Imagine a distributed public message board system. Similar to Usenet but each node does not need to store the whole thing. Each node only stores only part, and if it wants a message it doesn't have, routing happens.
In order not to be overwhelmed by spam, each node does spam filtering, choosing not to store or forward some messages. Some of the spam filtering (the hard cases) might even be done by the gold standard: the human user of the node.
Can such a system work? What are the conditions?
An interesting, probably unrealistic, simulation is to create a network, simulate messages sitting and bouncing around in it, and spam messages having a higher probability of being dropped. Can we avoid spam from overwhelming the network?
No comments :
Post a Comment