ECTO A Novel Proof-of-Stake Cryptocurrency 6 . 824 Final Project Writeup

Content Provider	Semantic Scholar
Author	Aspegren, Henry Lovejoy, James
Copyright Year	2018
Abstract	Currently the most popular cryptocurrencies use Proof-of-Work. However this consensus mechanism consumes large amounts of energy: Bitcoin alone uses more energy than the entire country of Cuba [1]. Proof-of-Stake has been proposed as an energy-efficient alternative. Much like Proof-of-Work, in Proofof-Stake, a single participant known as the leader, staker or block proposer, is selected to add a block to the chain. Leaders are selected probabilistically in proportion to how much of the currency the leader has ”staked”. Ideally this ”stake” constitutes a bond that the leader will lose if she ”misbehaves.” Realizing Proof-of-Stake in practice has been challenging. Indeed at the time of this writing, no Proof-of-Stake currency has ever achieved widespread adoption. One of the main reasons for this is called the Nothing-at-Stake problem. In Proof-of-Work, equivocating by proposing multiple blocks to extend competing chains requires the consumption of additional energy and computational resources with real-world cost linear with the number of chains being extended. So to extend 5 competing chains simultaneously requires 5 times the real-world cost. Furthermore, any blocks that do not end up in the canonical chain will not receive the associated rewards and the cost incurred to produce them is permanently lost. This creates a strong incentive for a rational actor to mine on only one chain: the chain that the network will eventually reach consensus on, where their mining rewards will be honored. However, in existing Proof-of-Stake designs, the real-world computational cost of producing a block is deliberately negligible. As a result there is no additional cost associated with producing 5 competing chains. In fact it may be completely rational for a block proposer to extend as many possible system states simultaneously to ensure that they will put a block in whatever chain the network eventually accepts. Because of the Nothing-at-Stake problem, current Proof-of-Stake systems provide more limited guarantees of eventual consistency than Proof-of-Work. Any block proposer can trivially show equivalently valid system states to different clients or in the case of a staker (or group of stakers) at any point having a majority of the stake in the system, arbitrarily re-write long-established history. Existing Proof-of-Stake implementations have solved this problem by using a developer-operated coordinator (sometimes distributed) that endorses a canonical chain for clients to follow. In this work we propose a novel Proof-of-Stake consensus protocol that disincentivizes equivocation and does not require a centralized coordinator to operate.
File Format	PDF HTM / HTML
Alternate Webpage(s)	https://pdos.lcs.mit.edu/6.824/projects/ecto.pdf
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article