Is collateralised mining just proof of stake?

[bielie]

There is talk of "collateralised mining" which means only masternode owners will be able to mine.

1) Why would this be good for the network, apart from being hugely profitable for the MNO's? (who will therefore probably vote for it)

2) Since there will be no competition for more hashpower, mining difficulty will remain stable. How does this differ in practise from proof of stake?

"Without the incentives and constraints created by the prospect of profit and the threat of losses, the same output might well cost millions of dollars more”

3) I would like to hear then why Craig Wright is wrong about POS in this article:


https://medium.com/@sumanthneppalli...not-cryptographic-it-is-economic-7a8042be6b07

Thanks! I would really like answers, since I am heavily invested in Dash!

 

[flare]

There is talk of "collateralised mining" which means only masternode owners will be able to mine.

Collateralised mining was just a conceptual idea - there are no plans to implement it in Dash anytime soon. And as the premise is wrong, i won't discuss questions 1) - 3)

 

[bielie]

Thanks!

 

[Macrochip]

The term "Proof of Stake" has to be defined first before we can equate our concept to it. If you're talking about the "Proof of Stake" mining scheme which is used to mint new coins, then no: Collateralized mining has nothing to do with it. Simply because you'd still be mining by hashing, not staking.

If you're talking about the concept of "proving a stake" ("I got X coins") as an "entry ticket" to mine, then yes: Coll. mining sure comes close to it.

Requiring proof of ownership has a myriad of upsides when you think about it. Not only "skin in the game" and less volatility/market stabilization. The network is also given a tool to control its own total hashpower, which is a ground breaking concept. Energy friendly blockchain anyone? While being (virtually) fully secure at the same time? Others dream of that, we can actually do it.

The mechanism is simple: you have to own X Dash to be able to mine at Y MH/s and these values can be adjusted dynamically by the network itself. Consume too much power? Increase the requirement. Done.
Ofc you'd still be required to have real world hardware able to mine at whatever speed your Dash stash allows you to, so the economic constraint becomes two-fold, but not overbearing. It's quite the improvement, because currently all an attacker needs to launch a 51% attack is having more hashpower than the entire network combined. But with collateralized mining, he'll not only need that but additionally more Dash than the entire mining network combined, which is an economic impossibility. Sounds awfully familiar to some other wildly successful incentive scheme, right? Expanding past successes onto other network layers is only a logical step imo.

"Without the incentives and constraints created by the prospect of profit and the threat of losses, the same output might well cost millions of dollars more”

That sentence still holds true with collateralized mining. People will always be incentivized to get more Dash in order to mine more Dash faster.

Btw: Thanks for linking that article. I only skimmed it, so I don't know what you mean by Craig Wright being wrong. My takeaway was PoW is future proof and requiring healthy competition & (re)investments is how the economy should work. That includes buying more Dash to be able to mine more as well, so going by that the Collateralized Mining principle is economically very sound.

 

[bielie]

Since collateralized mining gives you access to x megahashes and no more it removes the race for hashing power completely. Whether you create coins by hashing or by staking makes no difference: The race has shift from buying hashpower to buying coins regardless of what method you use to actually write the blocks. So in reality there is no real difference between POS and CM.

Hashpower is potentially infinite, and is only capped by hardware cost vs coin value. As coin value goes up, haspower goes up because more money is available for hardware. Coins are limited. There will only ever be 16M of them, and already more than 60% are in masternodes. That means the current MNO's will be there ad infinitum and get the full non-treasury block reward ad infinitum. (90%) It removes competition.

On an economic level I really do not see the difference. If CSW is right about POS then collateralised mining would be a mistake for Dash.

This is my worry.

As I said, the MNO's would make a killing; it would be the equivalent of a state granted monopoly... So I can see the MNO's voting for it.

 

[bielie]

Read this article, especially what CSW says about Casper and the oligopoly.

https://coingeek.com/dr-craig-s-wright-proof-work-trumps-proof-stake/

I think the current Dash system is good. It provides good governance with POW.

 

[Macrochip]

So your argument boils down to CM=PoS and CSW said PoS is bad so CM must be bad, too. I agree that PoS is flawed and I've also explained that CM has nothing to do with PoS.

