Why Proof of Stake? A dissection of the transition from RandPoE to PoS
It is well documented now that Witnet is transitioning to Proof of Stake from its current consensus mechanism. But what went into this decision and why was Proof of Stake sought after?
What is Witnet?
Witnet is a layer 1 decentralized oracle network, incentivized by the $WIT coin, and single-purpose built for data requests.
This means that Witnet brings real world, off-chain information onto a blockchain that helps smart contracts to execute their deterministic programs. Because Witnet was single-purpose built for this function, it does its job incredibly well via a series of security parameters that we set in motion when the network launched.
Witnet is fully decentralized with thousands upon thousands of nodes running on it at any given time, all incentivized to provide truthful information and enforce the rules of the network.
The Witnet development community recently announced the development of Witnet 2.0 that will transition the Witnet network into a full Proof of Stake oracle, making Witnet the first layer 1 PoS oracle ever.
What is Random Proof of Eligibility?
Random Proof of Eligibility is Witnet’s current consensus mechanism. It is a very unique and exciting mechanism that was built for the layer 1 in the early days, but not designed to scale at the rate of growth we’re seeing for Witnet.
What’s Proof of Work?
Proof of Work is a consensus mechanism that requires someone to spend real world resources to add a block to the blockchain. In most cases, they are rewarded with units of the currency of that network. As with Bitcoin, blocks are non-deterministic, and are different for each identity working on the block, such that every block being worked on at the same time contains different transactions. Blocks are found rather than exposed, meaning block timing is not pre-determined, and the only way to add it to the chain is to find it first, which could take 1 minute, 5, 10, or an hour in some cases.
What’s Proof of Stake?
Proof of Stake is an alternative consensus mechanism that requires someone to stake in-network units of currency to secure the network. This means that entities must lock a specific amount of native coins in order to validate the next block. Blocks aren’t “found” but rather exposed to validators based on their percentage of stake in the network and on VRF sequences (the architecture is unique to many different PoS networks), and similar to PoW, blocks are different based on the transactions included by the independent validators. These blocks however, are deterministic and are produced at the same time, every time (for Ethereum, it’s every 12 seconds).
In Witnet
Under a Random Proof of Eligibility model, Witnet nodes are drawing a random number within the protocol to mine the next block in Witnet. Blocks are deterministic and happen every 45 seconds.
Blocks are mined based on what is referred to as the “Active Reputation Set” or ARS. The idea based around this is that nodes who have acted honest in the past are most likely to act honest in the future, so they should be referred to for sources of truth more often than nodes who are new, or those who are less reputable. Nodes calculate their right to mine a block or resolve a data request within the network via a cryptographic sortition algorithm.
This is a great system for an oracle configuration like Witnet. Because it is a layer 1, it made a lot more sense to have an approach like this that negated frontrunning and kept nodes anonymous, in turn keeping them honest. We are now at a point where Witnet has grown at an astronomical rate- making scaling security an issue.
Why choose Proof of Stake?
Proof of Work would secure an oracle like Witnet, no? It undoubtably could. But the problem with Proof of Work, especially a hashing system like SHA-256 is that it isn’t scalable- it’s designed as such. Proof of Work is incredibly efficient in a system that creates incentive around the main problem: double spending. The basis is that double spending can be done with digital money so you either need a centralized authority to prevent it, or you need to create high costs associated with doing it.
For an oracle, there exists a number of different problems that aren’t specifically a double spend (which of course is a problem in all blockchain networks, but that’s not the number one concern for an oracle). These problems are referred to as the blockchain oracle problem. In essence, it is the problem that real world information can’t be brought on chain without breaking the consensus mechanism or trusting a third party to do it, read more about the oracle problem here.
Witnet tackled this problem by creating a layer 1 with its own unique properties for decentralization and ensuring truthfulness without trust. The most important mechanism that allowed Witnet to go after the issue is the explicit staking mechanism. This wasn’t inherently a Proof of Stake mechanism, because percent share of stake didn’t determine eligibility in the network. Staking in Witnet was simply posting some collateral that would be burned if you were dishonest. This ensured nodes remained honest via cryptoeconomic incentives. Not only did you lose some $WIT for being dishonest, you also needed to control a significant amount of nodes with staked $WIT in order to successfully attack the network.
The fact that the network has been resilient and hasn’t been attacked yet is evidence that the network works as intended.
But there has come a point where the current consensus mechanism doesn’t scale to secure all the protocols Witnet is tasked with securing. In the current moment the Witnet network and its secured protocols remain strong, but if the total value secured were to multiply quickly (as is the case of a bull market), an attack could be profitable.
A more fundamental Proof of Stake mechanism would ensure that collateral necessary for attack would scale with the amount that is staked in honest validators hands. There are a few different ways this mechanism can be implemented, that will be released in the coming days. The important thing that remains constant is that incentives work in Proof of Stake: consensus rewards outweigh the cost of attack purely because that cost of attack grows with nodes who want to remain honest and earn a healthy reward for providing a valuable service. If you want to attack Witnet or its protocol users, you must outbid them on the open market.
That is why Proof of Stake.
