Author: Sergio Lerner 2014-04-27 12:36:11
Published on: 2014-04-27T12:36:11+00:00
In an email exchange on April 27th, 2014, Mark Friedenbach discussed the definition of an "SPV proof" with another individual. He defined the most general definition as a non-interactive payment proof (NPP), which is not dependent on moon math or assumptions of honest peers and jamming-free networks. SPV nodes trust that the most-work chain is valid based on economic arguments about the opportunity cost of mining invalid blocks. However, Friedenbach argued that assuming honest peers is necessary to talk about the most-work chain and compute economic arguments correctly.Regarding a use case where a client starts asking for parent blocks until all parents agree, Friedenbach preferred to avoid the linear scan of block headers by using back-links and making it have log(N) space usage. He clarified that this method uses NPPs, not SPV proofs, and that going back means skipping a potential short fork. The linear scan is actually O(1), and its exact value can be approximated by the sum of the convergent infinite geometrical sequence of forking probabilities. Without considering selfish-mining, this value is about 1.03, and probably less than 2.03 with selfish-mining considered.
Updated on: 2023-06-08T21:47:05.937378+00:00