yeah, lots of future timestamp fuckery going on. At least testnet4 requires one asic block per difficulty period, that at least prevents the blockstorms like on tn3.
Discussion
Yeah I think we can do better. My block erupter from 2013 is getting hot! I feel there's too many spoofed timestamps. But at least it's testnet so we can war game some defenses. Ive been writing a stratum server and client in the last couple of days. It's almost working but the hex formats can be tricky. I think the community should get together and try to defend #testnet4, even with CPU, but if someone has a bitaxe, that would be huge.
I don't think its a matter of attacking/defending testnet. The root cause is the '20 minute rule' active on testnet3/4. This can be abused (spoofing timestamp up to 2 hours), but it's there for a reason: to not stall the entire chain. Coins have no value, so normal miner game theory doesn't apply, because miners come and go while testing their software/hardware. If there was a better alternative to the '20 minute rule', we can just launch a testnet 5 with different (better?) rules.
I get the 20 minute rule. About 2 months ago basically the chain would move along nicely and if it stalled for 20 minutes wiz would mine a diff=1 block. PortlandHodl has an asic so he donates some hash to keep it ticking along. But spoofing 2 hours forward, I dont think is in the spirit of things. And what seems to be happening is that every diff adjustment we are getting high % increase (double digit or 90% sometimes). I think its good to play around on, because we are going to have a lot of problems when the block subsidy goes away, so this gives us some early warning of what sort of things can be done. If that makes sense.
Not sure if testnet could be indicative of what happens on mainnet (security budget wise). Because there is no financial incentive on testnet, miner game theory is quite different between those two.
Just thinking out loud:
Would a testnet without 20 minute rule work, now we have a lot of 'hobby' miners also? Maybe not, because difficulty stays too low, and an adversarial miner with enough hashpower can cause huge reorgs.
What if the difficulty adjustment would not take into account difficulty 1 blocks, and increases/decreases relative to the previous difficulty. I think this makes 'fair' mining using an asic easier, but blocks will flow in much faster than 10 minutes.
Many things to ponder. We'll know much more by year end, I think. Do you know why there is a 2 hour rule, and whether it could have been make just 1 hour? That would have helped.
I think I read somewhere (no clue where), that 2 hour was chosen arbitrary. But it must be at least an hour (+ a little bit) to allow for node operator clock issues related to daylight saving time (+/- 1 hour).
got it, makes sense ... 1h10m would be better then
I'm watching it closely, very curious how this evolves :D We also need more wallets and lightning implementations supporting testnet4 (looking at you LND!)
Already on it :
https://github.com/lightningnetwork/lnd/issues/8966#issuecomment-2275072404
Also watching closely. Please feel free to add anything you want here: