Disposing of "dust attack" UTXOs

The discussion revolves around the concept of maintaining a "single input, single output" transaction pattern as a means to prevent a Denial of Service (DoS) vulnerability through batching at the relay level. By adhering to this model, it's possible to avoid relying on the standard minimum fee rate and still preserve relay-level batching. This can be achieved by implementing an append-only rule where a batched dust transaction is only replaceable by another that includes all previous inputs plus additional ones, always in a deterministic order. This method would help limit the replacement space to a linear growth, effectively managing the potential for an exponential increase in complexity.

The approach also contemplates the realities of miner behavior and the practicalities of transaction aggregation under the 'ANYONECANPAY' flag, which could be seen as offering a weight discount for single-input-single-output transactions. In scenarios where at least one such transaction is included, there would be no discount; instead, there would be a penalty involving an additional virtual byte to account for the worst-case scenario increase in compact size due to multiple inputs. Despite the theoretical framework, it seems miners have not widely adopted this practice, possibly due to insufficient incentives. The data shared in the charts suggests a maximum revenue potential of about 4 BTC under optimal conditions, which might not be compelling enough.

Moreover, there is an expressed concern regarding the security implications in environments where dust disposal transactions are infrequent. In such cases, an attacker could potentially correlate dust Unspent Transaction Outputs (UTXOs) that are confirmed or broadcast around the same time, posing a significant risk. This highlights the delicate balance between operational efficiencies in transaction processing and the overarching need for robust security measures to mitigate potential attack vectors in blockchain transaction management.