On Friday, August 24, Ethereum's core devs and other stakeholders within the community gathered for meeting 45. Aside from updates related to testing, clients, and research, the crew discussed the upcoming Constantinople hard fork and the topic of issuance reduction (that is, reducing miners' block rewards).
Issuance of block rewards, although an evergreen, seemingly nebulous issue within the Ethereum space, took center stage at the meeting. Important to this subject is the difficulty bomb, or the gradual increase in mining difficulty to the point where an Ethereum "ice age" (or lack of mining) would occur. This works so that as the network's hashrate increases, the difficulty increases. This bomb was originally put in place to incentivize the move toward proof of stake, but as transfer to Casper continues to be pushed back, the community must decide what to do in the meantime. Apparently, simply delaying the difficulty bomb will create a sudden increase in issuance of block rewards, leading to inflation. To counteract that, developers are proposing a number of EIPs to reduce block rewards.
According to the discussion this past Friday, there are three issuance-related EIPs in competition to be included with the Constantinople hard fork. Each EIP seeks to delay the inevitable bomb rather than defuse it, but the proposed block reward amounts vary: EIP 858 would reduce the reward to 1 ETH per block, EIP 1234 would reduce it to 2 ETH per block, and EIP 1295 would maintain the 3 ETH reward. In addition to supporting different EIPs, those in attendance spoke to the nuances and complexities present in the issuance-reduction debate.
Based on the meeting, though, the two prevailing proposals appear to be EIP 1234 and EIP 1295. Eric Conner, a representative of the investor community, supports EIP 1234 because he believes reducing the mining reward to 2 ETH would better align the value of the Bitcoin and Ethereum networks.
According to Conner, Ethereum is overpaying miners to maintain its security. The Ethereum network is at 34 percent of the market cap of Bitcoin, yet Ethereum miners are paid "at an 80% rate compared to Bitcoin." He argues that, theoretically, the ratio of market cap to payment rate "should be close to 1," but that number currently sits around 2.5.
Additionally, Conner maintains that in the case of another delay in Casper implementation, it is safer to reduce the inflation rate of ETH now. Ether's inflation rate is near 7.4 percent, but Bitcoin's "sits around 4.25%."
Blockchain specialist Danny Ryan chimed in and said that the move from 3 ETH to 2 ETH seemed "like a reasonable compromise to" him. Matthew Light, who authored the EIP to initially reduce the block reward from 5 ETH to 3 ETH (which was eventually included in the Metropolis hard fork), also supports EIP 1234, stating that this proposal "feels like the most prudent response." Apparently, according to Afri Schoedon at Parity Technologies, a reduction would help maintain a "stable issuance" of rewards.
However, Brian Venturo, CTO of US-based GPU mining organization Atlantic Crypto, believes a decrease from 3 ETH to 2 ETH would "force GPU miners off the network." According to Venturo, this reduction would only leave ASIC mining rigs in the game, which require less energy and have higher hash rates. The GPU vs. ASIC debate is another can of worms in and of itself.
Venturo and his organization have instead proposed EIP 1295 to address a different problem they see with the Ethereum network: the misalignment of incentives. He maintains that the mining system is set up "to have the highest uncle rate possible," thereby increasing the rate of ETH issuance.
To give some context, an uncle block arises when two miners generate a block simultaneously. This happens when mining pools are not synchronized, which happens in part because Ethereum has such a fast rate of block creation. The uncle is the block that is ultimately rejected from the mainchain. In the Ethereum network, the production of uncles is rewarded and incentivized.
Venturo's proposal seeks to combat high rates of Ether creation through a reduced uncle reward structure that would decrease the rate of ETH issuance by 11.4 percent. This modification, in turn, would spur "investment in and optimization of ETH network node infrastructure," because "we shouldn't be rewarding subpar infrastructure," he elaborated. Additionally, and perhaps less controversially, EIP 1295 offers an alternative to addressing Ether's inflation problem: Instead of reducing rewards for miners of canonical blocks, rewards are reduced for uncles. The overall effect of issuance reduction is maintained.
Further, according to Venturo, the issuance schedule of EIP 1295 is close to that of the Casper proposal.
A handful of stakeholders went on to show their support for the EIP. Alex Thorn, director of blockchain research at Fidelity Investments, said that he and his team generally favor EIP 1295, although they "aren't ready to take a firm stand" yet. Xin Xu, CEO of the mining pool SparkPool, also prefers this proposal out of the three. In fact, Xu believes that a reduction could adversely impact the security of the network.
And Jean M. Cyr, a developer for Ethminer, indicated their preference for EIP 1295 and its "slight reduction on uncles."
The last option up for debate, EIP 858, was discussed the least. In short, blockchain developer Marius Van Der Wijden explained that this proposal's reward reduction to 1 ETH per block would help alleviate the environmental costs associated with mining operations.
Despite the attendees' deep dive into block reward reduction, there was no consensus by the end of the meeting. Many on the call did agree, though, that a decision needed to be made and that the issue should be investigated further. With a topic as complex as issuance reduction, perhaps the best conclusion is that more research needs to be done.