3 ms·

This is a very nice primer but I felt like it glossed over the issue of transaction fee incentives and why grouping transactions into a block is sufficient to e

by wsxcde 12y ago

This is a very nice primer but I felt like it glossed over the issue of transaction fee incentives and why grouping transactions into a block is sufficient to ensure security. Thinking out loud here, how does a transaction fee that is less than the transaction value lead to a secure system? Suppose we have N transactions in a block each with value: value_1, value_2, ... , value_N and corresponding fees: fee_1, ..., fee_N, are we making the assumption that (fee_1 + fee_2 + ... feeN) >= value_i for each i? I guess not and I suppose we are relying on the fact that even if the above isn't true, a dishonest transactor/miner would not able to attack the system because s/he doesn't have 51% of the network? If so, then it appears there is a somewhat surprising relationship between the largest transaction that can be securely carried out on the network and the hashrate of the network. Because when a transaction above than a certain value needs to be carried out, it might become profitable to just buy up enough hashpower to be able to reverse it later! Is the above train of thought logical or am I missing something? And if all of the above is correct, I'd be curious to know what an estimate of the largest safe transaction for the current bitcoin network is.

patio11 12y ago
The assumption is that that a) the amount of hash power one would waste to tinker with transactions retroactively is prohibitively expensive when you control less than 50% versus just cooperatively mining and b) miners care about the price of Bitcoin enough such that the short-term returns of tinkering with any number of transactions are unattractive compared to the massive hit they take when it becomes apparent that Bitcoin is insecure and the price crashes. I am reporting widespread beliefs rather than endorsing them. Assumption A is problematic if you believe selfish mining is practical. Assumption B is problematic because the history of the largest pool using double spends against a disfavored merchant without meaningful consequences contradicts it.
- wsxcde 12y ago
  I hear what you are saying. What I'm interested in exploring the consequences around rational behavior should if you make the assumption that (a) each unit of GH/s has a certain cost, (b) the miners are expecting to recover this cost, (c) for an ethical player this cost can be recovered through the sum of the block reward and transaction fees, and (d) an unethical player could also attempt to recover the mining costs by performing a double spend attack. My question is for what transaction values does (d) become a viable and rational strategy. The reason I ask this question is because mining is hugely more expensive operation in comparison to a centralized transaction processing system. It is way more inefficient to have hundreds and thousands of computers chugging away trying to shove a SHA puzzle than it is to just do a simple atomic database update. So somebody must be paying for this and the question is who? And what is all this really costing them. And from this, the next question that follows is what is the impact of the mining reward/cost on what can be secured by the network? So that is the question I was really trying to get at. I'd be interested in hearing your thoughts on this or pointers to analyses along these lines.

by wsxcde 12y ago

patio11 12y ago

The assumption is that that a) the amount of hash power one would waste to tinker with transactions retroactively is prohibitively expensive when you control less than 50% versus just cooperatively mining and b) miners care about the price of Bitcoin enough such that the short-term returns of tinkering with any number of transactions are unattractive compared to the massive hit they take when it becomes apparent that Bitcoin is insecure and the price crashes. I am reporting widespread beliefs rather than endorsing them. Assumption A is problematic if you believe selfish mining is practical. Assumption B is problematic because the history of the largest pool using double spends against a disfavored merchant without meaningful consequences contradicts it.

wsxcde 12y ago
I hear what you are saying. What I'm interested in exploring the consequences around rational behavior should if you make the assumption that (a) each unit of GH/s has a certain cost, (b) the miners are expecting to recover this cost, (c) for an ethical player this cost can be recovered through the sum of the block reward and transaction fees, and (d) an unethical player could also attempt to recover the mining costs by performing a double spend attack. My question is for what transaction values does (d) become a viable and rational strategy. The reason I ask this question is because mining is hugely more expensive operation in comparison to a centralized transaction processing system. It is way more inefficient to have hundreds and thousands of computers chugging away trying to shove a SHA puzzle than it is to just do a simple atomic database update. So somebody must be paying for this and the question is who? And what is all this really costing them. And from this, the next question that follows is what is the impact of the mining reward/cost on what can be secured by the network? So that is the question I was really trying to get at. I'd be interested in hearing your thoughts on this or pointers to analyses along these lines.