Group Questions

  1. How have show-stopping bugs or disruptions to the network been handled in the past? document with Sebastian’s notes on this.

  2. At what point should a bump to the min cumulative work, or the last checkpoint height be considered?

    This means are the checkpoints in the right place? I think they might be pretty much useless now that we have headers-first IBD which can be cheaply cross-checked with multiple nodes. But not sure on this…​

  3. What is a Sybil Attack, and how has it been solved in the past? How does proof of work enable Sybil resistance for new nodes joining the Bitcoin network?

    Sybil attack in distributed protocols is where the attacker is generating fake identities to overwhelm the honest node. If the protcol is operated based on the number of participants, a quorum, then as long as we have more than a quorum of honest node then the network is safe. In Bitcoin anyone can join or leave or make new identies whenever they want. This means an attacker might be able to easily dominate the network.

    To defend against this type of attack bitcoin used PoW which does not rely on number of honest nodes, but how much work is done by honest members. New identities cannot easily be used to generate new identities on the network — you need to expend real energy with economic cost.

    When new nodes join the network, they first learn history of the chain before starting new work on the most-work chain. A malicious attacker cannot easily overtake a new node who has the most work chain already.

  4. What is 'network Sybil attack'? Is Bitcoin safe from it? If not, how can we make Bitcoin more resistant to that?

    "Network" here refers to sybilling the networking stack. Erebus attack is one example of this type of attack. As we have 125 incoming connections there is some scope for these to get sybilled.

  5. What do you think will happen when the mining reward runs out?

    Technically this question is incorrect; presumabley it means to refer to block subsidy and not "mining reward", as block reward = block subsidy + fees.

    On the technical side there is not too much interesting to explain, we expect that an organic transition between block subsidy and block reward (fees) will occur. According to the onion model of blockchain security this has some potential to affect the "Economic guarantees" layer of security in the case that fees do not compensate as well as the previous subsidy + fees did.

    However there are some counter-points to this, namely the fact that because the subsidy is removed logarithmically, there is no sudden transition to this state. In face to date, 89% of coins have already been mined and fees already account for approximately 9% of block reward.

    Bitcoin’s block space is a scarce and unique commodity. It will continue to accrue demand and blocks should remain full indefinitely now.

    Regarding developers changing the fee "weight" for certain transactions a-la SegWit. We thought that this simply increased transaction density slightly and that, unless some improvement suddenly offered multiple orders of magnitude improvement that blocks would continue to be full of fee-paying transactions.

    Finally, bitcoins incentives are setup so that hash power should roughyl correlate with the value of the network, not with the mining subsisdy and we don’t see this changing after the subsidy runs out.