Author: Matt Corallo 2012-09-10 19:53:33
Published on: 2012-09-10T19:53:33+00:00
In a message thread from 2012, Matthew Mitchell expresses his skepticism towards segmenting blocks as a way to improve block relay times. He argues that it wouldn't make much of a difference unless the block sizes were unrealistically large, and even then, relaying a list of transaction hashes and transactions separately would be more efficient than using segments. Mitchell also doesn't see blocks ever becoming that large and believes that upgrading protocols for the few full nodes remaining in such a scenario would not be difficult. Instead, he encourages focusing on decreasing block relay times for the current network. In response, Luke-Jr suggests that most of the block propagation problem lies in implementation rather than protocol. He acknowledges that distributing missing transactions on an as-needed basis could be a possible improvement at the protocol level but there hasn't been enough research to determine if the benefits outweigh the costs. Luke-Jr questions why six new messages are needed instead of adding a single new type to getinv. In his earlier message, Mitchell discusses the six new messages proposed, including getseginv/seginv, gettreelevel/treelevel, and getsegment/segment.
Updated on: 2023-05-19T04:11:50.162395+00:00