Author: Binance Research; Translator: Blockingcryptonaitive
1. Main points
The new trend in blockchain scalability is the evolution towards Layer3 networks, superchains, and hyperchains. These developments are being adopted by several Layer2s and are expected to become the basis for the next generation of Ethereum scalability, simplifying the development process, enhancing security, and promoting greater ecosystem interoperability.
The OP Stack of Optimism is a pioneer in moving towards highly interoperable blockchain networks such as Superchains, and the recent Bedrock upgrade represents one of the first steps towards realizing their vision.
Arbitrum is also at the forefront of Ethereum scalability, developing Layer3 networks through Arbitrum Orbit, providing a permissionless framework for deploying custom chains on top of Arbitrum L2.
zkSync introduces the concept of hyperchains, a customizable and trustless chain-based blockchain network, achieving ultra-high scalability, improved composability, and enhanced security.
StarkWare, leading the innovation of fractal scalability, is developing a multi-layer solution, exploring L3 for custom scalability and leveraging L2 for universal scalability.
Polygon 2.0 aims to unify its set of L2 solutions, including Polygon PoS, Supernets, and zkEVM, through cross-chain coordination protocols, to create the “value layer of the Internet.”
Over the past few years, Ethereum co-founder Vitalik Buterin’s rollup-centric vision of the future has gradually been realized. In fact, Layer2 rollup technology has laid the foundation for solving the Layer1 scalability problem. Successful projects such as Optimism and Arbitrum have effectively utilized optimistic rollup technology to provide reliable services to users for nearly two years. Meanwhile, zero-knowledge (zk) rollup technology, represented by Starknet, zkSync Era, and Polygon zkEVM, has also gained recognition in recent months after successfully launching their mainnets. With a considerable level of market validation achieved, the focus of L2s is now shifting towards the new theme of blockchain scalability to accommodate the potentially hundreds, thousands, or even millions of rollups that may arise in the future. With this shift, we may finally realize blockchain scalability, attracting a large number of developers who can work together to attract millions (or even billions) of users to cryptocurrency.
As the L2 market moves into its next phase, rollups are beginning to define a vision of dominance in the rollup-led future. In this report, we explore the rapidly developing field of blockchain scalability. Our focus is on dissecting the complexity of different L2 approaches, evaluating their current position in the market, and delving into the new themes that are defining the field.
Challenges facing the current scalability paradigm
Ethereum, as an L1 network, must be able to scale to the same size as the internet, but currently no single L2 chain can achieve this goal. In fact, as expanding applications and ecosystems look for blockchains to deploy on, their choices are very limited and far from ideal. They can choose to deploy on Ethereum, but sacrifice scalability; or they can choose to use an L2 solution, but there is a risk of dependence on that ecosystem; or they can decide to deploy and maintain their own chain, leading to fragmentation of liquidity and giving up the potential of network effects. In addition, interoperability has long been a key challenge, and our solution to this challenge, cross-chain bridges, is fragile and often a target of malicious attackers.
What are the new themes?
The vision for the next stage of L2 evolution is to simplify the process of deploying various types and modes of rollups for developers, whether application-specific or general-purpose. Leaders in the current L2 space seek to achieve this new stage by reducing technical complexity, increasing resource availability, and providing a platform for developers to share existing overhead.
In our current system, each rollup has its own infrastructure overhead (bridges, sorters, node providers), common standards, and design structures. This is unsustainable, leading to continued fragmentation and resource waste. Many teams do not want to consider running infrastructure, they only want to focus on building applications. The ultimate goal is that deploying new rollup chains should be more or less like creating new web pages; expect to achieve something as close to “one-click deployment” as possible.
These second-generation rollup chains, whether the OP chain on Optimism, the L3 chain on Arbitrum Orbit, or the hyperchains on zkSync, will allow different entities with different intentions to build their own custom environments, while reusing the existing and battle-tested infrastructure of the parent L2. By doing so, they hope to create a seamless, secure, and highly interoperable environment that truly achieves the next level of scalability.
4. Key Players
Optimism and OP Stack
Optimism is the collective behind the OP Mainnet, an Ethereum Virtual Machine (EVM) equivalent solution based on Optimistic Rollup that began running in December 2021 and is one of the leading Ethereum L2 solutions. As of the writing of this report, the total locked value (TVL) of the OP Mainnet is over $2.2 billion, making it the second largest market capitalization in all Ethereum L2 solutions with a market share of over 23%. In October 2022, Optimism launched OP Stack, which they describe as a “highly scalable, highly interoperable modular, open-source blueprint for various blockchains.” This marks them going beyond just running an Optimistic Rollup solution in design and vision. It also introduces us to the concept of a so-called Superchain, which refers to a highly integrated and unified L2 blockchain combination built on OP Stack. The latest development in the implementation of this new phase by Optimism is the migration of their flagship L2 Rollup to Bedrock, the first formal version of OP Stack, which brings many operational and user improvements to their product.
