How Bob gets his money in NTumbleBit

[osagga]

• After Bob finishes the Promise Protocol here, he gets a puzzle that needs solving.

• He passes the puzzle to the Solver protocol (Alice) here.

• Alice runs the Solver protocol to get the solution for the puzzle here.

• Now the Solver protocol can end in two ways:

  • If the Tumbler is cooperative (The Tumbler sends the solutions k_j For all j in R before it closes the channel):
  • Alice gets the solution keys from the Tumbler directly here
  • Else if Tumbler is not cooperative (Tumbler disconnects right after Alice posts the T_puzzle):
  • Alice gets the solution keys from the Blockchain (From the tumbler's fulfillment transaction with the pre-images) here

• Alice runs "CheckSolutions(solutionKeys)" to find at least one working solution from the obtained keys here.

• Bob takes the solution key from Alice and runs "GetSignedTransaction(solutionFromAlice)" (here).

  • This function takes the signature (solution) and applies it to the hash of the cashout transaction and returns a transaction that is ready to be posted on the blockchain.
  • This function works as follows (from here):

• puzzle = Unblind(blindedPuzzle)

• epsilon = solutionFromAlice

• for i in range(RealTransactionsCount): #Mu

  • h = RealTransactionsHashes[i] # equivalent to h_t_i from paper
  • if (i == 0): quotient = 1, otherwise: quotient = QuotientValues[i-1] #From step 10 in the paper
  • epsilon = epsilon * quotient mod(N)
  • signature = MGF1(epsilon) XOR promiseOfPuzzle_i
  • if (ECDSASignature_Verify(TumblerKey, signature, h)):
    • Add signature to transaction T_cash
    • Return T_cash
  • else: continue

    
    Now after Bob gets the signed transaction, he posts it to the Blockchain [here](https://github.com/NTumbleBit/NTumbleBit/blob/2f92176bbcd931615299b775b53d99f590140fdf/NTumbleBit/ClassicTumbler/Client/PaymentStateMachine.cs#L485) and [here](https://github.com/NTumbleBit/NTumbleBit/blob/2f92176bbcd931615299b775b53d99f590140fdf/NTumbleBit/ClassicTumbler/Client/PaymentStateMachine.cs#L444) (depending if tumbler is cooperative or not).

[goldbe]

Ok. Wow.

For the q version of the protocol we have a LOT more work to do, unfortunately. I wonder if maybe we should just build the cooperative version of the protocol during your internship and defer the noncooperative version for later.

[osagga]

@goldbe Hopefully we can finish this in time. I think we will have to start by designing a Q version of the Solver protocol. Then maybe decide how would Bob spend the puzzles. What do you think?