cashu is a centralized IoU and also “offline” payments could be fake until you get online
Discussion
Sure, but at least you have some chance of getting that donation, since donors don't tend to hang around and attempt 15 times. Different market than with direct payments to a supplier or something.
I'd like the ability to have a secondary LN wallet, so that I could keep the self-hosted one up, and it automatically falls back to #Minibits, if the payment doesn't go through or the wallet is offline, or something.
the subject of resilience of nodes on LN has loomed big over my horizon just now thinking about this... this is something that could be added to LN to make it simpler to run multiple peers replicating the channels on different IP addresses that can pick up the slack if the primary goes down
of course, this would also be needed for cashu mint nodes as well, though they don't have to p2p it exactly, they have to signal it to wallets
yeah, i think it's a false argument that it solves the liveness problem of lightning at all, i'd say the opposite, cashu mints are probably more likely to be of lower quality (at this point) and even in ideal circumstances, they are just as prone to downtime, at least, as a lightning channel (which is one or both nodes in a channel down) which even sorta suggests that the cashu mint may be actually less stable, since where lightning channel failure is a 1 out of 4 probability the offline probability grid for cashu mint is 1 out of 2 (because it's only one)
oh, sure, you could have a dns round robin or you could have a loadbalancer but that itself is a centralization vulnerability, in both cases
i think that a strategy for failover for lightning nodes would be a great thing to have exist... such as a backup that heartbeat broadcasts proof it can maintain a given channel end and the peer falls back to it in such a case. and the backup can act as a mirror, just needs a sync protocol, and it could act as a signed broadcast message
when i was working with the Steem protocol there was this possibility to set up a peer to sign over to another peer to cover for a particular pubkey that was in the stake-to-validate race, i could cook up such a cryptographic scheme so that it is even less noisy, one node syncs state to another(s) and a protocol for a takeover message to move the channel addresses to a new IP
I can't really follow this, but I like the idea of a backup node.