By Research Team
The piece can be downloaded as a PDF or read in its entirety below. The rest of the series can be found here.
Sam Lee – Director of Research
Chris Robichaud – Research Analyst
Tao Tao He – CEO of HGR Digital Asset Group
For questions, contact Sam Lee (email@example.com)
- Despite its ubiquity, Ethereum faces scaling and other challenges that pose a significant risk to its future adoption and success.
- Governance disagreements are becoming cumbersome as the number of Ethereum stakeholders grows. Existing community polarization could prompt more hard forks, similar to 2016’s DAO fiasco.
- The Casper upgrade will switch Ethereum’s core consensus mechanism from a computationally expensive Proof-of-Work system to a more efficient but unproven Proof-of-Stake system.
- Casper further enables incremental scaling solutions such as sharding, The Raiden Network and Plasma, but these all have hurdles to overcome as well.
- Proof-of-Work miners may not migrate to the Proof-of-Stake system post-Casper. Instead, they may choose to mine a different Proof-of-Work coin altogether, reducing the size of Ethereum’s following.
- The stakes have never been higher for the Ethereum project. The success or failure of various solutions are important factors to consider when thinking about a long-term valuation model for Ether.
Part IV: Voyage of the Ether
- Key Takeaways
- Scaling Ethereum for Global Adoption
- Arguments for POS vs POW
- Arguments for POW vs POS
- Appendix: Timeline of Ethereum’s history
Ethereum’s open-source code base has enabled the project to crowdsource innovation from the masses, heightening expectations for its impact. Now more than ever, Ethereum has the most to lose – a vibrant development community and a large investor following epitomize the platform’s progress and reputation to date. Still, existing hurdles pose substantial risk to the technology’s success.
In 2018, Ethereum is slated to face what many consider to be the most challenging and defining moment of its young existence. The implementation of the Casper upgrade will involve switching from a Proof-of-Work (POW) to a Proof-of-Stake (POS) consensus mechanism. It is a wildly polarizing topic within the community. Miners who have invested in POW machinery purport that it is unwise to abandon a battle-proven POW system for an unproven, more complex POS system. Activists and developers suggest that the elimination of energy costs and other benefits more than offset these risks.
The very nature of the Ethereum mining community poses challenges to a smooth POS transition. Ethermine and Nanopool, POW mining consortiums that are responsible for approximately two thirds of Ethereum’s devoted hashrate, have effectively established a duopoly on Ethereum’s POW mining operation. Despite developer efforts to reduce the profitability of POW mining, pools such as Ethermine and Nanopool are incentivized to keep the system the way it exists today to drive additional return on their mining machinery investments.
As recently as April 2018, mining equipment provider Bitmain Technologies announced an Ethereum version of its POW application-specific integrated circuit (ASIC) hardware, a decision that is particularly notable in the context of Casper’s planned switch to POS. The announcement somewhat brazenly suggests that a future version of Ethereum may run using POW, a situation that could only emerge with either  a successful hard fork or  a failure to integrate Casper.
The Monero cryptocurrency’s recent history provides a proxy for understanding how the addition of ASIC mining technology can disrupt a crypto community. When Bitmain announced its Monero ASIC release, Monero developers became concerned that it would centralize mining operations. To avoid this, multiple developers independently modified the Monero code to prevent ASIC technology from being utilized, resulting in a hard fork that created not one, but five different Monero cryptocurrency projects (Monero Zero, Monero Original, Monero Classic, Monero-classic, and Monero).
Time will tell whether a similar forking phenomenon will occur for Ethereum. Bitmain Technologies generated between three and four billion dollars in 2017 profits. With this much capital, Bitmain could theoretically continue supporting development of the POW version of Ethereum by itself, even without the consent of Ethereum leader Vitalik Buterin.
To complicate matters further, the Ethereum community is concurrently experiencing flashbacks of its 2016 DAO hard fork debacle. A 2017 Parity Technologies wallet hack of $264 million has many people calling for a separate hard fork to return stolen funds back to the rightful owners. Again, this topic has been met with contentious opinions from many members of the Ethereum community.
However, regardless of the disagreement in question, the important theme to note is that blockchains do not do well when their community is divided as clearly as is the Ethereum community’s today. Polarization leads to internal dissent and ultimately, to hard forks. As the Casper integration looms, many question whether Ethereum can sustain yet another split that would create even more new competitors, as was the case for Ethereum Classic in 2016.
While the recent influx of users and capital are positive signs for Ethereum’s global adoption, an effective scaling solution will need to be implemented in order to support commercially viable decentralized applications (dApps).
