Today, the Ethereum Core Dev team held a meeting to discuss lowering the block reward given to Ethereum miners. The primary motivations for this seem to be:
- The cost to the network in the form of inflation
- The environmental impact of energy spent on mining
To their credit, the team invited a number of miners of varying sizes and scales to give their input on the various proposals on the table. One such person was Xin Xu, who represented a large, China-based mining operation. He made this point [1:20:07]:
I want to point out...it’s more of a network security level. I posted a link before in the chat about crypto51.app, which is a website that helps you to calculate how much it costs to actually, do a, do a 51% attack to the network. For Ethereum, it’s not that much. However, why is there no one attacking the Ethereum network?
I think from my perspective, or our study as a whole team, is there isn’t any hashrate you can use in the market to attack the Ethereum network…. So, like there are a lot of people [saying] on the internet, “What if we cut the issuance? How much hashrate will go out?” However, at least to me, it’s not like if you’re going to cut [the block reward from] 3 to 2, only 1/3rd of the hashrate will go out, or maybe we’ll cut the Ethereum network’s security to 2/3rd.
It’s more like now we have 100 soldiers, and that’s all the soldiers we have in the world. But if we cut 40 of them out, then it’s not about if the enemy or the bad people have enough soldiers to attack us. Because before they had 0. Now they have 40.
They have the chance, once they have the capital, they always have the choice to hire those soldiers, or those hash rate, to attack, then. So my major point is, the Ethereum network has been secure or safe to now. It’s not because attacking Ethereum costs too much, it’s simply because now all the hashrates are on the Ethereum network, and no one has extra hashrate to attack it. So even if you have enough money, it’s not going to happen.
The issue he brings up is rather interesting. Essentially, he's arguing that attacking the Ethereum network is difficult today primarily because the world is hardware constrained. There literally aren't enough GPUs for you to buy, or hire, to achieve the hash rate needed to attack.
Thus, when considering the security effect of forcing some percentage of miners off the network, one can't simply look at the hash rate reduction. One also has to consider that a bunch of GPUs suddenly become available for attack that weren't before.
I'd take this a step further and point out that new GPUs will continue to be manufactured. If these are no longer profitable for mining, they won't be bought up for mining, and will become less expensive and more widely available. This makes it significantly easier and cheaper for a would-be attacker to acquire the hardware that would actually be needed to attack the network.
In short, Xin Xu is making the case that considering security to be equivalent to the hash rate is naive. The network hash rate must be put in context of the global supply of GPUs and other mining hardware, and that supply is itself effected by the issuance rate.
To my surprise, no one engaged Xin Xu on his point. In fact, the very next person to speak was Carl L., who had this to say [1:22:14]:
The point I wanted to make was that, going forward, I think we can sort of concern ourselves with the current hashrate and the current miners and their running efficiency. In order to….you know, do we need additional hashrate? Like are we talking about wanting to continue to grow the hashrate? Is that desirable?
In my mind, that’s absolutely not desirable. And in that case, looking at the kind of equipment cost, we want to be discouraging that equipment, that investment in new equipment. And one way to do that is to, you know, to lower the block reward, and that you know, worsens the economic argument for that investment.
In case it wasn't crystal-clear that discouraging investment in mining was an explicit goal, Tim Coulter had this to say [1:25:28].
If people are continuing to invest in mining, then I completely agree that we should de-incentivize that.
As I've said before, I'm rooting for the Ethereum Foundation to achieve it's goal and bring a secure, truly decentralized PoS network into the world, even though I remain skeptical.
While we wait for PoS to materialize, it seems prudent to tread carefully with regards to network security. I was disappointed today that no one engaged Xin Xu on this discussion of the second-order effects a reward reduction might have on hardware availability, and thus network security.