Author: Jeremy 2021-07-11 00:02:35
Published on: 2021-07-11T00:02:35+00:00
In a discussion on improving the sequence space for Eltoo channels, Jeremy Rubin proposed two solutions. The first solution involves creating a Taproot tree with N copies of the state update protocol, where the 1st copy is used until MAX_STATE is reached and then signing with the 2nd copy while removing the 1st copy from the tree. This allows for a random path through which leaf is being used, which increases privacy. However, this approach has a downside as it achieves only 5 bits of added sequence space compared to the prior approach that achieves 64 bits with 2 txns. The second solution involves using a continuation state, where for the final update, one signs a move to a new update key which essentially allows for 32 bits worth of updates before forwarding to a new contract by paying an extra transaction. Using this trick for every update would provide 64 bits of sequence space, but would also increase resolution time linearly. To mitigate this, Rubin suggests doing a "semitrusted burst mode" with a counterparty, where parties move to sequence M+1 and enter a burst with 32 bits of sequences to blow through. Then, they indicate to the other party to terminate the burst at "internal state number" Q and sign M+2, reflecting the last state at internal state number Q, to get rid of the temporary extra locking time for when parties are offline. This solution also provides a benefit for privacy as top-level state numbers do not strongly reflect the number of payments made. However, entering burst mode with malicious peers can result in double funds lockup time. To mitigate this, Rubin suggests only entering burst mode with long-lived peers, when the initiator has more funds in the channel than the other party, or when only a certain percentage of liquidity is moved during a burst.
Updated on: 2023-06-03T05:07:37.221191+00:00