hard_forker
Member
No.
Development Update - Oct 1, 2014
- Thread starter eduffield
- Start date
No.
What I would like to see WRT mining, in steps:
1. p2pool becomes the only pool from which blocks are accepted. If MN's can enforce payments then they should be able to enforce p2pool use. This solves the centralisation problem and removes the threat of 51% attacks. No one has yet offered me a sensible argument against this. Please do if you have one.
2. The logical next step - p2pool nodes need to be run on Masternodes. This removes any need for 'enforcement' in the first place. And miners who choose to leave their mined DRK in the pool account could receive pro-rata MN payments automatically at the protocol level. eg, MN starts off with 1000DRK, it has the usual probability of getting paid, but as miners increase the MN/pool balance, that probability gets weighted upwards accordingly relative to other MNPools. This makes MN shareholding practically automatic, creating scarcity and increasing price.
edit: alternately, it could allow the MNPool op to free up some or all of their 1000DRK to deploy a new Masternode, as long as the MNPool balance was > 1000DRK.
1. p2pool becomes the only pool from which blocks are accepted. If MN's can enforce payments then they should be able to enforce p2pool use. This solves the centralisation problem and removes the threat of 51% attacks. No one has yet offered me a sensible argument against this. Please do if you have one.
2. The logical next step - p2pool nodes need to be run on Masternodes. This removes any need for 'enforcement' in the first place. And miners who choose to leave their mined DRK in the pool account could receive pro-rata MN payments automatically at the protocol level. eg, MN starts off with 1000DRK, it has the usual probability of getting paid, but as miners increase the MN/pool balance, that probability gets weighted upwards accordingly relative to other MNPools. This makes MN shareholding practically automatic, creating scarcity and increasing price.
edit: alternately, it could allow the MNPool op to free up some or all of their 1000DRK to deploy a new Masternode, as long as the MNPool balance was > 1000DRK.
Last edited by a moderator:
hard_forker
Member
I see absolutely no problem with MN doing the processing of the transactions. Not only there won't be threat of 51% attack, there is more than enough processing power to deal with the transactions and it has a positive effect on the environment. In addition, MN owners who keep coin will be receiving new coins rather than multipool miners who sell the coin for BTC and drive the DRK price down. All we need is a stable high number of MN.
We only needed the miners for the begining phase. It was a win-win situation but their age comes to the end now and DRK will do better without them.
This is revolutionary speak right here!
This would essentially create a masternode server arms race. I don't like the idea one bit. I understand the ideology but not the practicality.
Why not have both? Anybody can CPU or GPU(with the correct server) mine on their masternodes right now. Why not let them do it, enable an option to idle CPU mine in the QT wallet and also let the mining pools do their thing? Surely diversity is better?
This would essentially create a masternode server arms race. I don't like the idea one bit. I understand the ideology but not the practicality.
Why not have both? Anybody can CPU or GPU(with the correct server) mine on their masternodes right now. Why not let them do it, enable an option to idle CPU mine in the QT wallet and also let the mining pools do their thing? Surely diversity is better?
hard_forker
Member
I believe that diversity with 900+ MN is sufficient. Remember - all it takes is one feable CPU to process all the transactions. Everything else is just a wasted power to assure decentralisation. But - why wasting electricity when we get the decentralisation through MN ?
If you're not using the CPU at 100% util you aren't wasting power.
Maybe thats a feature that could be considered for a future daemon though. I've been doing some benchmarking on a C3.XLarge and CPU mining is affecting the performance of the darkcoind instances on there quite badly. AWS is reporting oer 300% CPU UTIL. I'm sure that isn't a good thing.
Are 900 (mostly t2.micro) CPU mining going to do the job? I don't think so. This is why all hash should be welcomed imho.
Maybe thats a feature that could be considered for a future daemon though. I've been doing some benchmarking on a C3.XLarge and CPU mining is affecting the performance of the darkcoind instances on there quite badly. AWS is reporting oer 300% CPU UTIL. I'm sure that isn't a good thing.
Are 900 (mostly t2.micro) CPU mining going to do the job? I don't think so. This is why all hash should be welcomed imho.
I imagine if Masternodes were doing the block generation the process would not resemble "CPU mining", it would be a different beast entirely and would not need to eat up the entire CPU load. The reason the CPU load is so high when CPU mining is because you are competing for blocks and trying to squeeze every little bit of processing power out of it.