Since collateralized mining gives you access to x megahashes and no more it removes the race for hashing power completely.

That's incorrect. The race for hashpower is still there, only the requirements to get it have increased. You may have the coins, but that doesn't mean you have the hardware to mine at the level your coins enable you to.

Whether you create coins by hashing or by staking makes no difference

It makes a world of difference. Expending real world resources always trumps the flawed concept of staking. The article you linked yourself explains it.

The race has shift from buying hashpower to buying coins

Only momentarily. Once you're mining at hardware maximum, you'll still need to buy more/better hardware.

So in reality there is no real difference between POS and CM.

False, as described by my points above.

Hashpower is potentially infinite

And so is the waste of energy. CM enables us to limit the expenditure of electricity while remaining secure at the same time simply by constraining it through coin ownership.

Just like the MN network it's just an added layer not a replacement.

There will only ever be 16M

18.9M.

That means the current MNO's will be there ad infinitum and get the full non-treasury block reward ad infinitum.

Two things: 1) You're implying every MNO will automatically become a miner assuming they all have the necessary hardware to mine and 2) even if that were the case (it won't), what would be so terrible about the fact that those who have the most to gain/lose in our network have the highest responsibility as well? We have 3 years of empirical evidence how solid and successful that model is. Masternodes are being copied left and right.

It removes competition.

Not true. An MNO may have 1000 Dash but a CPU that allows him to mine at an equivalent of only 50 Dash, while a miner with 250 Dash may have an ASIC that allows him to mine at an equivalent of 500 Dash. So the miner is going earn 5 Dash for every 1 Dash the MNO mines despite owning less Dash in total. The MNO's total amount of Dash and his MN rewards do not offset this in any way. The only way for the MNO to compete with the miner is to buy better hardware just like the design of Proof of Work requires it. The economics with CM are still sound.

On an economic level I really do not see the difference.

You should. I've explained it twice now.

If CSW is right about POS then collateralised mining would be a mistake for Dash.

Non sequitur. CM!=PoS

As I said, the MNO's would make a killing

Again, you're assuming with no data that they have a vast mining array at their disposal. Even then I'd rather have them in power than anyone else.

 

[bielie]

Thanks @Macrochip

I will look at your answers and see if they make sense to me.

So CM does not award a specific amount of hashes to e key holder but only allows him to mine?

 

[Macrochip]

Correct!

Nothing is changed about the hardware requirements. You could own 1 million Dash, it wouldn't help you mine any faster than your existing hardware allows you to. All you can do is "fill up" the hashrate these 1 million Dash grant you to mine at.

 

[bielie]

So there IS a max hashrate associated with a mining key?

 

[AnarchicCluster]

So there IS a max hashrate associated with a mining key?

As long as what macrochip is saying is true, yes. However, when I saw flare's comment I got an impression that collectivised mining is something that is in very distant future (by cryptoworld standard) if at all. CM seems to be not definite yet.

 

[Macrochip]

Ok, first of all: Nothing I say is official. I am not part of Core. I am only independently theorizing here how CM would work and what the benefits would be. Nothing else. What flare said is official. He knows more than me and if he says there's no definite plan to implement it at all, then that's how it is. Only voicing my opinion on how it could be. After all these kinds of community discussions are the reason we have Trustless Masternode Shares on the roadmap now. The idea did not come from Core in case anyone thinks that...

So there IS a max hashrate associated with a mining key?

I'm confused. Where did that come from? I think I missed when that was the issue. Anyway, to answer: Of course there is a theoretical maximum hashrate for the entire network. It's (total_amount_of_collateralized_Dash * Hashrate_granted_per_Dash).

Note that the research we're funding at ASU might find a sweetspot ratio one day for consumed electricity vs. security provided by PoW-mining, so we could cap the hashrate (and thus energy consumption) by adjusting variable 2 down to a sufficient value.

To summarize the concept (and I'm most likely repeating myself here) one simple example: To mine at 100 MH/s you must prove ownership of let's just say 1 Dash. Now you have a mining rig that allows you to mine at 500 MH/s. But since you only own 1 Dash, the network disregards every hash above 100 MH you provide (*see edit). To be able to mine with your rig at full capacity, you must acquire another 4 Dash -> race for coins.

