As cryptocurrency platforms continue to garner more attention and mainstream adoption, a key issue that has arisen has been the need for long-term scalability solutions. Ethereum has been at the forefront of this dilemma as its network has recently been plagued by network congestion. This heated a debate over how the platform should effectively scale to support its ambitious goals and lofty expectations.
Casper is a proposed protocol upgrade dating back to September 2014 that would essentially transition the Ethereum network into a Proof of Stake (PoS) consensus system through a series of incremented implementations. Currently, Ethereum operates as a Proof of Work (PoW) consensus system similar to Bitcoin. The first implementation of the Casper protocol in its alpha testing stage. If you’ve been following the crypto space closely, you are sure to have heard the term Casper come up many times and the subsequent debate revolving around the long term viability of Proof of Stake consensus models.
With a goal to provide a framework for the Ethereum platform to scale efficiently and securely, Casper aims to become one of the first mainstream platform PoS protocols. Regardless of your position on whether or not Proof of Stake is the right solution for long-term scalability of decentralized networks, Casper’s research, application, and resulting effects will be a useful case study for analyzing effective solutions for the larger industry’s scalability issues.
How Casper Protocol Works
The Casper FFG (Friendly Finality Gadget) protocol has been running in an Ethereum Testnet environment since the Metropolis upgrade. This implementation of Casper is a hybrid Proof of Work/Proof of Stake model where the PoS consensus mechanism functions as a finality system that overlays the existing PoW blockchain. The intention is to provide an additional layer of security against chain re-organizations while also allowing for analysis of the PoS mechanism’s functionality within the network without full migration to PoS on a live network.
Empirically, the Casper protocol was initially designed as a full migration of the Ethereum network from PoW to PoS. The way this system would function is by replacing miners with “validators”, who would basically perform the same functions as the miners through a staking system of the validators’ funds. However, these validators would not need to consume a real word resource (electricity for PoW) in order to secure and validate the network, but would instead stake their own personal ETH as collateral in order to provide an incentive to act honestly within the system. This represents a shift from using physical resources as a cost of network consensus to the system becoming entirely virtual through using virtual tokens as the implied cost of providing consensus.
While the amount of ETH needed to become a validator has not been fully established yet, lets use an example of needing 1,000 ETH to become a validator in this system. As a validator, you would lock up the 1,000 ETH as your stake to ensure that you act honestly. When discovering a block that can be added to the chain, you will place a “bet” on it. If the block is validated and successfully added to the chain, then you would be rewarded with an amount proportional to your bet. If you act dishonestly and say, attempt to validate blocks on competing chains, then your stake of 1,000 ETH (or a percentage of it) would become forfeit through a process known as “slashing”.
This system ensures that the validators have an invested interest in the success of the platform as they are receiving their rewards (in transaction fees) in the native token and have large sums tied up in the network (their stake). The creator of a new block is chosen in a deterministic way based upon the size of their stake. Where Casper differs from most other implementations of PoS protocols is in its punishment of validators that attempt to act maliciously.
A common criticism of PoS models is the concept of what is known as the “Nothing at Stake” dilemma. Essentially, as a validator it is possible to place a bet on two competing blocks that are produced at the same time. Each conflicting block can then build up stake with eventually one block having more stake and becoming the winner. The issue is that since staking tokens on a chain is not connected to a finite real world resource, validators can be incentivized to stake their tokens on both competing chains to increase their rewards. In a Proof of Work consensus model, this problem is inherently solved by the need to allocate real world resources to whichever block is built on top of one of the blocks and becomes the chain with the most work. This incentive is created through random process and has a cost (electricity) for not adhering to it.
Casper aims to solve this issue by slashing validators’ stakes if they attempt to take advantage of the “Nothing at Stake” problem. Clearly, the punishment is harsh enough to deter anyone from attempting to manipulate the system. Validators can even lose their stakes due to being careless or lazy with their node uptime, thus reducing censorship and helping the network to become more Byzantine Fault Tolerant.
Overall, the proposed Casper protocol has been somewhat dynamic in nature as developers attempt to adjust the protocol in order to improve it as more about PoS systems is learned and understood. The second implementation of the Casper protocol is the Casper CBC (Correct by Construction) proposal created by Vlad Zamfir. Its aim is to be a kind of dynamically adjusting protocol improvement by using an oracle known as an “Ideal Adversary” to constantly develop and adaptively adjust the protocol. The general idea is that the final version of Casper will be a combination of both the CBC and FFG implementations.
The current proposal of Casper is slightly different than what was initially designed. The system is used as more of a “checkpointing” protocol used every 100 blocks. The PoW mechanism still functions as the overall consensus model of the network and helps to mitigate many of the initial deficiencies seen with early PoS models that led to some important lessons on their long term dynamics. This version is a simpler rendering of the original framework. An often criticized aspect of PoS systems is that their long term viability and evolution necessitates ever increasing logical and protocol complexity.
Specifically, the protocol has a more profound focus on the Byzantine Generals Problem that can be leveraged to compromise the network. With the current proposal, if more than ⅔ of participants (validators) in the network are acting honestly, then the algorithm governing the model cannot finalize conflicting blocks. This goes back to decades of research into Byzantine Fault Tolerance consensus algorithms while providing some contemporary adjustments. The continued modifications of the protocol and proposed enhancements will be interesting to watch as Proof of Stake consensus models gain more traction.
