A Comparative Study Of Proof of Stake Algorithms

Abstract: Cryptocurrencies are growing at a rapid pace. It is well established that the mining operation fueling the cryptocurrency Bitcoin consumes a significant amount of energy. This study aims to determine if there is an alternative to the consensus model, maintaining security, integrity, and decentralization. It includes a deep dive into the Bitcoin network and the consensus model Proof of Work (PoW ). Additionally, four cryptocurrencies running the alternative consensus model Proof of Stake (PoS) are examined. To test if a PoS consensus model implementation could be a viable option, four distinct PoS implementations are described. Bitcoin is compared to four PoS implementations in the different aspects: security, scalability, and decentralization. How the different implementations differ and how the various properties are affected when changing the consensus algorithm is the main target of this thesis. The results show that the Bitcoin network suffers from a high fraction of the mining consumption originating from a small concentration of mining entities. Upon dissecting the PoS implementations, the largest attack vectors stem from a problem called the long range attack vector involving the lack of cost for proposing a block to the blockchain, in contrast to the Bitcoin blockchain that uses the PoW model that consumes external resources. This problem is handled by different security mechanisms depending on the implementation, mostly consisting of checkpointing schemes and finality on the network. However, despite not seeing any problems with this in practice, there is no fully trustless way to mitigate the problem. Additionally, despite PoS not being built primarily as a scaling solution, a cryptocurrency can implement it in that setting. For example, with EOS, significantly increasing the throughput compared to Bitcoin, with the cost of losing parts of the decentralization properties by maintaining 21 block producers that produce blocks for the entire blockchain. Additionally, Algorand is significantly improving throughput compared to Bitcoin, under the assumption that at least 2 3 of the users are active, meaning that they are either validating blocks or delegating their vote to an active validator.