Figure 1: A brief timeline of Optimism to date.
What is OP Stack?
OP Stack is a standardized, shared, and open-source development stack (i.e., a set of software) that drives OP Mainnet. It consists of various software components (i.e., modules or code libraries) that together make up Optimism’s L2 Rollup and can be used to create a set of shared, interoperable, and coordinated L2 blockchain networks. It is worth noting that OP Stack is a constantly evolving concept that grows with Optimism itself. Essentially, OP Stack aims to simplify the process of creating an L2 blockchain and can be thought of as a “supermarket” for building an L2.
Figure 2: The different conceptual layers of OP Stack.
Data Availability (DA) Layer: This layer defines where the raw inputs based on the OP Stack chain are published. The most commonly used DA module in OP Stack is the Ethereum DA layer, which is used by the OP Mainnet Rollup. It is worth noting that we have recently seen the release of the Taro testnet, which is the first OP Stack Rollup chain to use Celestia as its DA layer instead of Ethereum.
Sequencer Layer: This layer determines how transactions on the OP Stack chain are collected and published to the DA layer. In the current state, the sequencer module of OP Stack is a single sequencer setup. Future proposal modifications will include multiple sequencers to enhance the decentralization of the platform. Economically, sequencers generate protocol revenue through transaction fees paid by OP Mainnet users, which is primarily used for the traceability of public goods funding. In the future, as sequencers become more decentralized, there are also plans to distribute a portion of the revenue to token OP holders.
Derivation Layer: This layer defines how to process the raw data of the DA layer to form inputs sent to the execution layer. It is closely related to the DA layer because it must understand how to parse the raw data of the DA layer.
Execution Layer: This layer defines the state in the OP Stack chain and how it changes after receiving inputs from the derivation layer. The current execution layer module in OP Stack uses a slightly modified version of the EVM but can be modified to include the Move virtual machine, among others.
Settlement Layer: Traditionally, this layer processes the withdrawal of assets by first proving the state of the target blockchain and then processing the operation of extracting assets from the blockchain. From a broader perspective of OP Stack, the settlement layer allows third-party blockchains to understand and establish a view of the state of the OP Stack chain. OP Stack currently includes modules around proof-of-stake proof-of-fault tolerance to establish this view. However, Optimism is also researching their minimal-trust Cannon proof system and added ZK validity proofs to the recent Bedrock upgrade.
Governance Layer: This refers to general tools and processes for managing upgrades, design decisions, and system configurations. Relative to other layers, this is a more abstract layer and can contain various mechanisms. The two modules displayed in OP Stack are multisignature contracts and governance tokens. A multisignature contract is a smart contract that executes a given operation once a predefined signature threshold is reached by the participant set. This is currently the mechanism used to manage the OP Mainnet bridging contract upgrades. Governance tokens are typically used for decentralized decision-making, and their exact function varies depending on the situation.
It is important to remember that builders can easily modify existing modules or create new modules to fit the needs of any application they are focused on. For example, as mentioned in the DA layer section, Taro is a Rollup that uses Celestia instead of Ethereum as its DA. The OP Stack essentially deconstructs the various components needed to build an L2 chain and packages them as separate modules. Builders can then combine the most appropriate modules to create their own L2. Ultimately, Optimism foresees a proliferation of highly compatible L2 and L3 chains that they refer to as OP Chains. They believe that these chains will eventually lead to a super-chain. The first step towards realizing this future is their recent Bedrock upgrade.
Advantages of the OP Stack
Optimism has always focused on aligning with Ethereum, so the OP Stack is an Ethereum Virtual Machine (EVM)-equivalent Rollup development toolkit. This brings some key advantages that are crucial to helping them achieve their potential super-chain future.
Scalability: The OP Stack code is designed with the goal of other builders wanting to use and build on top of it. As such, their code is open source and often modular so that other developers can easily build on top of it. This means that Ethereum Improvement Proposals (EIPs) and future upgrades should be easy to implement.
Simplicity: Optimism’s philosophy is that complex code does not scale. They strive to make the code as simple as possible and tend to reuse existing battle-tested code. That’s why they chose to use Geth fork as the default execution engine for OP Stack (tested on Ethereum for many years). Generally speaking, the simpler the code, the less vulnerable it is to potential errors and attacks.
Familiarity: Existing Ethereum developers should find it relatively easy to build on top of OP Stack because it aligns with Ethereum and its code.
Client Diversity: OP Stack can have multiple client implementations. This diversity brings benefits in terms of vitality and security. We have already seen the mainnet release of OP-Erigon, another execution client for OP Stack, as well as the announcement of Magi, an alternative OP Stack Rollup client.
