Author: David A. Harding 2023-01-10 20:14:47
Published on: 2023-01-10T20:14:47+00:00
In a discussion about decentralized coinjoin implementations like Joinmarket, Peter Todd raised the question of how to implement "conflict monitoring". Two methods were suggested in response. The first method involves running a relay node with a conflict-detection patch and using Bitcoin Core with -debug=mempoolrej to identify conflicting transactions already in the mempool. This allows for quick detection and attribution of fault, but requires Alice's node to be online at the time of conflict propagation. The second method involves assuming a conflict exists when there are otherwise unexplainable failures, such as new blocks with lower feerates than an unconfirmed coinjoin transaction. This method is slower (taking hours) and does not attribute fault, but does not require Alice to have any information beyond recent block copies. Both methods can be used to evade the $17.00 attack and the $0.05 attack without the need for full-RBF.
Updated on: 2023-06-16T03:53:25.588516+00:00