In its current state, Ethereum can support approximately 15 transactions per second, meaning that the platform is still several orders of magnitude away from transaction speeds that will sustain the load of more than one dApp that processes a million transactions per day or more. CryptoKitties, a dApp which allows players to breed and trade cartoon cats on Ethereum, highlighted these scaling issues in a direct manner. As a result of its adoption, the number of unprocessed transactions in queue rose from an average of 2,000 to more than 10,000, causing significant delays in all transaction confirmations. While Ethereum may not be able to fulfill its founders’ global philanthropic vision in its current form, a variety of different projects and proposals exist to optimize and scale the platform in pursuit of this goal.
Casper is the transition mechanism that will eventually migrate Ethereum to a POS consensus method – a change the Ethereum Foundation believes is necessary to solve the various problems associated with POW. This modification will mark a monumental shift in one of the core technological elements that defines Ethereum today, and its execution will likely have large implications for Ethereum’s future.
Casper will be implemented in an incremental format, meaning that POS mining rewards will be slowly increased over time while POW mining rewards will be slowly decreased through planned mining difficulty bombs,  such as the one contained within the Constantinople update.
More broadly, Casper will enable the concept of finality, which is necessary to make the Ethereum blockchain truly immutable. Transaction finality will theoretically make it more difficult to rewrite Ethereum’s history, helping to prevent future splits of the Ethereum community such as the one that resulted from the 2016 DAO hack.
To further delve into the concept of finality, it is important to understand the “nothing at stake” problem associated with POS consensus mechanisms. POS establishes consensus by fielding bets from validators related to which transactions are believed to be valid. Because of this, there is a chance that a POS validator bets on the wrong set of transactions and therefore does not receive an Ether reward. To ensure they receive a reward, POS validators could choose to validate all proposed transactions regardless of which proposed transactions they actually believe are correct. In this scenario, the validators’ incentives are not structured appropriately because they are not necessarily preserving the correct transaction history, threatening the blockchain’s concept of immutability.
In addition, Ethereum is coded to build upon the longest chain available in the ecosystem. One potential attack for both POS and POW involves a miner maliciously overwriting the longest chain by disconnecting from the internet to create empty, transaction-less blocks offline. Empty blocks can be mined more quickly than their online counterparts, which take longer because transactions require additional time to fill and process. Upon reconnection to the internet, the miner’s empty block version will be longer than the actual blockchain, and therefore, Ethereum will continue to build new transactions on the incorrect, longer version. In the process, valid transactions are purged from the blockchain. This attack effectively spoofs the system into an illegitimate hard fork.
In order to avoid these issues, Casper punishes any validator who goes offline or incorrectly proposes a modified blockchain history through the use of finality checkpoints, which is a key part of the Constantinople upgrade. After every 100 blocks, a finality checkpoint will be formed, at which time POS validators will choose to send some Ether into a smart contract for the chance to be rewarded for confirming future transactions in consensus with other validators. Two-thirds of validator node votes are needed to finalize transactions associated with each finality checkpoint. Validators that correctly validate are rewarded with Ether. If a validator proposes a transaction history that is not in consensus with all finality checkpoints, that validator’s Ether will be confiscated and destroyed. This way, miners are incentivized to protect blockchain history with the promise of new Ether upon correct validation, and at the cost of losing their Ether in the event of malicious behavior.
Casper is expected to reduce Ethereum’s transaction block time from its current 15 seconds per block to anywhere from 2 to 7.5 seconds per block, providing a direct increase in transaction capacity on the network. Additionally, Casper will lay the groundwork for sharding and other major scaling improvements that are intended to further increase transaction speeds.
In order to validate transactions on the blockchain, Ethereum’s nodes need to reference every past transaction as well as process new transactions from each block. Although this is a secure approach, it has some unfortunate scaling properties. One immediate theory on how to improve Ethereum’s transaction speed is to increase the number of transactions processed per block, which is currently limited by a block ‘gas limit’ or fee. Miners have the ability to increase or decrease the block gas limit by a factor of three to help with throughput delays during times of large transaction volume. The problem with allowing miners to increase the block gas limit more than three-fold is that it would also increase the computing power necessary to process each block, precluding consumer grade hardware from mining and therefore centralizing network consensus to professional crypto-mining operations. Adding more nodes to the network would not ameliorate the issue because each node needs to individually process transactions as part of the consensus mechanism. With the current process, new blocks of transactions are created and added to the blockchain through the process depicted below:
One of the Ethereum community’s plans to address this prohibitive processing power issue is called sharding. The idea is that by splitting Ethereum’s nodes and transactions into smaller groups called “shards,” all running separately but concurrently, the entire network can function much more efficiently. Each transaction is assigned one shard with which it interacts. This approach increases the number of transaction processing queues that exist for the blockchain – instead of having one queue for many transactions, the blockchain would have many queues for many transactions, helping to balance the processing load. Nodes within each shard would only need to verify the small fraction of blockchain history that is relevant to that shard before processing new transactions.
