Reading Blockstream's introduction article on Eltoo and one paragraph says:

[..] in eltoo all participants share a common set of transactions, unlike LN-penalty, which requires asymmetry in which participant has access to which transactions, in order to tailor the reaction to the misbehaving party. This change eliminates what we call toxic information in Lightning. Toxic information comes from transactions belonging to outdated states, which if leaked will result in the loss of funds. This happens not only if a party misbehaves, but also if a node forgets about an update (e.g., when being restored from a backup). With eltoo this is no longer possible because only agreed-upon states can be settled (i.e., eltoo is penalty-less).

What this seems to imply to me is that it's not possible to broadcast an old channel state and have it accepted by the network, and this would seem to greatly simplify some design issues we're dealing with today. I strongly feel like I'm misunderstanding something here and would love to get clarification from someone more knowledgeable. Thanks


With eltoo you can publish old state and the Bitcoin network will accept it. (As every state is a valid and signed transaction.)

Let me recall the key idea of eltoo in short (if it's too short read the full paper it is well described there. Also I hope I remember the details correctly) : You can imagine states in eltoo encoded as a linked list of time locked update transactions. Each comes together with a settlement transaction which spends the update transaction and encods the current state. any state can be forked with a later state (within the time lock). This happens by Sorenson the update transaction with a larger update transaction preventing the precis settlement transaction to be mined. these states are known to both parties. If the state is not forked the update transaction can be settled by the settlement transactions (which are child's of the update transaction and part of the protocol)

That being said there is no need for a penalty as the tricked party can just enforce the newest state on chain if the fraudulent party tried to publish an older state.

The party still needs to monitor what happens on chain as the forking is only possible within a time lock and afterwards if the settlement tx is not published. So watchtower services might also be useful in eltoo.

One side note on the ability to omit the punishment. Peter Todd criticized the missing penalty as the game theory now encourages one to publish old state. Either the other side catches it and corrects it or you win. This would basically lead to ask stats eventually being written on chain. So even though eltoo works without penalty it might very well include one. This would however make the protocol asymmetrical again as I believe it is currently unknown how toascribe blame. This in turn would mitigate many of the advantages of eltoo

  • 1
    So eltoo might actually make watchtowers more necessary because it might decrease the deterrence that classic LN channels pose. At the same time, it would actually make watchtowers much cheaper/more scalable because the watchtower need not remember infinitely many revocation keys/txs, but only the latest update tx that can update any old state. Feb 24 '19 at 16:11
  • I would agree with this. Maybe @cdecker an weigh in? Feb 24 '19 at 20:24
  • Thanks @RenePickhardt. Could you explain further the pair of update and settlement transactions. I assume the settlement transaction is broadcast to the bitcoin network, however you also said that it "spends the update transaction", does that mean that the update transaction needs to be broadcast as well, thus requiring two on-chain transactions to close a channel?
    – tMGK
    Feb 26 '19 at 19:57
  • Also, can you offer any more insight into the whole idea of Eltoo removing penalties, and how that might affect the network. It seems that Peter Todd's concerns are very valid, since it seems to make most sense to try to steal money by broadcasting an old state since you'd have nothing to lose (except reputation & closing the channel). Has there been any discussion about this for example on the LN mailing list that you could refer to? If penalties are introduced with Eltoo, do the issues of Watchtowers needing ever growing storage, and old channel backups being dangerous resurface?
    – tMGK
    Feb 26 '19 at 20:02
  • Well, you would lose some fees, plus the things you named, and as I explained, watchtowers would become cheaper so maybe they would be universally used. So it might not be much of a problem, but we'll probably only know when we try it. There is lots of discussion about this on the LN dev list, just look for the word eltoo. And no, I don't think penalties would reintroduce the problems you named. Feb 26 '19 at 20:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.