Things To Consider: Is There a Meaningful Difference Between Proof of Work and Proof of Stake?
While there has been recent discussion about the current misalignment of incentives on the Steem blockchain, I have been thinking about blockchain at a more fundamental level and trying to understand the mechanics beyond the meaningless word "blockchain" has inevitably become. I'm trying to tackle the big questions.
The first topic that I want to address is the continual debate between Proof-of-Stake (PoS) and Proof-of-Work (PoW). Rather than simply give my opinion, I want to try and present both parts of the debate, add a few additional thoughts to be considered and then open up the discussion to all who have something meaningful to say about it.
The overarching goal of this series of posts is to encourage discussion and promote understanding of these higher level topics and hopefully give all of us (including myself) an opportunity to explore new perspectives and gain new insights about this new technology that none of us really truly understand to this point.
.jpg)
Image Source
First, I feel that is appropriate that we contextualize the big question of consideration. I will make two simplifications to frame this discussion. The first being that PoW is an effective way of securing blockchains with the major drawback being the wasted electricity and a major appeal of PoS is to provide an alternative and equally effective way of securing blockchains without that drawback.
The major point in the debate centers around the effectiveness of PoS in securing the blockchain. We will now take a look at two different sides of the argument and try to unwrap the key points in this discussion.
First off, we should start with why PoS essentially behaves the same as PoW with less of the negative side effects. The argument here being that PoS essentially does the same thing that PoW does but more efficiently in terms of energy and speed.
PoS works by using stake-weight as a consensus mechanism rather than computational hash power. In both cases, it is very expensive to attack the network as one either needs to purchase a lot of computing power or a lot of coins in circulation. However, in the case of PoS, there is no energy wasted in the selection of block producers. The argument here is that not only are PoS coins more eco-friendly but reduced cost of block production makes the producing more accessible to more people and allows for faster blockchains.
Also, theoretically speaking, the cost of an 51% attack is higher in PoS coins than PoW coins since one would need to obtain 51% of the total staked coins versus needing only 51% of hash power over a limited period of time which would cost significantly less than 51% of value of total PoW coins in circulation. To simplify, a 51% attack is harder on an established, decentralized PoS coin than an established, decentralized PoW coin.
So there has to be some catch, right? Not only does Proof of Stake essentially replicate Proof of Work at a much lower cost but if developed correctly, PoS coins are more secure against attacks on the network. But now we will take a look at some of the drawbacks of the technology to see what those who argue against Proof of Stake see.
The major argument against PoS boils down to the Nothing at Stake problem. The issue here is determining the validity of a chain if a fork occurs. In PoW, multiple chains are resolved through voting via hash power. The chain which has the most hash power wins. In turn, the chain that is the longest is the valid chain as the most hash power is behind it.
With PoS such a notion doesn't really exist because work is never performed. Anyone with stake can stake both chains. Thus if some party arbitrarily produces their own block and forks the network, there is no real way of determining the legitimate chain as there is no cost to staking on a particular network given the fork. Sure the chain with the most stake behind it will eventually win out but what are the potential weaknesses of such a fork?
Consider the following attack: Given a fork with equal amounts of stake behind both which is now possible as block producers can now mine both chains, you now have a situation where you can double spend. You spend on one chain and stake on the other and given you have some amount of stake (less than 51%), you can get away with this type of double spend.
This means that during such a fork, PoS would be less secure than PoW as there is now a way to exploit the system that simply isn't a problem in PoW blockchains. So, the argument here is that because of the nothing at stake problem and a lack of a fix to PoS that eliminates it, PoW is still the superior method that requires the least trust.
So, there are the arguments for and against Proof of Stake and its relationship with Proof of Work blockchains. I know that it got pretty technical above so feel free to ask questions if you need additional clarification or add your own clarifying points to give everyone a better idea of some the ideas presented above.