Bedrock Upgrade was released on June 6, 2023 and is the first official version of the OP Stack, representing a complete modular rewrite of the core components of the OP Mainnet Rollup architecture. Optimism’s flagship Rollup now uses modular OP Stack construction. Bedrock not only represents an upgrade to the OP Mainnet Rollup, but also provides the tools needed to launch a production-quality optimistic L2 Rollup blockchain.
Lower Cost: Bedrock implements an optimized data compression strategy to minimize data costs. This significantly reduces transaction costs, with the average cost per transaction now reduced by over 77%. This also makes Optimism’s L2 Rollup one of the cheapest Ethereum L2 Rollup solutions.
Figure 3: After the Bedrock upgrade, the average gas fee for the OP mainnet has dropped by about 77%.
Figure 4: The OP mainnet is now the cheapest option for token exchange on the Ethereum L2 Rollup.
Shorter Deposit Time: Bedrock introduces support for L1 restructuring in node software, reducing deposit wait times. Deposit times have been reduced by about 70% from about 10 minutes to about 3 minutes. This is an important improvement for users who are new to or less experienced with L2 Rollup, and improves the user experience.
Improved Proof Modularity: Bedrock abstracts the proof system from the OP Stack (the settlement layer in Figure 2), allowing the OP Stack chain to use error-correcting proofs or validity proofs (i.e. ZK-proofs) for transaction verification. It is worth noting that there is already an RFP (Request for Proposal) to build an OP Stack ZK-validity proof program.
Improved Node Performance: Bedrock makes it possible to execute multiple transactions in a Rollup “block” instead of the previous “one block, one transaction” model. At current transaction volumes, this will reduce state growth by about 15GB/year.
As we discussed earlier, OP Stack helps break down the various components needed to build an L2 chain, and Bedrock is the first implementation of this software. The key here is modularity. Builders can now use standardized modules similar to Bedrock to build their own L2 chains or create custom L2 chains by recombining various components. Overall, the Bedrock upgrade not only brings many important improvements to Optimism’s L2 Rollup, but also brings Optimism one step closer to its Superchain vision.
Following the Bedrock upgrade, Optimism’s next step is to upgrade itself to a superchain. The superchain is envisioned as a decentralized L2 chain network (called the OP chain) where these chains share security, communication layers, and open-source technology stacks (OP Stack). These chains will be standardized and intended to be used as interchangeable resources. This standardization will enable developers to create applications for the entire superchain, not just the underlying chain that runs the application. It should be noted that the superchain is currently just a concept and is still evolving. In fact, the Optimism team sees this as a “multi-year (even decade-long) journey.”
Figure 5: Visual Representation of the Superchain
What Will the Superchain Bring?
Strengthened and Secure Codebase: As the chain grows, each chain shares and contributes to a modularized and standardized codebase that strengthens the system. More iterations and development on the same standardized code will help simplify it, making it more attack and error-resistant. Given the prevalence of hacker attacks and security vulnerabilities in the system, this will be a key advantage of structures like the superchain.
Cross-Chain Atomic Composability: This refers to seamless transactions between different OP chains without the need for bridging or intermediaries. While the superchain is made up of multiple chains, end users can have the experience of using a single, unified chain. In practice, this can bring improvements such as a universal block explorer (instead of each chain using different explorers) and no network switching when using applications (e.g., no dropdown menu for MetaMask).
Shared Ethereum Infrastructure: This makes the work of developers easier and enables existing Ethereum developers to seamlessly switch to building on OP chains.
Chains that replace standardized modules with experimental alternatives are called “Hacks” in the OP world. These chains, which do not fully conform to the OP Stack, may have security vulnerabilities and cannot join the superchain. However, they provide a useful platform for developers to experiment and create novel applications. OPCraft is one such example, running a modified EVM on the execution layer to create a whole-chain 3D voxel-based gaming platform. Optimistic Game Boy is another example, where developers used a Game Boy emulator in the execution engine.
Optimism’s L2 Rollup chain became the first member of the superchain after the Bedrock upgrade. Coinbase’s upcoming Base L2 will become the second member, expected to be officially launched this year. Worldcoin has also promised to build on the OP Stack. Conduit is another interesting project that aims to help developers launch their own OP Stack Rollup more easily and eventually become part of the superchain. Aevo is a decentralized options exchange that recently partnered with Conduit to launch a Rollup based on the OP Stack, with Conduit running the Aevo Rollup Sequencer. Interestingly, BNB Chain has also announced the testnet for opBNB, an OP Stack-based EVM-compatible L2 chain. In the NFT space, the decentralized NFT marketplace Zora recently launched the Zora Network. The Zora Network is an L2 chain based on the OP Stack, designed to make NFTs cheaper and more accessible. Their documentation states that the cost of minting on Zora can be less than $0.50 and transactions can be confirmed in seconds.