Opposite example: You have 1000 Dash, which would allow you in theory to mine at 100 GH/s. But you possess a mining rig that is capable of only 20 GH/s. So your only option to mine at full or at least near full capacity is to buy more and/or better hardware -> race for hardware.

Is that clearing things up now?

(*edit: To clear up a potential misunderstanding: The reason we'd save power here is the idea that miners would be incentivized to lower their hashrate voluntarily if they don't own the required Dash amount to mine at 100% hardware usage, otherwise they'd be wasting electricity for zero gain)

 

[camosoul]

The success of the Proof of Service model lies in the deliberate lack of granularity. The huge chunk of 1000 DASH leaves much liquid through the difficulty of attaining 1000 DASH.

This must be maintained or it will essentially become Proof of Hoard, and there will be too much deflationary value. Everyone will hodl and no one will actually use it for commerce.

This is why integrating MN sharing and savings accounts in Evolution concern me. It works because it's difficult. Making it easy breaks it.

If collateralized mining is implemented, it should be in the same all-or-nothing fashion, not tiered. Scaling decreases utility as a currency, and increases Ponzi by doing nothing more than requiring an artificial prop via hoarding for mining permissions. Of course, that's no concern for crypto-at-large, because cryptotards only care about the Ponzi.

DASH would be foolish to break itself based on the thoughts and theory of persons with an opposing objective.

We see obligations continuing to heap upon MasterNodes. And this model works. The more feature set requirements MasterNodes service, the more they get paid. Why shouldn't they absorb mining into that service burden? Mining is just one of the many things a crypto must do to be useful.

But, with so many other services, the entropy that could be gathered from DAPI alone... I'm beginning to question the value of mining over other proof-sources now available, and on the near horizon.

We're not accustomed to that because the Satoshi experiment, and all it's Ponzi clones, failed to have any of these other dimensions. We've been conditioned by dysfunction being "good enough." DASH is, or was, aiming for function over dysfunction. I'm not sure what they're up to anymore. They seem to have lost their way...

Evan knew this from day one. It's why I came to DASH back when the client still had an X-Coin logo on it.

I question if he has lost his way, but DASH is still the closest to on-track with this fundamental principal than anything else.

Function over dysfunction philosophy is the only reason I'm still here. Please don't break it.

 

[demo]

It makes a world of difference. Expending real world resources always trumps the flawed concept of staking.

Why you consider the concept of staking flawed?

 

[dash_gt_urcoin]

To summarize the concept (and I'm most likely repeating myself here) one simple example: To mine at 100 MH/s you must prove ownership of let's just say 1 Dash. Now you have a mining rig that allows you to mine at 500 MH/s. But since you only own 1 Dash, the network disregards every hash above 100 MH you provide (*see edit). To be able to mine with your rig at full capacity, you must acquire another 4 Dash -> race for coins.

Opposite example: You have 1000 Dash, which would allow you in theory to mine at 100 GH/s. But you possess a mining rig that is capable of only 20 GH/s. So your only option to mine at full or at least near full capacity is to buy more and/or better hardware -> race for hardware.

How exactly do you "limit" someones hash rate to your example of 1 Dash = 100MH? I would love to see some technicals behind this. What happens when someone has 1 Dash, but a 1GH rig? How does the network distinguish between "too much" hashrate, and someone just getting incredibly lucky?

The only functioning hash rate limiter I can see is if we select a node (based on the pseudo-random PoS philisophy of more coins = more probability of being selected), and then allow them to mine as many blocks in that time period as their max_hashrate allows. But then we can just throw out hashing completely right? What purpose does it serve here?

Secondly, if there was some sort of functioning hash-rate limiter due to the amount of Dash you own, why don't we just drop it to 1 Dash = 1 Hash? Then the entire network is supported by only 18.9MHash, and is super "eco-friendly"... But then we are basically just reverting to PoS... but since you have to use some sort of selection process for this type of collateralized mining to work at all, then we should just revert to pure PoS. right?

