The Casper Proof-Of-Stake Algorithm: Prepare To Commit

The measure of intelligence is the ability to change. – Albert Einstein

When the creators of Ethereum first grasped how underutilized the Bitcoin blockchain was, a breakthrough occurred. That breakthrough was the conceptual framework behind the idea for the Ethereum blockchain, a platform for decentralized applications and the nexus point for what could evolve into a new type of internet. Although Bitcoin’s first-mover advantage created the modern crypto-ecosystem, Ethereum has perhaps shown the world more of what’s possible in the space. The dichotomy between the two most prominent cryptocurrencies has created entrenched maximalists on both sides and both platforms are still evolving. However, there could not be a more stark contrast between their evolutionary paths.  

Bitcoin is being forced into creating reactionary solutions to address the challenges of its own growth. Limited by technological locks in its core architecture, Bitcoin has seen its market capitalization slashed nearly in half this year. Alternatively, Ethereum has grown in excess of 3000 percent since the start of 2017. While there may be various factors that contributed to Ethereum’s growth, a portion of its success must be attributed to the calculated development plan laid out by its founders.

Whereas Bitcoin was designed in relative totality, Ethereum developers have taken a marked approach to growth by breaking down monumental challenges into smaller, more achievable goals. One of these goals has always been to marry the advantageous attributes of Proof-of-Stake algorithms with an economic mechanism that incentivizes users to reach consensus. Researchers and developers of Ethereum have been working on this topic for over two years now. Their solution is lightheartedly called Casper, after the cartoon ghost, because the new protocol is adapted from principals of the Greedy Heaviest-Observed Sub-Tree (GHOST) protocol.

Reaching Consensus In Distributed Systems

Blockchain technology is not necessarily as new as the current hype would have us believe. Generally speaking, as a data structure, it shares similarities to the linked-list developed by the RAND Corporation in the mid-1950s. Moreover, distributed systems research has been a field unto itself within computer science since the DEC Systems Research Center in Palo Alto, California created the first internet-based distributed computing project in 1988. Taken to the limit of what is historically credible, distributed systems can actually trace their origins further back to the top-secret precursor of today’s modern internet, The Advanced Research Projects Agency Network, or ARPANET. This computational network, run by the United States Department of Defense, was designed in the early 1960s as a distributed system primarily so that it could survive a nuclear war. Today, our globalized civilization utilizes distributed systems to facilitate a massive amount of infrastructure. Telecommunication networks, massive multiplayer online gaming, aircraft control systems, peer-to-peer networking and the World Wide Web are all examples of distributed computing systems.

All distributed computational systems essentially facilitate communication between a network of computers. This communication is important because the baseline idea in distributed computing is to link two or more computers together to accomplish a common objective. If the network of computers is not synchronized uniformly, the network cannot function. Anyone who has ever played a videogame online has undoubtedly had to deal with latency issues. When the game you are playing loses its synchronicity with the other computers on the network the user experiences every adolescent gamer’s nightmare – lag. When the computers running the game are functioning harmoniously, you could say that the network is in agreement, or functioning in “consensus.” This concept of consensus is fundamental to understanding how Ethereum works as a decentralized blockchain platform, and how its method of reaching consensus is about to change.

Proof-Of-Work & Pre-Casper Consensus

As a type of distributed system, blockchains have four primary ways of arriving at a state of network consensus: the Practical Byzantine Fault Tolerance Algorithm (PBFT), the Proof-of-Work algorithm (PoW), the Proof-of-Stake algorithm (PoS), and the delegated Proof-of-Stake algorithm (DPos). Prior to Bitcoin in 2009, cryptocurrencies capable of interacting with cryptographic protocols hadn’t yet been invented. This is why Proof-of-Work was initially utilized by blockchain enabled cryptocurrencies. You could start with nothing and, by solving the cryptographic puzzles that unlock block rewards, you could “mine” that block and reap the prize. Other users on the network could simply verify the work performed by your computer to solve the puzzle and the entire data record could then be appended to the blockchain database for immutable record keeping. While this is a novel approach to reaching consensus about a canonical chain, there are significant pitfalls associated with long-term PoW mining.

Proof-of-Work doesn’t actually require a user to start with nothing. Technically speaking, before a user can mine a block, they must input the electricity required to run their computer. They must also spend their computer’s computational cycles in order to crack the puzzle to unlock blocks. Both of these resources are forms of expenditure. Crunching the numbers on PoW reveals that global mining operations draw nearly the same amount of electrical power as Ireland. Considering that all of this valuable electrical power is being expended to solve trillions of SHA-256 computations every second, the electricity expenditure isn’t just inefficient – many perceive it as unethical.    

The Advent of Validators

Casper is a unique approach to PoW’s strongest alternative, PoS, in that it is a combination of PoS methodologies. The Casper algorithm will shift Ethereum into being a security deposit-based Proof-of-Stake system that secures consensus when networked users lock up an amount of their virtual currency, their “stake,” to gain proportional voting power. As defined by Casper’s minimal slashing conditions, the idea of “staking” virtual currency to incentivize honest behavior within the system is ideal because bad actors lose their staked investment, along with their ability to vote and participate in the network. This is a superior methodology because it not only solves the inefficiencies of PoW, but increases security by making attacks on the system expensive. This is one of the reasons why Casper is a true accomplishment for Ethereum; the best attributes from several PoS methodologies have been uniquely adapted to update the network from a technical standpoint, while incentivizing users to act properly within the network.

The Casper PoS algorithm will work around the pitfalls of PoW mining while achieving many of the same results in a different way. Several familiar terms have been rebranded to describe their new role in the updated system. Miners, as they will no longer be mining blocks, have taken on the so-called role of a “bonded validator.” This new role not only serves the network by providing a path to consensus, it also solves the “nothing at stake” problem by ensuring that misbehavior is economically punished.            

The role of a validator will be to gather as much information as possible from the system about the transactions of other validators. Each validator will attempt to stay as up-to-date as possible on the stakes other validators are using to bet on blocks, compile the data, and produce a block from its own node. The validator will cryptographically sign the block it produced and send it out into the network. Validators can bet their stake at every block height by speculating the probability of a particular block being appended to the canonical chain.

Once a supermajority (somewhere between 67-90 percent) of bonded validators have all converged around a block with a high probability of being added to the canonical blockchain, the fork choice rule will determine the exact conditions that the canonical chain will be made under. This protocol will be run by every validator helping to ensure that bets are placed on legitimately forming chains. Should a node-validator place stake into a false or malicious chain, that stake will be destroyed.

Casper Implementation

Although the date for implementation is still pending, we now have a better idea of how things will change. The first implementation of Casper will be a true hybrid between PoW and PoS. Throughout this changeover, miner rewards will be decreased and validator rewards will be increased to help incentivize compliance. Work is being finalized on creating a Casper full node and deploying a Casper contract onto the main Ethereum network. Once transactions start to be validated, true regression testing can begin via ecosystem peer review. After testing, there will come another hard fork, most likely the update for Serenity, where the true shift in consensus algorithms will take place. This change will provide a foundation for future security and scalability of the Ethereum platform and its debut remains highly anticipated.