If you’re familiar with the Casper protocol, then the most prominent name that comes to mind when looking at the contributors to the project is Vlad Zamfir. Along with Vitalik Buterin and Virgil Griffith, Vlad has been leading the way on the implementation of the Proof of Stake consensus model on Ethereum.
The Casper protocol has also attracted a significant amount of other developers and community contributors who are either PoS consensus proponents, or are just looking to help find a viable scalability solution for one of the cryptocurrency industry’s largest platforms.
Interestingly, the implementation of a large scale Proof of Stake consensus model has also attracted attention of academics and technical researchers in decentralized and distributed systems that are eager to apply crypto economics and various game theory permutations to an innovative field of study. VMware researcher Dahlia Malkhi has voiced some interesting takes on the direction of consensus models and the general feeling seems to be one of excitement as research in the area continues to accelerate.
The advantages of Proof of Stake models and specifically the Casper protocol are profound. Their initial lure as a solution to the increasing energy consumption of Proof of Work models, ability to allow for sharding capabilities, and lower latency all helped to distinguish it as a logical proposal to help solve large scale and long term scalability.
Most prominently and often discussed when analyzing PoS models is the underlying concern regarding the increasing energy consumption of PoW models like Bitcoin. These platforms can consume enormous amounts of energy. For instance, Bitcoin consumes more electrical energy than 159 countries. With environmental awareness at high levels, this has raised concerns among people not just within the cryptocurrency space, but also outside of it to people who have maybe only heard of cryptocurrencies in passing.
By transitioning the consensus structure to strictly virtual, the Casper protocol, if and when it is fully implemented, will remove the need for hardware miners, even electrical energy consumption in order to secure the network. The network consensus will be governed and secured by a decentralized system of validators who stake their native tokens and act in the networks best interest.
Another key advantage of the Casper protocol is its ability to open the door to and create the framework for sharding to become an intricate infrastructure layer of the Ethereum platform. If you are unfamiliar with sharding, it is basically a form of horizontal database partitioning that separates large databases (such as a massive decentralized blockchain like Ethereum) into smaller and more manageable pieces called “shards”. This concept has been proven prior to blockchains entering the picture and if applied correctly to decentralized networks, can vastly improve processing times.
Finally, PoS consensus models provide an important and often overlooked advantage that has given corporations and hierarchical institutions with vast funds an advantage for a long time. A PoS protocol like Casper eliminates the possibility of economies of scale affecting the platform’s decentralization. Economies of scale refers to the concept that enterprises with large scale operations obtain a unique advantage over smaller competitors by having their cost per unit output decrease with increasing scale.
Applied to the cryptocurrency space, this can be seen with PoW based networks like Bitcoin where large mining pools, due to having a larger share of the hashing rate, have a better chance of mining blocks than smaller pools and individuals. Combined with the presence of specialized ASIC miners created by centralized institutions like Bitmain, mining pools are able to spend the profits from their larger proportion of blocks mined on better equipment to help them maintain their edge and even increase it. This can lead to centralization of the network. However, this problem does not exist within a PoS model such as Casper since there is no mining and the consensus process is completely virtual.
Despite the excitement surrounding Casper and other PoS consensus models being developed and implemented, there has been a consistent and valid criticism of the underlying technology and logic about the network’s long term scalability and security using a purely PoS model.
Specifically, the issue of “Nothing at Stake” has been one of the major concerns when it comes to Casper, even with its modifications to mitigate against this problem. For instance, staking does not add to convergence of the system as it does with PoW models. In Casper, validators who stake on multiple chains are punished through the slashing mechanism. However, it is very possible for a validator to receive a block first while the majority receives another block first, so even if the initial validator is switches back to the other chain with honest intentions, that person is still punished by the slashing mechanism.
A proposed solution to this issue was to create multiple rounds of staking in order to minimize the punishment for this type of honest behavior. The problem with this approach leads to the fundamental criticism of PoS models that they need to become more and more complex over time in order to satisfy fundamental issues that will continue to cause problems.
Another concern of Casper is the unfair economic model of the underlying PoS model. Essentially, validators with larger wealth that is staked gain higher returns so “the rich get richer”. The amount needed to stake to become a validator in Casper has not been set yet but initial rumors and estimates put the number at 1,000+ ETH at least. This is supposed to eventually decline over time but the initial barrier for mainstream users to become validators will be huge.
The latest proposed version of Casper has mitigated some of these concerns by primarily using the PoW consensus mechanism and having a PoS model overlaid on top of it to help ensure against malicious miners by providing an additional layer of consensus every 100 blocks. However, the problem with this proposal as outlined by Bitmex recently is that validators “can reach the necessary ⅔ support for a lower PoW chain than the current PoW chain, a new way of causing a reorganization.” This essentially undermines the PoW system through what is seen as an arbitrary extra layer of security by some.
As research into Proof of Stake models continues to evolve and more is learned about their viability as a scalability solution, Casper remains an important flagship protocol upgrade to watch closely. It has some interesting and profound benefits as well as deficiencies that are important to the overall decentralized and distributed network industry improving in the long run.
Ethereum’s vision and network size is ambitious and their expectations from users and observers are even more so. With a loyal, open-source community behind the platform’s development, opportunities are abundant to help contribute to the platform and suggest proposals that could improve the underlying protocols further. If you’re looking to stay more informed on Casper or to help contribute you can find all the information you need available on the Ethereum Wiki.
The debate revolving around which consensus models are the best as long term solutions to scale will continue to persist. Whether you remain on the sidelines watching these developments or are actively involved in building and improving them, the opportunity for a novel solution to catapult decentralized platforms into their next phase awaits.