From alejandro.ranchal_pedrosa at etu.upmc.fr Wed Apr 11 08:43:28 2018 From: alejandro.ranchal_pedrosa at etu.upmc.fr (Alejandro Ranchal Pedrosa) Date: Wed, 11 Apr 2018 10:43:28 +0200 Subject: [Lightning-dev] An Idea to Improve Connectivity of the Graph In-Reply-To: <878ta034mm.fsf@gmail.com> References: <878ta034mm.fsf@gmail.com> Message-ID: Hi Christian, > That's not as bad a tradeoff as people usually interpret, the DMC > construction has parameters that allow tweaking the number of > invalidations, and with parameters similar to LN we can have 1.4 billion > updates. Which is years of operation without need to > re-anchor. In addition penaltyless invalidation has a number of > advantages, As far as I understand, long-lasting DMCs require either: ??? (a) an initial Refund transaction with a very distant relative locktime ??? (b) periodic updates in the form of a Refund transaction pointing to a new Refund transaction resetting initial the locktime, instead of actually refunding. ??? For an extreme case of (a), if one party goes unresponsive and decides not to sign new commitments then the counterparty in the DMC will have its funds locked for a significant amount of time, without penalising the unresponsive party. In the extreme case of (b), either if as a result of a malicious, unresponsive, or honest participant, each new refund transaction that resets the refunds may end up hitting the blockchain, which means the worst-case utility of the channel itself decreasing due to accumulative blockchain fees. Is this the trade-off you speak of? if so, can you point at any resource where this trade-off is tackled to get worst-case utility similar to that of LN channels? Best, Alejandro.