Open TheBlueMatt opened 4 years ago
See ongoing ML discussion : https://lists.linuxfoundation.org/pipermail/lightning-dev/2020-February/002505.html
And https://lists.linuxfoundation.org/pipermail/lightning-dev/2020-February/002513.html
On Wed, Feb 12, 2020 at 10:11 AM Antoine Riard notifications@github.com wrote:
See ongoing ML discussion : https://lists.linuxfoundation.org/pipermail/lightning-dev/2020-February/002505.html
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rust-bitcoin/rust-lightning/issues/495?email_source=notifications&email_token=AACPUA6OJGFIJQM7DRIWM7TRCQ3WFA5CNFSM4KT54KBKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOELRZQ4Y#issuecomment-585341043, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACPUAZIMJAR5BRBY4IFFCLRCQ3WFANCNFSM4KT54KBA .
Is this sufficiently covered by https://github.com/lightningdevkit/rust-lightning/pull/1359, or do we want even more randomization?
I don't think so - we're still pretty aggressive in practice about using the best-score path, which is pretty bad. It'd be really nice to get more randomization in somehow, but I'm not sure what the best approach is to doing that, maybe when we have some free time (hah!) we can look at providing a second router that works more like CLN's.
We should do (or have an option to do) something much more private than simply "cheapest path" wrt route selection. We may want to take inspiration from c-lightning's "select random route out of a bunch of random-biased route selections" approach, or do something else. We should probably check with some academics as to the latest thinking on this one.