 

[Macrochip]

How exactly do you "limit" someones hash rate to your example of 1 Dash = 100MH? I would love to see some technicals behind this. What happens when someone has 1 Dash, but a 1GH rig? How does the network distinguish between "too much" hashrate, and someone just getting incredibly lucky?

The only functioning hash rate limiter I can see is if we select a node (based on the pseudo-random PoS philisophy of more coins = more probability of being selected), and then allow them to mine as many blocks in that time period as their max_hashrate allows. But then we can just throw out hashing completely right? What purpose does it serve here?

Secondly, if there was some sort of functioning hash-rate limiter due to the amount of Dash you own, why don't we just drop it to 1 Dash = 1 Hash? Then the entire network is supported by only 18.9MHash, and is super "eco-friendly"... But then we are basically just reverting to PoS... but since you have to use some sort of selection process for this type of collateralized mining to work at all, then we should just revert to pure PoS. right?

Thank you for your input!

I'm neither a developer nor an engineer so I can't really give you any technical details, only describe the concept on a high level.

You wouldn't have to create an actual hashrate limiter. I don't think that's even technically possible in a decentralized network. You'd built an incentive to make miners limit their hashrate themselves instead. This could be done by linking hashrate to collateral and have the entirety of the network continuously check how many funds miner XYZ has and make it disregard the amount of hashes he submits above his allotted threshold.

He'd be strongly incentivized to limit the hashrate by himself then, because it would be wasted energy with zero ROI otherwise. That is until he accumulates more funds to make use of his additional hashrate of course.

Now I have no idea how to build consensus rules to ignore a participant's hashes, but I'm fairly convinced it is technically feasible. At least I see no insurmountable hurdle. Do you?

Of course you could limit the hashrate down to whatever you like and make our network extremely eco-friendly, but the point is that we must not and cannot get rid of proof of work. It is essential in creating the entropy and deriving the security required for Masternode quorums. Proof of Stake just can't provide that and most likely never will, because a closed system cannot validate itself without any external entity for verification. dPOS might be a solution, but it needs to prove itself in the field first. Meanwhile PoW is the way to go.

 

[dash_gt_urcoin]

Now I have no idea how to build consensus rules to ignore a participant's hashes, but I'm fairly convinced it is technically feasible. At least I see no insurmountable hurdle. Do you?

Yes, there is definitely an insurmountable hurdle here. Mining is, by definition, a random process. Therefore, you cannot distinguish between a user getting really really lucky and mining a lot of blocks, or a user with more hash rate than he is "allowed".

So the only thing you could do, is enforce protocol rules to say that a user with x% of total supply cannot mine more than y% of total blocks (calculated on some factors).

However, a user could still get to "choose" their block-hash if they have say 10x their "allowed" hash power. This would mean they have, on average, 10 hashes to choose from for a mined block. So now you lose all entropy with this scheme as well. So now you should just revert to plain POS mechanics.

Also, the lack of entropy isn't the issue with POS. It's the fact that there is not mathematical certainty about which chain the the "true" chain: download.wpsoftware (dot) net/bitcoin/pos.pdf

 

[Macrochip]

However, a user could still get to "choose" their block-hash if they have say 10x their "allowed" hash power. This would mean they have, on average, 10 hashes to choose from for a mined block. So now you lose all entropy with this scheme as well. So now you should just revert to plain POS mechanics.

How does choosing from a pool of hashes undo entropy? The hashes are purely random. If I have 5 million hashes and choose 1 million from them in what way am I harming entropy when I have no basis or priority on which I'm selecting them? They're just useless strings of text and numbers to me and I cannot know beforehand which one is going to contain the nonce the network is looking for. There is no advantage to the miner to be able to select from a certain pool of hashes. In fact there is a major disadvantage at creating such a pool in the first place, which is the wasted energy I was referring to earlier.

Also, the lack of entropy isn't the issue with POS

I was not arguing what the issue with PoS is, but why we need PoW. The problems with PoS are another topic. Also I don't understand how any of my concepts "revert back to PoS" as long as mining takes place.


btw: thanks @Friend for pointing out my typo! Corrected to "proof of work", which is what I meant to say of course.