axpoems
commented
3 days ago We already have a prevention mechanism in place for this, which is the proof of work factor applied every time a message is sent. I have made a PR to increase the cost for public chat messages. There is also an adjustment factor that the security manager can set and will propagate to the network.
suddenwhipvapor
It was probably just a matter of time, and now that I think about it, strange it didn't happen sooner. I noticed a moderation report in my instance about "spam", and then I went checking, there were 308 unread messages in Discussions or Assistance, can't remember which one anymore, as, after ascertaining the reported user was the one behind the garbage being posted, I immediately banned him and deleted everything. He still managed to flood the chat with 300+ messages. Other than being a UX nightmare, this could be used as a means for DDoS one day, so I think there should be a measure to have the local client limit the total number of outgoing messages per time unit (10 per minute and 20 per 5 minutes? to be defined), and also remote clients refuse to relay new messages coming from the same profile, that overcome that limit (in case someone modifies his local code), just like lightning nodes will not relay gossip when it's been broadcasted at a rate higher than a certain value.