regardless of the replacement cycling attack, it's a good practice to limit the size of your routed HTLCs anyway. few are the payments that need more than 1mil sats, and rebalancers will use whatever you give them anyway and do it in parts
default node settings are to have the limit at the size of the channel itself, which is wayyyy more than needed and is just needless risk. do it right now to reduce any kind of risk, as theoretical as it might be
what implementations support configuring a limit and what are the settings? i thought could only constrain channel size and max number of htlcs in flight
htlc size is part of protocol gossip. it can be set on all implementations
If a user is using a single, non-routing channel in a LSP model, would this attack be impossible to conduct? All the examples I’ve seen have involved at least one hop.
Is there a rule of thumb to follow when resizing them from default?
