Closed roeslpa closed 8 years ago
@roeslpa Your suggestion looks plausible to me. However, I know too little about proof-of-work protocols to do a in-depth analysis. For now, this should do, but we should keep this in mind for later auditing. Furthermore, I'd would be useful to read what inspired your suggestion (existing protocols).
@cburkert : was that an "ack"?
@cburkert I added my inspiration and modified the protocol to prevent precomputation. Would you ack for now and we keep it in mind for an audit?
I have several question about this spec until I can ack it. We should discuss this in real-time and post the outcome here.
In any case we need an issue (enhancement, future work) about independently auditing the proof-of-work spec.
My questions are answered and this spec looks sane to increase the amount of work someone must do to (D)DoS the (drop) system.
We could force this protocol only on suspicious
connections (e.g. on connections using IPs with more than r
requests per time unit).
Wrote a more specific description of the proof of work. Resolves #68