Despite Core Dev Agreement To Move Forward With ProgPoW, Dissent Continues
Concerns over the efficacy of the new algorithm – and that the move could impede ETH 2.0 developments – have fueled continued debate. Some even worry ProgPoW could lead to a chain split.
Core developers came to consensus during their call last Friday around the decision to implement ProgPoW. ProgPoW, or programmatic proof of work, is an alternative algorithm to the existing Ethash; it is designed to close the performance gap between ASIC and GPU mining rigs.
While debate about ProgPoW has featured in almost every core devs meeting since late August, discussion had cooled to a low simmer until Friday, when it was decided with more certainty that Ethash should be replaced by ProgPoW. That’s when things started to heat up again.
Let’s step back and look at the context.
ASICs are specialized, efficient mining machines. ASICs, produced by only two manufacturers, Bitmain and Ethmaster, have somewhere between a 20 and 50 percent efficiency advantage over GPU rigs. The catch: Although ASICs cost less to run and use less energy, they are much more expensive to purchase. The prohibitive cost and limited number of manufacturers has raised concerns that ASICs could make GPUs obsolete and run the risk of consolidating power to mining pools and rig manufacturers, a result that runs counter to the egalitarian ethos of Ethereum.
The debate about whether to implement ProgPoW – and if so, when – has been ongoing for some time. The Ethereum Foundation’s Piper Merriam posted EIP 958 for ASIC-resistance back in March of 2018, and Kristy-Leigh Minehan first posted EIP 1057 for ProgPoW back in May – but ASIC-resistance has been a topic of concern since the Ethereum white paper. In fact, the Ethash algorithm was designed with ASIC-resistance in mind.
ASICs have long been an object of ire, but that hasn’t eliminated friction from efforts to advance ProgPoW. Over the last few months, topics of debate around ProgPoW have included whether the change would be effective, time efficient, and desirable, as well as whether it should be included in the Constantinople hard fork. Before Devcon, the consensus seemed to be that, regardless of whether ProgPoW should be implemented, attempts to bundle the upgrade with Constantinople would only further delay the long-awaited hard fork.
It wasn’t firmly decided that the network would fork for ProgPoW, but the last few core dev meetings featured more complex conversations around ProgPoW development efforts and challenges, not simply the question of whether it should be implemented.
The controversy over ProgPoW implementation was rekindled by the announcement that a decision was to be made regarding a ProgPoW fork. Now that tentative consensus has been reached, voices of dissent have reemerged.
Without ProgPoW, Some Worry Constantinople Spells Bankruptcy for Miners
Ahead of the call, Ikmyeong Na, a ProgPoW contributor and self-proclaimed “voice of the miners,” began posting on the Github agenda page about his concern that failure to include ProgPoW in Constantinople would lead to a significant reduction of Ethereum nodes. He said, “[W]ithout ProgPOW bundled with EIP-1234 it will be only disaster with GPU miners and I expect more than 30% of ethereum miners will be bankrupted.”
EIP 1234, one of the five EIPs to be included in Constantinople, lowers ETH issuance from 3 ETH per block to 2 ETH per block, thereby reducing miner rewards – a significant problem for GPU miners that already struggle to compete with ASICs.
Despite Na’s concerns, the block associated with the Constantinople hard fork is set to be mined on January 16. There isn’t much room to negotiate this timing. As Eric Conner, voice of the Into the Ether podcast pointed out on Twitter, Ethereum‘s difficulty bomb “kicks in March 1,” at which point block issuance will significantly decrease. “With the difficulty bomb coming there is no incentive, even for miners, to stay on the current chain,” he says.
Even though issuance will decrease with the implementation of EIP 1234, post-Constantinople issuance will still be significantly higher than the existing chain after the difficulty bomb kicks in.
Though the course has been charted, some in the community share Na’s concerns, including the security lead at the Ethereum Foundation, Martin Holst Swende, and Andrea Lanfranci of SparkPool. Nonetheless, Swende and Lanfranci agree that ProgPoW has not been sufficiently implemented and tested to be deployed so quickly.
Some Worry ProgPoW Won’t Stop ASICs
Concerns also surfaced in the Fellowship of Ethereum Magicians forum that ProgPoW would only stave off Ethereum-specific ASICs for a year or so, before even more capable ProgPoW-resistant machines are developed.
David Vorick, co-founder of Sia, raised this issue in an ETH Magicians thread. “If the ethereum community at large greatly opposes the development of specialty hardware,” Vorick said, hardware developers will be forced to either “develop hardware that will be kept secret and released to specialty groups only,” or not develop hardware at all. “And while some manufacturers may choose [the latter] the amount of money at stake essentially guarantees that at least some hardware group is going to choose [the former].”
With this possibility in view, Vorick asks the Magicians, “[I]f a hardware developer manages to create a ProgPoW ASIC that outperforms GPUs by a surprising margin, let’s say 10x or even 100x, is it better for that manufacturer to keep their discovery secret and mine secretly, or is it better for that manufacturer to sell openly?”
Others Think ProgPoW Is a Distraction from Proof of Stake
Over the weekend, Twitter buzzed with debate around whether ProgPoW represents an efficient use of limited resources. Martin Köppelmann, CEO and co-founder of Gnosis, chimed in to express his opposition to ProgPoW.
Johann Barbie of
LeapDAO agreed. “To protect decentralization, steps towards PoS are more constructive than a likely contentious ostracism of ASIC miners,” he said.
Georgios Konstantopoulos, lead researcher of the Loom Network, posted in support of Köppelmann’s dissenting opinion.
Jorge Izquierdo, co-founder of Aragon,
echoed Konstantopoulos’ opinion: “I’m not interested in debating whether it is a good idea, just the fact that it shouldn’t be a priority at all for Ethereum in 2019. Let’s put all our energy towards POS pls.”
The founders of Harbour Protocol, Conner and Dean Eigenmann, even suggest that ProgPoW could lead to a chain split. The concern that ProgPoW distracts from proof of stake, or that it risks a hard fork is a minority position, despite its support by prominent thought leaders in the Ethereum development community.
Philippe Castonguay, a researcher at HorizonGames, asserts that ProgPoW would in fact decrease “the probability that a significant contentious fork will happen when moving to Proof of Stake.” He reasons that if expansion of ASIC mining continues, by the time PoS comes, ASIC miners will dominate the market. Because these mining rigs are so expensive, it will “increase the cost for miners to move to PoS,” and therefore increase “the probability of a significant fork.”
However, Castonguay also points out, “[T]his hypothesis is only valid if ProgPoW does indeed significantly reduce ASICs advantage over general computing hardwares like GPUs. If that’s not the case, then moving to ProgPOW would not be helpful.” As discussed above, it’s not clear that ProgPoW will significantly decrease the advantage of ASIC rigs, especially in the long term.
As for distracting from proof of stake, Castonguay argues, “It’s a false dilema [sic] that implementing ProgPOW means not focusing on ETH 2.0. People working on ProgPOW are not the same … ones working on ETH 2.0.”
It is unclear if the current level of dissent is enough to prevent a ProgPoW hard fork. The debate has been ongoing for months. Developments have already been under way for some time, and a testnet is already launched. In Friday’s core devs meeting, some estimated that ProgPoW could happen as quickly as February or March. Ethereum developments are notorious for arriving later than expected, but the short timeline suggests that concerns that ProgPoW is a distraction may be somewhat hollow – and that time for debate has ended.