On Sat, Dec 30, 2000 at 02:58:02PM +0100, Oskar Sandberg wrote:
> Freenet is a complicated system with a lot of aspects. Depending on what
> one is refering to, the amount of trust needed varies. If you are looking
> at the anonymity of the node operator, then it is 100% trust - you need to
> trust every node (which if why we have never claimed this as a feature).
It isn't even this simple. It is not just a case of Anonymous/Not
Anonymous, but of the probability that you will be compromised. With
the current mechanism the probability that you will be compromised on
any given request would be about (10*p)/N where N is the size of the
network, 10 is the number of nodes involved in answering the average
request, and p is the number of nodes in Freenet that are controlled by
whoever is interested in compromising your anonymity. As you can see,
the bigger Freenet is, the better your chances.
With Shadow nodes, the probability becomes equal to the probabilty that
the shield node chosen by you has been compromised.
Of course, this is an incredible oversimplification (ignoring factors
such as traffic analysis), but I think it might be useful to come up
with metrics for anonymity - and my point is that a yes/no metric is
insufficient.
Ian.
PGP signature
