Published on: 2019-02-18T15:34:47+00:00
The email exchange on the Lightning-dev mailing list revolved around the issue of frequent channel updates and the impact they have on routing tables. The participants explored various potential solutions to address this problem.One of the issues discussed was the enabling and disabling of channel updates. It was noted that when a channel is closed, enabling updates are sometimes flushed immediately, resulting in nodes not receiving these updates if the connection is lost before they are sent. To prevent nodes from discarding channels as stale, it was suggested that enabling updates should be held back instead of disabling updates.The importance of routing data and the usefulness of flapping channels were also discussed. Flapping channels, which provide information to nodes that don't require all the bandwidth to stay fully in sync, were seen as valuable. However, it was acknowledged that optimizing traffic efficiency and reducing unnecessary updates is crucial.In another email thread, set reconciliation schemes for reconciling channel parameters were discussed. Initially, a proposal was made to reconcile [channel id | timestamp | checksum], but it was pointed out that if the timestamp changes while the checksum remains the same, further communication rounds would be required to determine that the channel parameters haven't actually changed. Instead, it was suggested to reconcile [channel_id | checksum] and share the set difference if there is a different set of channel parameters.To address the issue of repetitive updates with new timestamps and signatures, it was proposed to locally update the timestamp and signature for the old channel_update instead of forwarding a new update. This would help reduce redundant information in routing tables.Regarding the synchronization of routing tables, it was hypothesized that adding a basic checksum to channel queries could significantly improve the process. This checksum would ensure that redundant information is not downloaded, making the syncing process more efficient. However, further data and analysis are needed to validate this idea.Overall, the participants emphasized the importance of optimizing network structure, improving traffic efficiency, and enhancing the resiliency and redundancy of the Lightning Network. Solutions such as locally updating timestamps and signatures, implementing local policies to delay channel_updates, and using a basic checksum for syncing routing tables were proposed. Importantly, these solutions can be implemented without requiring any changes to the Lightning Network specifications.
Updated on: 2023-07-31T21:21:50.166738+00:00