectivity between powerful nodes.

“This is [another] one of the things that make it more robust to attack,” says DeDeo. “The network elites are more spread out. In fact, the elites appear to be avoiding each other.”

This model of the dark net somewhat resembles a so-called “small-world network”, in which several heavily connected nodes link clusters of smaller local nodes – similar to how major air traffic hubs connect smaller local airports. Both systems exhibit similar resilience to catastrophic failure, although in-depth comparisons have yet to be completed.

Reconfiguring the entire internet to make it as robust as the dark net would be prohibitively expensive, but De Domenico thinks the pair’s work could still offer practical insights. “It is possible to rethink next-generation upgrades and the design of more localised communication networks, like the intranets of large companies,” he says.

Journal reference: Physical Review E, DOI:

Read more: Invisible: A visitors’ guide to the dark web