Author: eric at voskuil.org 2021-03-01 09:37:06
Published on: 2021-03-01T09:37:06+00:00
The discussion on Bitcoin-dev mailing list reveals that downloading relevant blocks from one node with gaps in between is possible, as only headers need to be downloaded sequentially. Libbitcoin v4 works on this principle by downloading and storing blocks in parallel. Block gaps are repopulated in case of a restart, and the most responsive nodes are selected based on standard deviation. Slowest peers are dropped, and load is rebalanced to new channels. However, pruning is not done due to its counterproductiveness, complexity, and linear growth of the chain with exponentially falling storage cost.
Updated on: 2023-06-14T18:39:32.236023+00:00