Monero’s hashrate has seen a significant spike since last month, nearly trippling within days to now not far off from all time high.
“More than 85% of the current Monero Hashrate is ASICs and each machine is doing 128 kh/s,” a report says.
They analyze a random number that is generated with each block, called a nonce. Looking at patterns and through some calculations, they conclude the network is effectively being run by asics.
The findings led to much discussion in Monero, with some suggesting the algorithm should perhaps be changed every three months to make the application specific integrated circuits (asics) non usable any longer.
Previously they decided to change the algo every six months, but it appears manufacturers have been very quick at getting out new asics, leading to a potential dilemma.
To Asics or No Asics?
Forking the network every three months might cause significant disruption. If nothing else, developers would not be able to work on much else but the algo change and related testing.
That would bring any innovation to standstill. Something the privacy focused coin might not afford as it faces challenges from Zcash and other competitors.
In this fast moving space, standing still might mean not standing at all. On the other hand, one of the unique proposition of this coin is that you can mine it on your laptop without even needing a GPU.
That has given it some bad reputation due to claims that much of the CPU mining is on botnets. There have been numerous reports of malware being installed on computers with the sole purpose of mining monero.
On the other hand, some news sites have tested monero mining as a potential alternative to ads or paid subscription. Unicef too has launched a site where people can donate CPU which mines monero, with 27,000 people donating at the time of writing.
However, according to Riccardo Spagni, monero’s lead dev, the fight against asics is only a temporary measure, rather than a permanent proposition. He says:
“All that’s happening is Monero is stalling ASICs, not preventing them. Hopefully it’s stalled until ASICs are commoditised!”
The plan therefore appears to be just delaying an asics run network until a future time when asics become so widespread that one can buy them at their supermarket or computer store.
Asics, Inevitable in Proof of Work?
The bitcoin whitepaper has a much quoted paragraph where Satoshi Nakamoto, bitcoin’s inventor, says one CPU is one vote.
That only lasted for about one year with a coder proposing a method to effectively game the Proof of Work (PoW) algorithm by using GPUs which can give the miner significantly more hash. FPGAs then followed, which are programmable asics, then the specialized asics themselves which are so refined as to do nothing else but solve the PoW calculations in the most efficient way.
Numerous attempts followed to go back to that one CPU, one vote, but all have failed so far, with monero now seemingly settling for just a delaying tactic.
That wasn’t their stated plan a year ago when they changed the algo for the first time to brick asics. The implication back then was that the threat of the algo change would be sufficient to prevent the development of new asics.
The threat, however, had a counter-threat. An 80% hashrate drop after the algo change fork in spring 2018 was quickly followed by 21 block deep reversals of history in one of the first 51% attack of a major coin.
Riccardo Spagni, Monero’s lead dev, tried to dismiss the attack at the time, but following reports that the cryptocoin is now back to 85% asics controlled, Sagni said:
“I’m not involved in the discussions about changing the PoW algorithm.”
They claim their plan is to change the algorithm every six months, but when it was changed in October there was hardly a real drop in hash.
It did fall by 50% within five days as can be seen, but within two weeks it was back to near pre-fork levels. Suggesting, as far as data goes, that perhaps there wasn’t any real algo change as in the spring fork, hash dropped by close to 90%, with 50% of it disappearing for months.
Now the asics are back, and instead of a brash Sagni riling against them as he did last year, we have a muted dev who claims he effectively has nothing to do with changing the algo.
So raising a fundamental question that in many ways was at the root of the blocksize debate alongside many other considerations.
To fork can be a very easy affair. You just write some code, and now you have your own network. To get anyone to recognize or even know your network exists, however, is a pretty difficult endeavor. To then get others to name your fork as monero or any other main coin can be close to impossible.
That’s for anyone who is not a prominent dev of the coin in question. If they are a lead dev, then the name goes with it as if by default.
In bitcoin, whatever Blockstream says is arguably bitcoin. In ethereum, whatever Vitalik Buterin says is ethereum. In monero, whatever Sagni says is monero and in bitcoin cash, whatever Amaury Sechet says is arguably BCH.
In all of these coins they all claim that’s not the case and in most circumstances they’re probably right, but dev power is a recognized potential weak point in non completely black and white cases.
Ideologues in bitcoin, therefore, have this idea of no forks, no upgrades, no changes. Effectively a frozen piece of code.
In ethereum too they dream of a time when changing the code would be practically impossible because there would be so many customized systems or old systems that perhaps no one even knows how to change them or changing them requires huge resources. So freezing the blockchain in stone.
The answer in bitcoin (or subversion depending on your view) is backwards compatibility or soft-forks. Basically, old nodes can keep knowing after the upgrade everything they knew before the upgrade, but they can’t know or can’t utilize the new features.
This can allow for changes, and sometime quite fundamental changes. Like the blocksize which was changed in bitcoin through a mathematical trick that makes the old nodes think it wasn’t changed.
The idea, however, is that you can’t change fundamental aspects, like the 21 million limit, but whether that actually applies in practice is unclear.
For now, there is much work and many improvements that have to be incorporated. Making developers a privileged class in the blockchain space, kept in check by only other competing projects.
Yet it is probable that in twenty or thirty years, at least one blockchain will reach the stage where there are no “real” improvements save for “cosmetic” modifications here and there.
At that point, the blockchain might be set in stone. As it stands, when to fork and for what reason is significantly influenced by what the lead dev prefers, at least where such choices matter.