Interacting across shards will likely require a longer and more complex method of confirmation. However, overall network speed would theoretically remain high because cross-shard interaction is expected to be less common. That said, the Ethereum Foundation is interested in developing intra-shard operations before trying to solve inter-shard logistics. If implemented correctly, an effective sharding solution could improve transaction throughput by many magnitudes without sacrificing security or decentralization, unlike other solutions that involve using altcoins to process off-chain transactions or using more powerful centralized masternodes to increase speed.
Sharding plans are still under development and are likely months or years from implementation. Because a specific approach has not been finalized, an accurate estimate of improvement and potential challenges is not realistic. In addition, development teams will need to address skepticism from the community related to the core premise of sharding’s usefulness as a scaling solution:
“The intent is to relieve the amount of work a single validating node must do so there can be more of them, but it only results in prolonging the issue [of requiring full blockchain validation at a later point], and not fixing the problem.”
The Ethereum Foundation is actively working on creating formal specifications and identifying teams to implement and deploy sharding solutions onto the testnet. Grants ranging from $50,000 to $1 million are available to various groups that dedicate time to facilitating the development of protocols such as sharding.
The theory behind the Raiden Network is similar to the theory underpinning Bitcoin’s Lightning Network in that both use designated transaction channels to alleviate challenges associated with high transaction volumes.
Individuals who transact using Ethereum’s platform have the option to set a max gas limit fee they are willing to pay a miner. In times when a transaction backlog develops, miners will select those transactions that offer the greatest fees, and other transactions will be deprioritized. Therefore, as the number of network transactions increases, a transaction backlog develops, miners select the most lucrative transactions to process, and average transaction fees rise, as depicted in the simple scatter plot and linear model below:
The Raiden Network intends to ameliorate this issue by initially recording transactions externally to Ethereum (i.e., off-chain) via smart contract functionality, hence why it is known to developers as a ‘second layer solution.’ At a later time, the net result of all off-chain transactions is updated onto the main blockchain. This approach reduces the number of times transactions are recorded onto the primary Ethereum chain, therefore reducing the congestion on the network. Transactions sent though Raiden’s payment channels are designed to be fully confirmed within one second, a marked improvement from the 30 to 180 second range that exists on the Ethereum blockchain today.
As demonstrated in the illustration below, Raiden’s technology works through what are known as ‘state channels,’ which are smart contracts that enable secure off-chain transfers without blockchain consensus. Effectively, these channels allow for Ether to be sent back and forth between two parties multiple times before any blockchain update is finalized, letting transactions to be recorded on a ‘net’ basis.
As seen in the above figure, Alice and Bob transact between themselves outside of the Ethereum network. It doesn’t matter that each transaction is not recorded on the main blockchain as long as both parties agree on the correct balance to be written to Ethereum when one party finally chooses to close their state channel.
In order to use the Raiden Network, participants must move some amount of Ether off-chain by placing it into escrow. This off-chain Ether is then shifted back-and-forth between participants according to smart contract logic until the channel is closed. The requirement for placing tokens into escrow means that opening unique payment channels for every counterparty relationship could lead to liquidity constraints associated with multiple Ether deposits. However, Raiden addresses this challenge by enabling ‘multihop’ transfers, which provide the ability to route transactions through pre-existing state channels shared by both parties in a particular transaction. For example, if Alice wants to pay Bob, she does not necessarily need to open a new state channel with Bob if they both have another channel already open with a common receiver. While there are no mining fees associated with state channel transactions, intermediaries facilitating multihop transfers will likely charge a fee for providing access to the Raiden Network. Initial estimates suggest that multihop intermediary fees will be much lower than on-chain transaction fees.
In addition to the Raiden Network’s cost and scaling implications, its design enables some interesting privacy properties. Because transactions are only saved onto the blockchain when a payment channel is closed, only the net payment is visible on-chain. More importantly, most transactions are expected to go through multihop intermediaries, meaning that individuals can conceal their counterparty’s Ethereum address.
Though Raiden has great potential, it also faces several challenges of its own. Raiden is not well suited for bigger transactions. Escrow requirements necessitate that individuals deposit large amounts of Ether to engage in large transactions, potentially preventing individuals that have staked smaller deposits from participating. This suggests that many will turn away from the Raiden Network for transactions of magnitude. In addition, Raiden will need to generate significant network effects to become useful for the Ethereum community. If many people adopt Raiden’s solution, the larger community will enjoy the benefits of easy access to fast and cheap payments with many counterparties. However, without mass adoption, these benefits could be muted, causing users to view the time and monetary costs of setting up Raiden payment channels to be too high, ultimately preventing the solution from succeeding.
The Plasma solution is related to Raiden because it also allows for off-chain transactions that reduce network congestion. However, the manner in which those off-chain transactions are managed differs. Using Plasma, the main blockchain records when an asset has moved off-chain, and at that point a new sub-blockchain begins to manage the asset. Vitalik Buterin and Joseph Poon, the creators of the Plasma solution, have dubbed these ‘parent’ and ‘child’ blockchains.
Child blockchains operate using similar technological concepts employed by the parent blockchain. For example, consensus mechanisms, transaction fees, and distributed validation processes are still core components of child chains, but each may be customizable to some extent and will be executed on a smaller scale when compared to the parent.
Each child continues to build its own transaction history separately but in unison with other child blockchains. Periodically, the child chains will update the parent chain with their version of new transactions, thus preserving the integrity of the overall blockchain’s history while improving scaling ability. Further, Plasma will enable the ability to create multi-layer blockchains, meaning that a child chain could create another child chain, further enhancing the scaling on a sub-chain itself.
The addition of many smaller blockchains will inevitably segregate the mining community, leading to more mining pools operating on child chains. With this comes the risk that miners will be able to commit fraudulent transactions on a child chain and memorialize them on the parent chain. Plasma addresses this by adding safety mechanisms into the code. For example, Plasma requires that an individual’s assets originate on the parent chain so that user can always withdraw their assets to the main chain in the event of an attempted theft.
This type of complex security functionality, as well as other mechanisms to ensure security such as fraud proofs, child censorship, and Raiden-style state channels, are still under development today. Though a realistic timeline has not been finalized by OmiseGo, the lead Ethereum development team for Plasma, we remain more optimistic about Plasma’s potential than that of sharding or the Raiden Network. Plasma’s ability to preserve the blockchain characteristics of immutability and censorship resistance at an individual transaction level promises more transparency than the Raiden Network. In addition, the structural simplicity of Plasma’s sub-chain approach appears easier to implement when compared to the cumbersome process of splitting an active blockchain into different sections associated with sharding. Further, existing sub-chain solutions such as Ardor demonstrate that blockchain communities have had success implementing similar parent-child scaling solutions.
While many Ethereum developers are working on Ethereum’s aforementioned long-term scaling improvements, there is also significant activity around smaller, more immediate upgrades to the Ethereum Virtual Machine (EVM) – a turing-complete computing environment that enables external smart contract execution when compiled into the existing bytecode. Proposed advancements would reduce the size of today’s compiled bytecode, subsequently improving optimization and decreasing the gas and space required to process a block of transactions. While such changes will not improve processing scalability to the extent of the above solutions, its implications will be incrementally helpful.
Vyper also removes certain features that Solidity contains, such as operator overloading, modifiers and class inheritance. While these features have traditionally helped developers create better architecture and design, they also provide avenues for cybercriminals and corrupt developers to hide malicious code. Today’s headlines around improperly designed smart contracts and exchange hacks accentuate the need for security within smart contract coding itself. Vyper’s approach prioritizes security and therefore could improve the likelihood of long-term smart contract adoption on Ethereum. Smart contract integrity that is verifiable through human-readable and auditable code will be paramount in enabling smart contracts to replace, for example, lengthy legal documents. In this sense, Vyper’s simplification of coding language and improvement in security could eventually help to secure Ethereum’s future.
To be successful, Ethereum must provide economic incentives for people to allocate their computers’ processing power to Ethereum. In the existing POW system, that incentive comes in the form of miners solving complex cryptographic puzzles for the chance to receive Ether payment. Meanwhile, their computers are simultaneously verifying new transaction blocks to maintain the blockchain’s consensus, an approach that has several well-documented drawbacks.
Expanding upon our description of POS and POW in Part II: Return to Classic, solving these POW cryptographic puzzles requires an exorbitant amount of electrical power. The mining of Ether alone consumes about $1.2 billion in electricity per year.
“The top 500 most powerful supercomputers on Earth combined only have 4% of the processing power of the Ethereum network, which wastes 99.99999999999997% of its energy hashing random numbers.”
Ethereum aims to address this potentially unsustainable construct by moving to Casper’s POS system, which uses much less energy through its staking process.
POW mining is centralized because of barriers to entry created by the high computing power needed to be profitable. Regular desktop computers do not have the specialized, advanced, and expensive hardware to realistically compete. This has led to mining pools, or networks of computers that aggregate mining power in order to collectively reap crypto rewards, which are then shared between participants. Proponents of POS argue that it will help decentralize the mining process because prohibitive processing power requirements will be eliminated, allowing smaller validators to thrive alongside larger ones.
