livepeer / research

Organization of the current open research tracks within the Livepeer Project
MIT License
7 stars 2 forks source link

Efficient Value Transfer - Payment Channels, Checkbooks, Probabilistic Micropayments #4

Open dob opened 6 years ago

dob commented 6 years ago

Livepeer's POC currently uses a deposit+claim payments system with custom logic for when broadcasters can withdraw their deposits and when transcoders get to draw from a broadcasters deposit. More efficient, secure, and attack-proof methods are likely available via generalized solutions like:

These solutions should be considered in not only the context of transcoding, but also for fair compensation for bandwidth in P2P data transfer as referenced in #3

dob commented 6 years ago

Related to this, but not exactly the same topic: efficient protocol-to-many fee release and payments. Each user needs to claim all the fees and rewards they're owed for each round, and this can get expensive. A mechanism like merkle-root based pool payments may be interesting to consider.

https://ethresear.ch/t/pooled-payments-scaling-solution-for-one-to-many-transactions/590

https://github.com/cardstack/merkle-tree-payment-pool

There are all sorts of challenges around making sure the proofs are available offchain, but also potential game-theoretic solutions involving staking, acceptance, and slashing in the case of invalid or non-available proofs. For example, transcoders (validators), could play an additional role in producing these proofs and accepting proposed ones, and then watchers could slash their deposits for any faults.