Before you go off to leave a comment or read something else, I leave you with one last thing to consider. Despite the nothing at stake problem, is there acceptable amount of trust in other parties that makes the efficiencies gained in PoS over PoW acceptable? Is it okay for a blockchain to trust that people won't exploit the chain if they are given large enough incentives or disincentives? These are things to consider. The floor is now yours.
Both proof-of -work and proof-of-stake are fledgling primitive versions of how a blockchain should operate. I prefer proof-of-stake, but only because there are so few options at the moment.
Proof-of work trusts you if you have computing power. Proof-of-stake trusts you if you have stake (money). I expect much more complex reputation systems to arise that put both of these methods to shame.
Proof-of-reputation/trust will continue to evolve indefinitely. As soon as someone finds a way to exploit the current system someone will come along with a new system that is even harder to defeat.
I've been thinking about this for a while and I've yet to see something as elegant as the original Proof-of-Work consensus mechanism. Sure, it is really wasteful but the incentives align correctly. I think that any future consensus mechanism will require some wastefulness although I can't really explain at this point why I feel that way. I guess it is just a feeling.
But I'm interested to see new composite consensus systems. I'm sure that as more people enter the space there will be more creative solutions to the problem. I'm curious at what new types of cost we can come with as a security counterbalance (as opposed to electricity or stake).
Proof-of-work was the perfect solution to bootstrap the movement. Now, coins like Bitcoin are too valuable and the competition for the hash lottery is too fierce. Proof-of-stake came along just at the right time. And I expect proof-of-whatever's-next to feel the same way in retrospect.
I think DPOS resolves the nothing at stake problem quite satisfactorily. Witnesses certainly have something at stake and will destroy their reputation and forfeit their positions if they sign conflicting blocks that fork the chain.
It solves the nothing at stake problem by comprising through it necessitating trust in some third party entity. The third party in this case is the community and it's elected witnesses. Witnesses certainly have something to lose if they misbehave, but if the community doesn't really care a lot about the witnesses then witnesses may be able to get away with some things. We trust that enough people care.
The current blockchain is vulnerable to voting collusion among whales (less than 51%) and general apathy of the majority of the users (some significant %). There are certainly small fixes that we can use to make it more secure, but that flaw with always be there. But hey, it might be worth it for such a fast and expansive blockchain.
POW requires trust in the community of block signers also, which in most cases means a few large mining pools and farms. Delegation and apathy are not unique to proof of stake, and it's best to design the protocol to account for realistic expectations about the role such factors will play.
Obviously in a POS network the distribution of stake is critical, and the distribution of Steem is quite centralized. That fact encourages even greater voter apathy in this network, but does not necessarily indicate a flaw in the consensus protocol itself.
POS has serious problems in that it is REALLY easy to take over.
You were describing an attack from the outside, but POS attacks come from the inside. They already have a large stake. They only have to convince a couple of other people, and they can change things. Subtly of course, because they do not want to affect the whole ecosystem.
In fact, a POS attack has been successfully done on the STEEM chain.
There is too much trust in the witnesses in a POS system, and those witnesses rely on the central authority. And this creates an entity that is really easy to manipulate.
It does not solve the Byzantine General attack problem.
I wouldn't say there is too much trust in witnesses as they are elected by stake-holders. But I can see the argument that there is too much trust in the community to properly select and manage their witnesses.
I'm curious, however, when and what type attack on the Steem blockchain are you talking about? Is it a major attack? (overwriting blocks, altering the history) or something less serious? (spamming, collusion) One could argue that those less serious attacks aren't really attacks but roundabout ways of mining the blockchain. I've seen arguments that such behavior could be seen as a "roundabout" Proof-of-Work.
It was altering history. And you will have to try to find the posts on it. Some people posted about it.
IN POS systems, ... well, in all human systems, cliques form. And, it is not unlikely that a clique will find itself with enough power to change things.
It is because those who get in early are usually friends, or friends of friends. They are not independent actors.
I like Reddcoin working at paying people to all have their wallets open, and the wallets all verifying. I wish to look into it more.
And i agree, that next year, there will be something that blows POS and POW out of the water.