Additionally, when a validator stakes more coins during the POS consensus process, a reward proportionately large compared to the validator’s stake in the total Ether pool is received. This is different from POW, where laying out additional capital can theoretically create disproportionate advantage through acquisition of better mining equipment relative to the community’s equipment. In this regard, POS inherently encompasses a hedge against further centralization risks related to quantum computing breakthroughs.
POS proponents argue that economic security is preserved because each validator must deposit their own ETH into a smart contract before the next block mining process begins. Validators have lower incentive to maliciously attack something in which they have a personal stake, especially because POS is programmed to systematically destroy an attacker’s Ether when a malicious attempt is recognized.
For example, Vitalik refers to an attack called “spawn camping,” whereby a malicious POW miner with more than 50% of the network hash power can repeatedly send bogus transactions to the network, preventing it from processing legitimate transactions and rendering the network useless. With Casper, malicious attackers cannot spawn camp without destroying their own Ether every time. Repeated, large-scale attacks of this nature would theoretically reduce Ether supply and therefore increase the cost of buying ETH, making the attack more expensive every time it is executed. This POS feature serves as a deterrence to denial-of-service style cyberattacks.
POS will also allow faster transaction processing, both directly and indirectly. Without time consuming computation to be executed, the block time on Ethereum can be cut by an estimated factor of 50% to 90% of existing times. In addition, POS provides important infrastructure upon which additional scaling solutions can be built, including sharding, Plasma, and/or Raiden Network options.
As described in the Casper section above, transaction finality requires that the nodes in the Ethereum network all accept the blockchain’s history at periodic ‘finality checkpoints.’ This prevents some of the nodes from diverging into an alternate chain via a hard fork and helps to keep the community unified. The implications are important to preserving transaction immutability, one of the core value propositions of blockchain technology.
While Casper brings a variety of benefits to the table, there are some associated challenges as well. We have outlined several below to provide an overview of some points that dissenters reference as shortcomings of the POS system.
During POS consensus, Ether transaction fees and validation rewards are distributed amongst the participating validators in accordance with the percentage of Ether that each validator contributed to the total staking pool. For example, if a validator stakes 2,000 Ether to participate in the POS consensus process, and there were 40,000 total Ether contributed to the entire POS consensus staking pool, then the validator will receive 2,000/40,000, or 5% of the resulting transaction fees plus 5% of new Ether rewards issued as part of the next block’s consensus. Given that rewards are proportionate to contributions, this type of system could favor those validators who have already acquired a large proportion of Ether. Said another way, validators that have a disproportionate amount of Ether today can more easily maintain their influential position. As such, some Ethereum followers claim that this leads to a ‘rich-get-richer’ type of scenario, whereby pre-existing centralization within the Ethereum crypto-economy could not be easily overcome.
Others argue that this is not a problem, but rather, is the norm for any free market. Those validators with strong conviction and a greater risk appetite will be rewarded for staking large amounts of Ether in an unproven POS system.
Casper relies on achieving a super-majority vote of two-thirds of validators, creating a risk that the chain may not come to consensus when the number of malfunctioning or misbehaving validators rises above a third. Compared to POW, this means that a malevolent node or coordinated group of malicious nodes need only acquire 33% of circulating Ether to wreak havoc on the system, which is lower than the 50% required by POW. This effectively lowers the threshold for fault tolerance across the entire Ethereum system. That said, this risk is somewhat mitigated by the fact that a successful attack would require the perpetrators to hold an untenably large amount of Ether.
Somewhat like a poker table in Las Vegas, there will be a minimum stake required to take part in the POS betting process. A party will need to own an estimated minimum of 1,250 ETH (possibly changing to 1,500 ETH) in order to be eligible to participate, a value that will likely be prohibitive for many Ethereum community members. In our opinion, if the Ethereum community is to create a truly decentralized system, barriers that restrict participation in the blockchain’s core consensus protocol should be removed or modified to be less onerous.
Also, Casper may include withdrawal delays of up to four months. As such, price volatility of Ether may expose validators to significant liquidity risk during that period – a very dangerous proposition for short and medium-term Ether holders.
While there are many whitepapers and Reddit feeds that granularly explain the complex nuances of POS pros and cons, the details of its final implementation are still largely under development. There appears to be an existing sentiment within the community that Casper and POS will not be fully understood by validators, which is reflected in a Tuur Demeester Medium article:
“[Casper provides]…a maze of rules to prevent abuse [that] will likely result in a much more fuzzy definition of what abuse really is…” 
Casper’s development team is trying to predict and prevent every type of attack that is feasible for a POS system, a noteworthy and admirable goal. However, the truth remains that new attacks targeting other aspects of the POS system will likely arise as the economic reward for hacking the Ethereum system increases. ETH developers will need to create complicated fixes to new threat vectors, thereby adding even more complexity to the system, reducing transparency for validators who are not subject matter experts, and potentially deterring new validators from joining the Ethereum network. Whereas the POS system is new and uncertain, the POW system has been tested for more than 10+ years, avoiding several POS-specific attacks with its simple, yet powerful economic incentive system.
