Author: Fabrice Drouin 2019-01-08 17:22:23
Published on: 2019-01-08T17:22:23+00:00
The email thread is about the problem of flapping updates in a network and how it can cause inconsistencies in routing tables. The discussion revolves around the issue of duplicate updates, and whether they should be discarded or remembered for future reference. There are concerns about nodes not being able to detect when a channel has been re-enabled if updates are discarded, and the possibility of nodes discarding channels as stale if they don't receive new updates when they reconnect. The email suggests different heuristics that could reduce gossip traffic and improve routing efficiency, such as "eager-disable" and "lazy-enable" policies. Ultimately, it is acknowledged that achieving consistency in views of the routing table may not be possible, and that the protocol should be considered a best effort gossip protocol.
Updated on: 2023-06-02T16:52:06.462123+00:00