Buterin’s “hard questions for any blockchain people” deal with some of the biggest obstacles that stand between the current state of blockchain technology and widespread adoption, highlighting hashpower centralization, the lack of “useful” large-scale apps, the high frequency of hacks, scalability and latency, issues with consensus methods, and the inefficiency of on-chain governance.
In this multi-part series, we will attempt to answer each of Buterin’s 7 questions.
“What are the centralization risks in proof of stake?”
Buterin highlighted the centralizations issues present within the Proof of Stake (PoS) consensus model in his first hard question for the blockchain world, noting that Bitmain and affiliated pools now control a significant portion of all Bitcoin — the largest PoW network — hashpower.
Blockchain centralization is a divisive issue within the cryptocurrency community. The core axiom of blockchain technology, decentralization, leads blockchain proponents to criticize any consensus method that could potentially be susceptible to centralization, whether engineered or organic. PoS is a consensus method that is potentially less susceptible to centralization than PoW, but presents different risks and challenges.
PoS, unlike PoW, doesn’t require network participants to use inordinate amounts of energy to solve complex computations in order to keep a blockchain immutable. Instead, PoS network participants are able to mine a percentage of transactions in accordance with their current “stake” in the network — or the proportion of network tokens that they possess.
In a PoS blockchain, blocks are assigned in a pseudo-random manner. Miners, referred to as “forgers,” must place their tokens at “stake,” which can be considered to be equivalent to escrow. Should a miner validate a fraudulent transaction, they forfeit their holdings as well as any future opportunity to forge new blocks.
PoS financially disincentivizes malicious network participation by putting cryptocurrency holdings directly at stake, rather than a disincentivizing malicious activity by forcing participants to spend money on energy. The manner in which PoS distributes rewards to forgers, however, creates the opportunity for monopolization in a very different way to hashpower centralization in PoW networks.
In a PoS network, forgers receive rewards in proportion to the amount of value they are able to stake. The odds that a participant may be selected as the next validator increase in proportion to the number of tokens held, which creates an ecosystem in which participants that hold the most tokens are most likely to benefit.
Staking algorithms also favor participants that are more active within the network, potentially leading to the unequal distribution of influence within a PoS network. By rewarding individuals that possess large amounts of tokens with more tokens, it’s possible that oligarchs may emerge in PoS networks in an organic manner.
In order to participate in a PoS network, forgers need only be online and possess tokens to stake. PoS still requires capital investment into operational costs such as hardware, electricity, and internet connectivity, but as PoS does not require hashing power competition operational costs don’t increase in proportion with the odds of being chosen to forge a block.
This creates a situation in which network participants that hold larger stakes benefit from lower operational costs — two separate participants may run identical operations, with identical costs, but possess stakes of different sizes. The participant that holds a larger stake is able to stake twice as much as the other, and thus has twice the chance of forging a block for the same operational cost.
The inevitable impact of economies of scale on smaller operators within PoS networks could lead to these operators dropping out of the network, resulting in the organic centralization of the network. Alternatively, smaller operators may decide to work together, pooling sources to share network rewards and operational costs — leading to the same result.
Data collected by PoS demonstrates that PoS networks are statistically less centralized than PoW networks. By analyzing the geolocation of wallets for PoS tokens, PoS shows that roughly 30% of all PoW Bitcoin hashing power and 45% of all Ethereum PoW hashing power is centralized within China.
Network participant geographic distribution for active PoS networks such as Stratis and PivX, however, reveals a far more distributed model, with network participation spread across multiple countries.
Ultimately, PoS is not immune to the threat of centralization, but it’s arguable that it’s more resistant to centralization given the lower barrier to network participation and data available on PoW versus PoS network distribution.
Next, we’ll tackle Buterin’s final hard question for the blockchain world: “Given how EOS governance has turned into an epic fail, doesn’t this mean that all on-chain governance including DAOs is fundamentally flawed? How can any DAO deal with bribe attacks, plutocrats and other risks?”
You can find all the answers for our “Answering Vitalik Buterin’s 7 Hard Questions For the Blockchain World” series here: