This article is an original Web3 annual review by Coinlive and GeniiData; please cite the source if you intend to reproduce the content.

Introduction

Since the conception of BTC in 2008, countless genius projects have emerged one after another, introducing critical use cases to the web3 industry. However, it is embarrassing that after 14 years of rapid development, there is still insufficient infrastructure in the industry. The scaling of Ethereum through Layer 2s is one of the first few attempts at solving this issue. Through L2s, web3 will truly enter a new era of hundreds of millions of users.

Data source: geniidata.com, from November 2022, the L2 daily transaction volume exceeded ETH main net transactions

The Web3 industry in 2022 is layered with black swan events and continuous turmoil. However, under such pessimistic market conditions, the growth curve of L2 is thriving, making it hard not to pay attention to it.

In this context, the GeniiData team teamed up with Coinlive and invited three senior industry analysts: Peter @peter3050, Tom @tomtsan79919541, Gary @garymetaz, and a community data analyst: 0xFalse @0x_false, to discuss the status quo of L2s. This discussion will be broken down into the following three chapters.

1. Macro level: Overarching Background of L2s

2. Mezzo level: Industry Trends

3. Micro level: Core Projects

1. Contextual Background of L2s

The Web3 industry has never been short of topics, and the scaling of Ethereum will always occupy a place in this list of topics.

(1) Industry demand

In the blockchain trilemma (decentralization, security, scalability), Ethereum chose to sacrifice scalability for security, which made it difficult for Ethereum to achieve more than 15 transactions per second (TPS). During the bull market, users can feel that the growing demand has brought about network congestion and rising gas fees. As shown in the figure below, in Q1 of 2022, the gas fee was once close to 0.01Eth, that is, the cost of each transaction was close to 20 dollars. For industrial verticals that require high-frequency and low-value actions such as socialfi and gamefi, this kind of cost is often unbearable.

Data source: geniidata.com, the Q1 gas fee in 2022 is close to 0.01 ETH, which is more than 20 dollars.

The high gas fee and network congestion has inhibited the development of the ecosystem to some extent, especially for socialfi and gamefi projects building on ETH. These issues were huge bottlenecks, and expansion is imperative.

(2) Expansion route

Ethereum 2.0

Ethereum 2.0 plans to improve the processing capacity and throughput of the network through sharding. In the Ethereum sharding architecture, each shard is a separate blockchain. They share a consensus mechanism, but in operation it is independent. This means that each shard has its own nodes, mining nodes, and miners, each running its own smart contract. The sharding architecture enables the Ethereum network to process transactions more efficiently, while also improving network security.

Ethereum sharding is a promising technology that is currently under development. The only drawback is there is no expected date of completion, it is always a work in progress. After all, changing the engine while the plane is flying is a very challenging job.

With no clear dates on the eventual completion of expansion of ETH, users and developers alike seek expansion through L2s. There are five main mechanisms for L2s: state channel, sub-chain plamsa, side chain, rollups, and state channels.

State channels allow users to make multiple off-chain transactions while only submitting transactions to the Ethereum network twice - once when the channel is opened and once when the channel is closed. In this way, TPS can be increased by 1000, but this mode only supports a fixed number of users and does not support smart contracts.

Sub-chain Plasma is a low-profile fork of Ethereum, which can be extended to multiple layers, and there are grandchildren or great-grandchildren chains on the sub chain. Unlike state channels, Plasma sub-chain can support a flexible number of users. However, the sub-chain does not have its own security mechanism and is highly dependent on the main network. It also does not support smart contracts.

Side Chain is a step further than the plasma sub chain, has its own consensus mechanism and security guarantee, and further optimizes the transaction speed, the most famous of which is Polygon. Sidechains are associated with Ethereum and can operate independently. However, there are still some obvious disadvantages in the side chain. First, the data is isolated from the main chain, and the degree of decentralization is worse than that of Ethereum.

Rollups. The core concept of rollup is to decouple consensus from execution, execute transactions in L2, aggregate many transactions to generate a large batch, publish the transaction results in L1, and verify the validity of the transaction results through the proof mechanism.

According to different proof mechanisms, it can be divided into optimistic rollups and zk rollups. The specific differences will be introduced in detail later.

Validium also adopts the zk rollup proof method, but unlike rollups, which store data on chain, Validium's data is managed by a group of data notaries stored off the chain, so its security is poor.

After the simple comparison above, most of the expansion solutions have obvious shortcomings. Although rollup is not perfect at present, it is the most reliable expansion method at present, so this article will focus on rollup to discuss the road to L2 expansion.

1. Optimistic rollup

Optimistic rollup operators package off-chain (outside of Ethereum) transactions into batches and send them to Ethereum for notarization. This compresses and reduces the amount of data passed through Ethereum, increasing efficiency. Optimistic rollup is considered "optimistic" because it assumes that off-chain transactions are valid, relying on the proveFraud() function to verify the correctness of the transaction.

To verify correctness, after submitting to Ethereum L1, a challenge period (usually one week) is set. If there is no challenge during this period, then this batch of data is classified as the final state and correct.

If the period is successfully challenged and it turns out to be an error state, it will be deleted and replaced with the correct result after recalculation. The packager who submitted this wrong transaction will be punished, and the successful challenger will be rewarded.

Source：ethereum.org

There are two star-level products built using optimistic rollup: Optimism and Arbitrum. Although they are based on the same rollup technology, there are some differences between the two, mainly in the processing of fraud proofs, proveFraud() function.

Proof of Fraud

Optimism adopts the interactive mode of single-round fraud proof. In case of a challenge, the entire transaction process is recalculated on L1 to find out the problematic steps. The whole method only needs one round of interaction. However, this method has an obvious disadvantage. The entire transaction is recalculated on ETH L1, and the resulting cost from gas fees is relatively high.

Correspondingly, Arbitrum adopts an interactive mode of multiple rounds of fraud proofs. Through the mechanism of dichotomy on chain, it first locates potentially problematic steps and then verifies the correctness of the entire sequence on L1. The entire dichotomous positioning mechanism has multiple rounds of interaction. Hence from a conceptual point of view, the model adopted by Arbitrum is more complicated and takes longer (1-2 weeks). But because it can accurately locate the disputed step eventually, it reduces the cost of on-chain calculations.

EVM compatibility

For users, optimistic rollups have greatly improved performance and can save a lot of costs. For developers, the EVM compatibility feature of optimistic rollups saves them a lot of work. Developers can use a lot of Ethereum's infrastructure and the learning cost is low. At the same time, the existing smart contracts on Ethereum can be directly migrated to L2 without extensive code modification, and the migration cost is also very low.

2. zk rollups

zk rollup uses zero-knowledge proof (zk) to verify the packaged transaction results. The core logic is to provide a zkp (zero knowledge proof) based on zero-knowledge proof technology when submitting data to L1. Compared to the fraud proof of Optimism rollup, zk rollup has higher privacy, security, scalability and capital efficiency:

1) Privacy: zk technology allows one party to prove to the other party that something is true without providing any information, so the user's privacy will be greatly protected.

2) Security: There are no observers in the zk rollup mechanism. By adopting the verification method of smart contracts, the security of cryptographic technology will improve.

3) Scalability and cost: For most applications, relatively less data needs to be submitted to the main chain through zk rollup, so this technology increases TPS and lowers costs.

4) Capital efficiency: Since the validity proof mechanism of zero-knowledge proof is adopted, there is no need to set a longer challenge period like Optimism and Arbirtrum rollups, so the withdrawal speed of funds will be much faster.

Source：docs.zksync.io

Due to the above obvious advantages, even Vitalik agreed that zk rollup is the ultimate path of rollup. However, the difficulty of zero-knowledge proof technology is very high, so at the current stage, zk rollups also have many shortcomings:

1) The lack of compatible EVMs. There is no practical zkEVM that can achieve consensus level (the highest level) compatibility, and Ethereum core developers expect that it will take several years of hard work. At present, the zksync project is relatively fast. zksync 2.0 already has zkevm with language level compatibility (minimum level compatibility), not yet consensus level.

2) Difficulty of development is high. The technical complexity of zero-knowledge proof is high, so the development of projects building with zk rollup is difficult. Understandably, the overall progress is behind projects using optimistic rollup. Also, zk rollup projects are harder to audit, thus the probability of serious risks and exploitable loopholes is higher.

3) Expensive hardware requirements. Since zkp needs to be generated every time during verification, projects building on zk rollups need better hardware equipment and even so, still run the risks of overheating.

In an attempt to solve the L1 trilemma through the introduction of rollup technologies, it seems that rollups have also fallen into a paradox that cannot be balanced.

Zk rollup, which is obviously superior in security and scalability is difficult to develop and is not compatible with EVM, which leads to the delay in launching mature products.

Optimistic rollup is easier to develop and is compatible with EVMs, so mature products can be deployed on it easily, capturing most of traffic on Ethereum L1 as the first mover in the rollup industry. However, in terms of performance, security and user experience, it is evident that it cannot be compared with the final form of zk rollup.

Considering the differences in the operating capabilities of many participating players and developers, it is difficult to draw a conclusion on which roll up will still be ultimately preferred by the market.

3. Prospects and Competition

The birth and development of an industry cannot be separated from demand and competition, with demand being the start of all stories, and competition deciding the ending of them.

1. Demand and Prospects

The need at the start of this story is derived from everybody's longing for high TPS (transactions per second) and low fees. If Defi stands for high-priced low-frequency chain activity, then it can flourish with performance support from L1. But with the introduction of L2s, with their high performance and low fees, will there be more projects that will experience explosive growth?

DEFI: In addition to the migration of old projects to L2, with the support of high TPS, it will also bring more innovation opportunities for the Defi track. Order book based DEXs such as DYDX are gaining more and more attention and potentially L2s can increase the use of DEXs amongst users as well. The order book DEX is naturally suitable for various high-frequency trading, and its market share in traditional finance can reach 30%~50%. However, in the past, due to high gas fees, such high volume transactions can only be performed on centralized exchanges. In the future, with L2 expansion, these customers can perfectly complete transactions on a safer decentralized exchange.

Gamefi & NFT: After the expansion, L2s will also support high-performance games, allowing for the development of more truly fun games, instead of solely relying on P2E games to attract users. Many classic games in Web2 can also be combined with NFTs to realize meaningful adoption and circulation of game assets. Layer2-based NFT gamefi such as Immutable X and ZKBox have received a lot of attention. Maybe Gamefi 2.0 is really not far away from us.

Socialfi & DID: Socialfi is the track with the most users and the highest frequency of use. If L2s can successfully integrate with socialfi aspects, it will bring huge traffic to web3. From the perspective of past history, a new batch of social platforms emerge every ten years, because young people are always looking for a more interesting social world that can break the shackles of the previous generation and establish their own voice.

At the same time, socialfi can capitalize on their users behaviors, contributions and social activities. These massive data assets will also enrich DID, so that web3 users can have a vivid virtual identity closely tied to their real lives, instead of a cold and distant one.

Every technological advancement may bring disruptive innovations that did not exist before, potentially creating huge new demand. Just like the birth of NFTs, what disruptive new technologies will L2 bring to the industry?

2. Competition and Ending

The industry demand for L2 is potentially huge as discussed above. This has given potential projects a good starting platform for building, but the competition determines the end game. If you cannot survive the fierce competition, then no matter how big the demand is, you cannot thrive.

Technology advancement is a big influencing factor when measuring competition, so rollups occupy a leading position in the L2 expansion competition compared to others like side chain or sub-chain plasma. This is undoubtedly good news for project parties adopting rollups. Also, within rollup, zk rollup is recognized as the final solution, but this does not mean that the current projects using this technology (zksync, starkware, etc.) are the final winners.

Here we want to introduce the competitive evolution of the intelligent driving industry and use this as a reference to think about the future pattern of L2.

Tesla vs Google

From the level of intelligence, the intelligent driving industry also has a classification from low to high L1-L5. Google believes that low-level unmanned driving is an immature product. They are aiming at fully unmanned L4+ from the beginning. Tesla's route is to launch low-level assisted driving L2 products first and use a large amount of data to push the products to mature L4 or L5.

Judging from the current situation, Tesla's route undoubtedly has a higher probability of success, because the advancement of technology itself relies on a large amount of data and market feedback, and purely limited road tests can never achieve real L4. As a result, Tesla was able to surpass google as the leading intelligent driving company.

The above example might help us think about the final outcome for L2s. Absolute technology is not the only factor that determines the success or failure of a project.

If we use high execution power, we can launch products that fit the current market as soon as possible, capture market traffic, and continue to improve in future operations. If a product delays launching till its final perfect form, it may have a lower probability of success.

Moreover, from a technical point of view, the combination of low-difficulty technologies can sometimes achieve the performance of high-difficulty individual technologies. Arbitrum and Optimism are major adopters of optimistic rollup. Although the performance of optimistic rollup lags behind the zk rollup route, some disadvantages can be solved by combining technologies. To overcome, for example, the lengthy withdrawal duration, it can be solved through a third-party cross-chain bridge.

L2 cross-chain bridge Orbiter can help Arbitrum to reflect the minute-level account transfer

2. Industry trends

There are some differences in the current definition of L2. In the broadest sense, all technical expansions outside the Ethereum main chain are included in the category of L2.

Here we will focus on a specific definition: L2 is the one that relies on the security of Ethereum L1. According to this, state channels, Plasma, and side chains are not strictly L2 since they have their independent security in place. Therefore, the subsequent discussion on the L2 industry will mainly focus on rollups only and their surrounding ecosystem.

(1) L2 Development Trend (Qualitative Analysis)

1. The development space for L2s is huge, potentially leading to the next explosion of the industry.

2. Multiple L2 solutions can coexist

For L2 infrastructure, Artibturm, Optimism, zksync, starkware. In terms of security, low rates, scalability, project availability schedule, there is no all-round leading L2 infrastructure.

The project party can only choose the corresponding infrastructure according to its actual needs, so there can be multiple solutions in the short term. Coexistence is a more realistic result than competition.

But what deserves attention is the operational and directional capabilities of each project party. Whoever can utilize great marketing and networking effects in web3 to secure larger market share will be the winner. With the help of first-mover advantage, projects can build a larger ecosystem, grasp more resources and take the lead in the competition.

3. Cross-chain is a necessity

The rise of L2 means that there must be a multi-chain ecosystem in the future. According to Vitalik, there will be L3 or even L4 in the future. From this point of view, cross-chain will also be an essential infrastructure.

The difference in technologies between different L2s highlights the issue of poor interoperability between L2s. Since these L2s are built with different smart contracts, there will be more inconvenience for users to interact with data and assets across the L2s. Hence, if the same application is deployed in different L2, it will also affect its fluidity. That is why in addition to the development of L2 itself, cross-L2 is also an important development direction.