Author: Mike Hearn 2013-10-31 12:01:37
Published on: 2013-10-31T12:01:37+00:00
In an email thread from October 2013, Gregory Maxwell raised a concern about the possibility of a node using priority queued rate limiting for its relaying to "accept" a transaction but then have it fall out of its memory pool before ever sending it on to other peers. However, the author of the email believes that this race condition is trivial and can be easily resolved. They suggest that there is no requirement for a transaction to be placed into a buffer before relaying and that the gap of a few seconds between relaying and placing the transaction into the buffer should not cause any issues. Additionally, Gavin's smartfees branch adds mempool persistence to disk, so restarting nodes won't clear the mempool in the future. It is also noted that even if all nodes are honest and well-behaved, it cannot be guaranteed that they will forward transactions, making measuring propagation a necessary part of Bitcoin wallets for the foreseeable future.
Updated on: 2023-06-07T18:41:19.773439+00:00