Threat Model (dist.ribut.us)

Threats

Malice

A malicious threat is one that can disrupt the network, but provides nothing to the attack (except knowing they did it).

Our goal is not to stop malicious threats, but keep their 'leverage' down so that a significant portion of the network (~10%) would need to involved to cause a disruption.

Greed

Greed is any form of free-loading. Failing to fulfill delivery promises, not providing onion routing anything of this nature.

Much like malice, the goal is not to stop it but keep it low enough that it will not disrupt the overall performance.

Vandalism

This any type of attack intended to either deface a target or leave a 'signature'.

Spam

Forcing content on users they do not want.

Spying

Collecting data that users would prefer the attacker not have. This will be sub-divided into coerced and covert.

Optimistic Deterrence

We need models that can handle attacks with-in some defined bounds. Controlling 50% of the network computation is a good example - once an attacker represents more than 50% of the network, they're no longer the attacker, they're the proprietor.

But there are good models that can operate at higher efficiencies when the threat level is low.