Policing other nodes #31
Description
To protect the network against cybil attacks, it is necessary to check if each node is doing their job.
For the actor networks, nodes will get reputation from delivering requested data, which should be enough.
However, the overlay network is also responsible for giving you some peers to join the actor network with. But whether those peers are constructive or malicious is not really as easy to check as verifying a hash.
The idea is that any node on the overlay network that doesn't exist in your own web of trust but where you stored the actor info and yourself as a participant, will need to be tried out by another trusted node, at a random moment in time. The trusted node will check if the node is still online, and if it is online but it does not deliver the actor info or your contact info upon request, it will consider the node a malicious node.
Then, a message will flood all the nodes in your web of trust, to signify that that node is not to be trusted, so that hopefully a significant part of the overlay network will stop interacting with it.