To address widespread misinformation and confusion surrounding Casper, Vitalik Buterin shared the history and state of Casper research, including the war between its two versions, design issues and more.
A deeper dive into consensus algorithms
As an alternative to proof-of-work, Casper belongs in the category of proof-of-stake algorithms. PoW establishes network consensus by measuring the amount of computational resources expended to support a particular chain of transactions. PoS seeks to replace physical mining with CPUs, GPUs and ASICs by measuring economic resources committed to a particular history.
PoS is believed to be superior to PoW in terms of security because the cost of an attack on a PoS system is much higher. An attack on a PoW system incurs the cost of electricity expended and hardware deterioration, which means it can be repeated once and once again with the attacker suffering only partial costs. PoS, however, does not allow for repeated attacks as ill-used resources are burnt as penalty, which provides a major disincentive for bad actors.
Two major flaws
However, PoS cannot be built simply by copying principles and intuitions from PoW. PoW uses computing power as its voting resource — a resource that once expended cannot be replicated. Therefore, in the event of a fork, miners need to commit to one, not both. A fork in a PoS system, however, creates a copy of a user’s current resources, which means that a validator can vote on both forks. When one of the forks is abandoned, the validator loses only forked resources and retains the ones in the main chain. This is called the Nothing at Stake problem.
2. Ethereum proof of stake research began in Jan 2014 with Slasher (https://t.co/ENIJsun6zn). Though the algorithm is highly suboptimal, it introduced some important ideas, most particularly the use of penalties to solve the nothing at stake problem (https://t.co/CbMXgQt0TI).— Vitalik Non-giver of Ether (@VitalikButerin) August 16, 2018
Long-range attacks are another flaw that pose a threat to a PoS system’s integrity. A 51% attack on a PoW system is executed by creating a longer fraudulent chain that can go back up to 6 blocks. This limitation is imposed by the amount of computational power attackers can potentially have at their disposal. There are no such limitations in PoS systems, and an attack can start far deep into the chain. This is why it is called a long-range attack.
6. If the attack chain diverges from the main chain at a fairly recent point in time, this is not a problem, because if validators sign two conflicting messages for the two conflicting chains this can be used as evidence to penalize them and take away their deposits.— Vitalik Non-giver of Ether (@VitalikButerin) August 16, 2018
Casper to the rescue
Casper seeks to eliminate these threats. At one point in time, however, Casper research split into two. Vitalik and his fellow developer Vlad Zamfir had certain disagreements on how to cope with these flaws. Thus, two versions of Casper were born: Vitalik’s Friendly Finality Gadget Casper and Vlad’s Correct-by-Construction Casper.
A PoS protocol is still to be fully implemented on the Ethereum network, but we can hope that Casper will inherit only the best out of both of its versions.