Author: email at yancy.lol 2022-10-30 11:06:32
Published on: 2022-10-30T11:06:32+00:00
In a discussion on the bitcoin-dev mailing list, Anthony Towns addressed the issue of transaction propagation failure rates. He stated that whether or not a 5% failure rate is terrible depends on how easy it is to retry after a failure and how likely the retry is to succeed. He suggested that it's conceivable for a 5% failure rate to be detectable and automatically rectified, but there's no clear way of doing so efficiently, privately, and in a decentralized manner. David A. Harding shared some napkin math, which showed that if only 30% of nodes have adopted a more permissive policy, lightweight clients would need to connect to over 50 randomly selected nodes to ensure one transaction per year fails to initially propagate. For a more permissive policy adopted by only 10% of nodes, the lightweight client would need to connect to almost 150 nodes. Anthony Towns added that with a target failure probability of 1-in-1e8 and no preferential peering, a majority of listening nodes would need to support the transaction features in most reasonable scenarios.
Updated on: 2023-06-16T02:22:38.462020+00:00