Author: Gregory Maxwell 2014-04-11 16:54:44
Published on: 2014-04-11T16:54:44+00:00
In a conversation from 2014, Tier Nolan and Peter Todd discussed the idea of error correction in the Bitcoin network. While Nolan mentioned it as a joke, he acknowledged that it could be an over-design for the basic protocol. He also noted that storing "random" blocks can have locality problems when verifying blocks, and that nodes with random spans of blocks could work out to the same failure rates as random for the non-error-coded case. Nolan suggested that diversity in p2p protocols would be good for the network, and that someone should develop an alternative protocol even though these techniques may not yet be compelling enough for the basic protocol.Todd shared a link about using forward error correction for faster and more efficient block relaying. Nolan added that suitable error coding can also provide an anti-censorship effect, making it difficult to provide part of the data without potentially providing all of it. Finally, they discussed the odds of a block not being stored if there were 10,000 nodes each storing 0.1% of the blocks at random. Nolan believed that the network already had a reasonable number of nodes that store all the blocks, which is sufficient to supply blocks that have otherwise gone missing.
Updated on: 2023-05-19T18:35:20.804094+00:00