Last edited by a moderator:
That, Sir (I assume Sir from your avatar), makes sense.
hard_forker
Member
Think outside the mining box. You do not need megahashes to assure processing simple transactions. MN would not try to brute force X11 - there would be voting in process and the selected MN would easily process few dozens/hundred transactions using CPU power.
Amount of hashpower is irrelevant, as hard_forker says. It's only the distribution that matters, and if p2pool is compulsory then that problem is taken care of too. If 1000 masternodes are 'decentralised enough' to process transactions then they are decentralised enough to 'secure the network.' 2000-3000 Masterniodes, even better.
Having MN's 'cost' 1000DRK achieves exactly the same thing as the GPU/ASIC arms race - it makes it extremely hard/expensive for any one party/bloc to dictate the blockchain. And anyone trying will have to spend that huge amount on DRK, not AMD/Nvidia and KNC etc.
Splitting the block reward 50/50 and implementing compulsory p2pool mining via MN's with automatic MN shares for miner balances left on the MNPool would do wonders for DRK. Maybe terrible wonders, but wonders nontheless... :tongue:
edit: The more I think about it, the current 80/20 split in favour of miners is way out of proportion to the value each group brings to the coin. The miners will all whine if the split evolves more in favour of Masternodes but let them whine. They mostly just dump for BTC/fiat anyway, what/how much actual USE are they is the question we should be asking.
Last edited by a moderator:
So MangledBlue's idea to put a tick box in the next version of the QT to enable "setgenererate true n" is something worth pursuing? That would distribute hugely. And every now and then, you might get a handful of coins out of it 
Can anyone explain from a security stand point what would be the downside to having only masternodes generate blocks?
hard_forker
Member
Nope! ;-)
M
MangledBlue
Guest
My 2duffs - I don't really see one besides DDOS
but soon the IP's would be hidden.
How would I get my DRK out? (noob question I'm sure)
I use VMPlayer with my ubuntu and terminal to get into AWS.....
I'd like to hear some dev team responses to some of the ideas we have come up with.
I still want to see the golden three happen while Evan works on the masternode changes and IP Obsufucawhatever thing though
1) Enforce
2) Market
3) Distribute/improve mining conditions.
I still want to see the golden three happen while Evan works on the masternode changes and IP Obsufucawhatever thing though
1) Enforce
2) Market
3) Distribute/improve mining conditions.
To make myself clear, I'm not advocating the abolition of PoW mining, at least not just yet, I'm just proposing what I think would be a better way of integrating it. I think people having the opportunity to acquire DRK by mining is important (although, if I want gold I buy gold, I don't go out and panhandle for it...) - I just also think that having them mine on MNP2pools would benefit everyone with no downsides I have yet thought of.
MNP2pool mining also lends itself pretty easily to the automatic MN shares for anyone idea...
I also like stonehedge's idea of dead-simple one-click in-wallet mining, with the ability to specify how fried exactly you'd like your i5.
Combine that with profit-sharing MNP2pools and suddenly everyone has a tangible stake in the coin, everyone's a shareholder, not just the MN ops and the hardware junkies (who are often the same people anyway.)We need to make the acquisition of DRK the number 1 priority of anyone with some spare guvpaper. :grin:
Last edited by a moderator:
M
MangledBlue
Guest
Let's how much I get beat-up over this one....
Proposed Qt changes for easy solo-mining
Talking about supporting network health - Finding a block would be a bonus
Proposed Qt changes for easy solo-mining
Talking about supporting network health - Finding a block would be a bonus
Your idea is brilliant. Though I'm not well versed in P2pool it does sound like your idea does wonders for distribution. I'm not certain I'm advocating abolishing POW either, yet. Just a cool train of thought at the moment. It just doesn't seem like we need the miners, could do wonders for security, greatly increase ROI for MNs etc. Also, if Evan implemented the ability to generate a key for a 100 drk vin, where 10 people could combine their 100 drk trustlessly for a MN, it would help address the distribution issue of abolishing POW.
hard_forker
Member
Let's how much I get beat-up over this one....
Proposed Qt changes for easy solo-mining
Talking about supporting network health - Finding a block would be a bonus
Not sure about charity. Would leave processing of transactions to MN rather than bothering users & asking them for help. Keep the end user wallet simple.
M
MangledBlue
Guest
I'd rather not see MN's processing (mining) blocks.
Would it not cause the MN cost on the server to increase?
I'd like to see the MN's cost as little as possible (cost effective)
Would it not cause the MN cost on the server to increase?
I'd like to see the MN's cost as little as possible (cost effective)