this is the miner block header signing code we implemented in wownero git.wownero.com/wownero/wownero/com…97f43d5bfb7d256125f30f522ca6154ea6e

it has been live on mainnet since 4th July 2021

it effectively killed off mega public pools, which over ran our network before the fork

but private pools and botnets are still possible

and i have no idea how dominate they are in the network compared to the percentage of solo miners

if you guys have any wild ideas on how to counter pool centralization in monero, we would be happy to test it out in our next hard fork sometime in April

protocol level? doesn't really seem likely

the only way to know you have a pool centralization problem is with a trusted oracle telling you the hashrates of all known miners

I guess, if pools all mine to constant addresses, you could tally a window of the last N blocks and see if certain addresses are winning too frequently?

but then what? ban those addresses for a time period?

then pools would just start rotating thru a set of miner addresses...

what about increasing the scratchpad in PoW to make running botnets more expensive and noticeable on unmaintained systems?

you think the pools are growing because of botnet hashrate?

we can certainly consider increasing the scratchpad, yeah

no, but they do make a large part of the hash in large pools

but it will most likely hobble the majority of PCs up to current generation

while the upcoming generation may have no trouble at all

but given the global chip crisis, getting hold of new CPUs will be difficult

Nobody is mining with very old hardware, except botnets

it'll be a different form of hardware exclusivity. maybe not as extreme as ASICs

I'm mining on a 6 year old phone...

lots of people are mining on old PCs

hyc: And it doesn’t matter vs botnet hashrate

get a billion people to mine on whatever spare phone they have at home

or, consciously decide to prevent a billion people from doing so.

I prefer the former, not the latter.

hyc: Good luck, we don’t even have million monero users 😅

with the former case, it might happen. in the latter, it definitely never will. that's not a good growth policy.

another idea increasing the size of ashing blob somehow in order to increase bandwidth requirements of running a pool

*hashing

hm, interesting.

or anything else we can crank up to increase bandwidth requirements

Bandwidth is cheap

*infinite, and latency is zero.

only to a certain point. even a 10Gbps interfacce will have a max packet-per-second rate

200g dedicated @6k

but not free and at scale, it becomes expensive

e.g. if we made the seedhash change more frequently so it was uncacheable, every job would require another roundtrip to monerod

or something along those lines

and roundtrip latencies would add up fast\

a mining pool could compensate by spinning up more monerods, but each one would eat up CPU/RAM/disk resources

Finding ways how to keep support for botnets, interesting times indeed

just increasing hardware requirement is no solution. today's firebreathing new PC is tomorrow's neglected-in-acloset-somewhere old PC

the solution to preventing botnets is fixing PC OSs so that malware can't install them in the first place

here's an idea: write a program that detects vulnerable PCs running a miner, infiltrate the system, and reconfigure its miner to use p2pool

hyc: We have forks, just update ram requirements after a few years

yeah the original plan was to re-evaluate parameters every 3-5 years

right now we're not convinced a tweak is warranted yet. next year when DDR5 systems are generally on the market, maybe