Ethereum’s development plans call for scaling solutions to be implemented in 2018. However, even with the best and brightest minds in the field working over multiple years, a seamless solution has yet to emerge. The Casper POS upgrade will add a lot of complexity to Ethereum’s system when compared to the current battle-proven, yet imperfect POW consensus mechanism. Further, investment risk associated with illiquidity of Casper’s staking process could prove to be more impactful than many realize if validators are not willing to accept Ether’s high volatility for month-long lockup periods or more.
The uncertainty of the regulatory environment has limited the explosive growth of Ethereum-based Initial Coin Offerings (ICOs) and threatens the blockchain’s core anonymity value proposition. The ‘know-your-customer’ and ‘anti-money-laundering’ regulatory requirements appear poised to become necessary evils for the future of many blockchain communities. While cryptocurrencies reduce frictions and increase privacy for regular online commerce, they can do the same for international criminal activity as well. Governments will be hard-pressed to accept a technology that allows for easy and untraceable illicit trade without some mechanism of centralized oversight.
In this sense, private blockchain solutions such as Hyperledger Fabric, an open source project originating from IBM, could become notable competitors to Ethereum despite the fact that private blockchain solutions are typically not associated with those of public blockchains. Companies like IBM have the resources, leadership, and talent to build powerful technical solutions on the blockchain, albeit in a centralized fashion. However, if the long-term reality for the crypto space involves regulation, then this implies that today’s underlying public blockchain projects must also be centralized to some degree. Ethereum will need to find innovative ways to balance institutional oversight with user privacy, a non-trivial task because it involves competing with some of the world’s greatest for-profit, centrally managed technology firms.
Though these challenges appear to paint an ominous picture, Ethereum development teams are well-incentivized to find meaningful solutions. Many prominent Ethereum developers are compensated, at least to some extent, in Ether. Without ameliorating these pressing and potent issues, they stand to lose much of their own personal wealth.
The stakes have never been higher for the Ethereum project. If the switch to POS does not go smoothly, Ethereum risks forfeiting much of its multi-billion dollar market capitalization as well as its head start on competitors in the race to become the web infrastructure of the future. These are all important factors to consider when thinking about the potential long-term value of Ether, which will be the topic of our final edition.
Below, we explore the timeline of the Ethereum project through the lens of its developer community. The organic evolution of the protocol has subjected Ethereum to various high and low points over the span of its lifetime. The main events, from our perspective, are reflected in the following chronology:
2013 – Ethereum Whitepaper
Vitalik writes the Ethereum whitepaper, finalizing the initial Ethereum concept.
January 2014 – Formal Announcement
The Ethereum project is formally announced at the North American Bitcoin Conference in Miami.
April 2014 – Ethereum Yellow Paper
Vitalik teams with Gavin Wood to publish the Ethereum yellow paper, an outline of the technical specifications for the Ethereum Virtual Machine (EVM) – a computing environment for smart contracts.
July 2014 – Ethereum Foundation
The Ethereum Foundation is registered as a non-profit in Switzerland. Its purpose is to oversee the management of funds raised to promote and develop the Ethereum ecosystem.
July 2014 – Ethereum Public Sale
Ether’s public sale goes live, raising more than $18 million in Bitcoin from July to September. The funds are used to pay legal fees and developers, as well as to finance ongoing development.
September 2014 – ETH DEV Non-Profit Status
Ethereum’s founding development team, known as ETH DEV, concludes that Ethereum will continue to be a not-for-profit organization. ETH DEV directors Vitalik Buterin, Gavin Wood, and Jeffrey Wilcke begin developing Ethereum proof-of-concepts to be evaluated by the broader developer community.
November 2014 – DEVcon0
DEVcon0, a five-day convention in Berlin, gathers developers from around the world to discuss the Ethereum project. Ideas from this convention become the foundation for important improvement initiatives related to reliability, security, and scalability. The convention brings the global developer community together for the first time, strengthening bonds within the project and aligning contributors with Ethereum’s broader goals.
January 2015 – Bounty Program
Ethereum’s bounty program is released. The program directly rewards developers with Bitcoin for finding bugs in the protocol or in any existing Ethereum clients.
April 2015 – DEVgrants
The DEVgrants program is introduced to offer funding and support for developers who are adding value to Ethereum’s core software by improving common services, application programming interfaces (APIs) and application binary interfaces (ABIs).
May 2015 – Olympic Testnet
The Olympic Testnet, Ethereum’s first public development sandbox, is released after ETH DEV shares its ninth proof-of-concept with the community. ETH DEV encourages the developer community to push the Olympic Testnet’s limits by implementing a rewards program meant to incentivize public stress testing.
July 2015 – Frontier
The Frontier release marks Ethereum’s ‘beta’ version. Features added with the Frontier upgrade include the ability to mine real Ether, transfer funds between addresses, and explore block history. Most importantly, Frontier enabled creation and deployment of smart contracts.
The upgrade was targeted at Solidity-proficient developers. Solidity is a programming language developed specifically for Ethereum smart contracts. The Frontier period gave miners the opportunity to set up mining rigs and foster a stable environment to test new applications. Developers took advantage of this functionality and started to create various dApps on the network, publicly demonstrating one of the network’s most powerful use cases and value propositions.
November 2015 – DEVcon1
The DEVcon1 conference promoted wider public adoption and corporate support for Ethereum by showcasing the project’s potential and technical development to date. The five-day event included more than 100 presentations, panel discussions, and blockchain talks. Four hundred developers, entrepreneurs, investors, and business executives attended, including firms like UBM, IBM, and Microsoft.
November 2015 – Microsoft Blockchain as a Service
Microsoft announces Ethereum blockchain as a service on Microsoft Azure.
March 2016 – Homestead
The direct result of extensive testing by Frontier developers, the Homestead release enhanced the user-friendliness and security of Ethereum. Changes were made to smart contract creation mechanisms and the block timestamp protocol. Additionally, Homestead provided software to simplify future upgrades. Though core functionality did not differ extensively from Frontier, these minor tweaks to the network allowed for further improvement of security and usability.
June 2016 – DAO Hack
A hacker drains the Decentralized Autonomous Organization (DAO) fund, which was created to fund Ethereum’s development, of approximately $60 million in Ether. Concerns relating to security and design of smart contract functionality are brought to the forefront of the many development projects underway at the time. The community becomes divided about whether to artificially erase the theft by executing a hard fork or to preserve the immutable characteristics of the blockchain by accepting the theft.
September 2016 – DEVcon2
The DEVcon2 event offered another chance for Ethereum’s global following of developers, investors, and other stakeholders to gather and discuss the project’s progress. A virtual accelerator program is announced to continue encouraging innovation on the Ethereum system.
October 2016 – Tangerine Whistle
Despite its smaller size compared to other releases, the Tangerine Whistle upgrade showcased the responsiveness and prowess of Ethereum’s development team. The upgrade was a reaction to a series of September 2016 denial of service (DoS) attacks, which increased the pool of pending transactions and therefore prevented legitimate transactions from being processed, thereby slowing the network. This attack was executed using smart contracts that were cheap to deploy in terms of transaction gas prices, but difficult for nodes to execute and compute. The solution rebalanced certain opcodes that made gas costs more expensive for this type of attack, rendering future DoS attacks of this nature less viable.
November 2016 – Spurious Dragon
While Tangerine Whistle provided immediate relief from the September DoS attacks, Spurious Dragon provided a long-term solution to prevent future attacks of this nature. As part of the upgrade, the Ethereum platform received additional protections against replay attacks, rebalanced more opcodes to adjust gas costs in order to create disincentives for future attacks, and removed empty accounts generated by past DoS attacks.
July 2016 – DAO Hard Fork
The DAO hard fork is executed to return stolen Ether to the accounts that fell victim to the June 2016 DAO hack. This fork did not introduce any protocol changes because its primary purpose was to reverse the illicit transactions.
April 2017 – Ethereum Enterprise Alliance (EEA)
The EEA, an organization connecting “Fortune 500 enterprises, startups, academics, and technology vendors with Ethereum subject matter experts,” was introduced by Microsoft.
October 2017 – Metropolis – Byzantine
The Metropolis release was intended to lay the groundwork for Ethereum’s switch to a Proof-of-Stake consensus mechanism. After realizing that its size would make it prohibitively cumbersome, the Ethereum Foundation instead split Metropolis into two separate releases. The first, known as Byzantine, implemented new functionality that included zk-snarks anonymous transaction processing, modifications to the mining difficulty formula, reduction in the block reward from five to three Ether, and delay of the “Difficulty Bomb” by six months.
November 2017 – DEVcon3
DEVcon3, held in Cancun, Mexico, touched on progress with respect to some of Ethereum’s most anticipated developments, including Plasma and the Casper POS solution. The conference became so popular that the organizers replaced the initial venue with one triple the size to meet increased demand.
November – European EEA
The first EEA is held in mainland Europe, illustrating the global reach of the Ethereum community.
