Topic: Proof of Virtual Miner - Dual Coin POS (Read 493 times)

Yep.

Thinking out loud.. 

I see a similarity with CASPER's Security Deposit based system.

The virtual rigs are the safety deposits, and in CASPER terminology every 'v-rig' owner is a validator. We can now penalise not only non-participating miners, but also misbehaving miners who mine on multiple chains and try other 'shenanigans'.

Also - small or large coin holders don't have to mine, and the network can still be secure. In a 1 coin POS system, large stake holders HAVE to mine, or the chain will definitely become insecure.



This is the bit I too am unsure of. Whether it makes a 51% attack easier.

You would now not need 51% of the 'Currency', but just 51% of the mining tokens, BUT this could be worth less or MORE than 51% of the currency. 

And finally - Someone may own a percentage of the v-rigs, and simply not wish to sell them at 'any' price (perhaps for security reasons). So that no matter how long the system ran, no miner would 'have to' eventually end up scooping up all the coins, which does eventually happen in all the current implementations of POS I have seen. (ie  you leave a miner with initial 20% mining power running. At some point in the future, if he doesn't cash out, he will own everything )

In some implementations of Proof of Stake you don't have to leave the client running "all the time", it kind of tries to catch up when it does get run.

So with those maybe as long as small holders of the currency do fire up their client occassionally, even if just to send some pocket-money to a kid or whatever, maybe the fact it will have to catch up to the blackchain, and maybe doesn't ring an alarm when it does catch up, tends tleave it sitting there staking long enough enough of the time that such folk do actually stake without necessarily even realising it or planning to?

I envision someone firing it up to send some coin or check if they received some someone was going to send them, and realising it will take a few hours for the blockchain to catch up, thus getting distracted by something else; so that when the blockchain does catch up staking can start and it might be gosh knows how long before the user realises the thing did catch up thus is ready to do what they fired it up to do.

You might suggest that they could just send coins without waiting for the blackchain to catch up, of course; but I think a larger weakness is the kind of small holder casual user in this scenario is actually probably unlikely to even use a client that uses the blackchain directly, thus their coins would most likely be actually held by some third party, who could do the staking for them (a staking pool) or just take the stake earnings for themselves.

-MarkM-

Depends on the value of the original currency and it's trajectory since I purchased the mining tokens I guess.


I kind of see where you're going with this now. I don't see how it solves any real problems but it is an interesting thought.
Your idea is basically to vritualize (as you aptly named it) POW-mining rigs by pegging mining-power to tokens that can then be purchased and tracked on the same chain as the coin that is being mined.

I don't think it would be much different from other POS. You could speculate on mining-power but I'm not sure just how great that is.
I'm wondering if it would make 51% attacks easier if people start dumping their mining-power because the coin goes to shiz. It just seems that it would make POS even more vulnerable to an attacker gaining a significant amount of mining power.
It would also seem to be even easier to join/create pools which isn't exactly desirable. Just send me your mining-power and I'll mine for you and we share the earnings.

Hmm.. 'Kind-of'.. It's not so clear exactly how much that is.

If the mining power you have generates 100 coins a year, how much is that worth ? 101 coins, 106, 200.. 50 ? Would a free market help to determine that..?


No - I am saying that small stake holders don't mine. If you want to mine you need to buy mining tokens. Then you can penalise those with mining tokens who don't mine, without affecting the regular small hold user. So that almost 100% of possible miners would then mine. I think..

Your assumption that


just doesn't make sense to me. To anyone that may want to "virtually mine" the mining power will be exactly as valuable as the fees they are producing. Now that value might change over time but it will always be directly linked to the value of the original currency and how much of that currency i can mine with it.

I also don't see why smaller miners would be more likely to mine. They don't mine with their original currency so why would they be mining with mining power that they had to buy for their original currency ?
 

Yes - I'm still not sure if the subtleties have eluded me..

Hypothetically -

Let's say there are a constant 100 TXN/s on this chain.

And that each TXN is for exactly 1 'currency' coin.

And that the fees are 1%.

.. So this generates 1 'currency' coin per second for the miners.

How many mining tokens, from the 100% available, 1 currency coin is worth is now floating. Not Fixed - as in single coin POS.

Maybe the miners want a lot for their mining tokens, even if the total collected fees are small, because they think the network has potential, or maybe they dump them for whatever they can get.

I know it sounds like the same thing as regular POS, but something about this 2 coin system seems more flexible..

Also - I think you would get a much higher percentage of possible miners mining. Unlike the current systems where small holders, probably the majority of users, who should still mine, don't.

Since if you bought this 'Virtual Mining Equipment', you'd be sure to use it, no? (You could demurrage unused mining assets to make sure those who got into virtual mining, err.. mined. Without penalising regular users.)

I don't see how introducing a second "currency" would solve the problem you are describing.
 
Say there is a second currency. Why do you think it would be more valuable than the original currency ?
There is no incentive to "mine" other than the tx-fees. So the miningpower-currency would only be as valuable as the tx-fees one can mine with it which is exactly the same as the original currency.

I'm probably missing your point...?

Seeing Vitalik + Vlad push CASPER as the new consensus model for ethereum,

https://blog.ethereum.org/2015/08/01/introducing-casper-friendly-ghost/

has made me revisit some of my own ideas about POS consensus.

Yes, you have to get a checkpoint from 'someone' when you first join the network (Weak Subjectivity as Vitalik puts it), but other than that the consensus rules seem to work just fine.

And Neucoin's white paper

http://www.neucoin.org/en/whitepaper/download

goes a long way to showing that the grinding attacks that are always possible in POS chains, are actually not as easy to pull off as was originally thought. The combinations and numbers involved are as astronomical as cracking SHA256 itself.

Now - here is the bit that I wanted to have a brainstorm about..

I don't like the way the Stake, the currency of the system, is the same as your 'Mining' power. This means that the fees paid for transactions, are directly linked to the re-distribution of mining power. This seems like there are conflicting incentives ? The mining power could be worth much more / or less than the fees are producing.

What if you had a dual coin POS chain. There are 2 currencies on this chain. One is the currency, and the other is mining power (A fixed amount - 100%). You can trade on chain form one to the other.

When you send a txn, you pay a fee in the currency, not the mining power.

An empty block (I would consider this a rare event that only happens at the bootstrap period of the coin.), pays nothing. There are still ways to make this worthwhile for the miner, with PPLNS (Pay-Per-Last-N-Shares) type chains.

Now there is a Free Market for the value of the mining tokens, and simply leaving a miner running does not increase your ability to mine, which is the case with ALL current POS models.

I can't work out if this is actually different, or the same as current models, just a bit more complex. 

Thoughts ?