Estimated 2018 – Metropolis – Constantinople
According to the October 2017 developer meeting, the second portion of the Metropolis fork, known as Constantinople, will include a shift from a POW consensus mechanism to a hybrid POS/POW mechanism via Ethereum’s Casper solution. Several core features of Constantinople are still being discussed and do not appear close to release. One such feature, Ethereum Improvement Proposal (EIP) 859, is a particularly important because it would allow for account abstraction on Ethereum’s main chain, making development on Ethereum more intuitive.
To Be Determined – Serenity
The Serenity update will use Casper to transition the network from a hybrid POS/ OW system to a POS only system. Though active development is occurring on this release, the majority of the community’s focus is on the more immediate Constantinople EIPs. Nonetheless, the Serenity code is mostly written and undergoing testing. For example, code repositories for sharding, a scaling solution that is associated with Serenity, currently exist but appear quite far away from implementation.
 “Monero Hard Forks Successfully: Four New Projects The Result” Coinbureau, Coinbureau. 10 April 2018 https://www.coinbureau.com/analysis/monero-hard-forks-successfully-four-new-projects/
 James, Adam. “BITCOIN MINING GIANT BITMAIN RAKED IN $3 TO 4 BILLION IN PROFITS LAST YEAR” Bitcoinist, Bitcoinist. 25 February 2018. http://bitcoinist.com/bitcoin-mining-giant-bitmain-raked-3-4-billion-profits-last-year/
 The difficulty bomb is intended to incentivize miners to switch from the Proof-of-Work to Proof-of-Stake consensus mechanism because it would make Proof-of-Work more computationally expensive.
 “What is gas limit in Ethereum” StackExchange, Bitcoin StackExchange. 7 April 2017. https://bitcoin.stackexchange.com/questions/39132/what-is-gas-limit-in-ethereum
 StopAndDecrypt. “Sharding centralizes Ethereum by selling you Scaling-In disguised as Scaling-Out” Hackernoon, Hackernoon. 7 June 2018. https://hackernoon.com/sharding-centralizes-ethereum-by-selling-you-scaling-in-disguised-as-scaling-out-266c136fc55d
 Stark, Josh “Making Sense of Ethereum’s Layer 2 Scaling Solutions: State Channels, Plasma, and Truebit” Medium, Medium. 12 February 2018. https://medium.com/l4-media/making-sense-of-ethereums-layer-2-scaling-solutions-state-channels-plasma-and-truebit-22cb40dcc2f4
 A Turing-complete system is one that can perform calculations to arrive at conclusions given an algorithm, enough time, and appropriate inputs; this is an important feature that allows for autonomous execution of many Ethereum programs.
 Farmer, Stuart “Proof of Work Kills the Earth” Medium, Lamden. 19 October 2017. https://blog.lamden.io/proof-of-work-kills-the-earth-e687d3e83ec9
 A core tenant of POW systems is that they incentivize many miners to spend long periods of time solving complex mathematical equations in order to earn cryptocurrency. Quantum computing could enable supercomputers to solve these equations almost instantaneously, allowing quantum miners to earn a disproportionate amount of cryptocurrency in an extremely fast manner, centralizing network influence to the quantum miners and giving them authority to modify the blockchain’s history, thereby rendering the blockchain and its cryptocurrency useless.
 Mamoria, Mohit “Is Proof of Stake Really the Solution?” Hackernoon, Hackernoon. 5 July 2017 https://hackernoon.com/is-proof-of-stake-really-the-solution-2db68487f4ba
 Buterin, Vitalik, Griffith, Virgil “Casper the Friendly Finality Gadget” Github, Github. 28 October 2018 https://github.com/ethereum/research/blob/master/papers/casper-basics/casper_basics.pdf
 Demeester, Tuur “Response to Buterin’s Criticism of my Proof-of-Stake Piece” Medium, Medium. 18 July 2017 https://medium.com/@tuurdemeester/re-buterins-criticism-of-my-pos-piece-4ee70d6fd289
 Gerring, Taylor “Cut and try: Building a dream” Ethereum Blog, Ethereum Blog. 9 February 2016 https://blog.ethereum.org/2016/02/09/cut-and-try-building-a-dream/
 Slacknation “Overview: Ethereum’s initial public sale” Medium, Keeping Stock. 15 February 2016 https://keepingstock.net/overview-ethereum-s-initial-public-sale-563c05e95501
 “Enterprise Ethereum Alliance Introduction and Overview” Enterprise Ethereum Alliance, February 2017 https://entethalliance.org/wp-content/themes/ethereum/img/intro-eea.pdf
 The difficulty bomb is intended to incentivize miners to switch from the Proof-of-Work to Proof-of-Stake consensus mechanism because it would make Proof-of-Work more expensive.
 A further explanation of the Casper system is available in the ‘Scaling Ethereum for Global Adoption’ section.