As we have seen, development around the OP Stack is unfolding widely in the crypto ecosystem. How their solution will perform compared to competitors Arbitrum, zkSync and Polygon, and the progress over the next few months, will be interesting.
Arbitrum Orbit is another popular EVM-equivalent optimistic Rollup in the L2 race developed by Offchain Labs, which launched on the mainnet in August 2021. In fact, Arbitrum is currently the largest and most dominant L2 network in terms of total value locked (TVL), with over $5.9 billion in value accumulated. Against this backdrop, the Arbitrum ecosystem includes several products, including Arbitrum One, Arbitrum Nova, Arbitrum Nitro, and the upcoming Arbitrum Orbit.
- Arbitrum One: The first core mainnet Rollup of the Arbitrum ecosystem. It is fully trustless and inherits Ethereum L1’s security guarantees without introducing additional trust assumptions.
- Arbitrum Nova: An AnyTrust solution, is the second mainnet Rollup for Arbitrum, suitable for cost-sensitive, high-volume projects.
- Arbitrum Nitro: The technical software stack that drives the Arbitrum L2 to make Rollup faster, cheaper, and more compatible with EVM. Nitro introduces interactive proofs that run on WebAssembly (“WASM”) code used in Arbitrum.
- Arbitrum Orbit: The development framework for creating and deploying L3 chains on the Arbitrum mainnet.
Figure 6: A brief timeline of Arbitrum so far.
What is Arbitrum’s vision?
Similar to Optimism, Arbitrum has laid out a development path that moves from relatively uniform approaches to more modular and component-based frameworks. The core strategy of Arbitrum is to attract a variety of applications to its general-purpose chains (Arbitrum One and Arbitrum Nova), and an important component related to this strategic vision is L3 Rollup, which is used to settle transactions onto the chain. L3 is the next stage of Arbitrum’s scaling journey and the solution to achieving its vision, which Arbitrum calls Arbitrum Orbit.
What is Layer 3?
L3 networks are sometimes referred to as application chains, although their scope extends beyond that term, as they are specialized networks built on top of L2 that host smart contracts supporting specific decentralized applications. They leverage the security of the L2 network, which in turn borrows security assurances from the underlying L1 network (such as Ethereum). Ultimately, L2 is used for general scaling, while L3 is envisioned as being tailored to specific applications.
The easiest way to understand L3 is to think of it as a Rollup on top of L2. L3 settles not on L1 but on L2. More importantly, they offer significant increases in scalability, making them an ideal solution for high-throughput applications, particularly in certain use cases in DeFi or crypto gaming. One notable example is dYdX, one of the earliest dApps in the DeFi space to move from L1 to its own application chain on L2 to better scale its product.
For example, if L3 increases throughput by a factor of 10 relative to L2 and L2 increases throughput by a factor of 10 relative to L1, then L3 increases throughput relative to L1 by a factor of 100. More importantly, L3 networks improve interoperability and communication between protocols by connecting different aspects of the blockchain and Web3 ecosystem.
Figure 7: Arbitrum Orbit is an example of an L3 solution.
What is the purpose of Arbitrum Orbit?
Set to launch in March 2023, Arbitrum Orbit provides a new permissionless frontier for developers to build chains on top of Arbitrum One or Arbitrum Nova, known as L3 chains. In short, Orbit is a permissionless development framework that allows anyone to deploy L3 chains on top of the Arbitrum L2 chain. The permissionless aspect is a key feature because it enables developers to create new L3 chains on the Arbitrum L2 chain without requiring permission or formal approval.
Arbitrum sees Orbit as a highly accessible product and an important component of Arbitrum’s public infrastructure. Offchain Labs CEO Steven Goldfeder said the team is committed to further developing and improving the solution, aiming to position Orbit as the simplest and most convenient platform for dApps to introduce L3. This is undoubtedly beneficial for application teams, giving them the opportunity to test and experience L3 before committing to their deployment.
Through Orbit, Arbitrum aims to support the following protocol-oriented use cases that want to launch their own L3 chains:
L3 Rollup: Launch an L3 Rollup chain similar to Arbitrum One.
L3 AnyTrust: Launch an L3 AnyTrust chain similar to Arbitrum Nova.
Customizable L3: Deploy a customized L3 chain built on Arbitrum Nitro to meet specific application requirements. Components include privacy, permissions, fee tokens, governance, and more.
Through this solution, Arbitrum aims to attract developers who want more control and seek customization, allowing them to fork and freely adjust Arbitrum’s source code according to their specific needs. Arbitrum values customization very much and even calls their solution “tailored chains”. Orbit will be compatible with the upcoming Arbitrum Stylus upgrade, providing developers with the possibility of building decentralized applications using standard programming languages such as C, C++, and Rust, thereby expanding Arbitrum’s diversity and influence. In addition, transaction fees on these L3 Rollups will be paid in ETH to Arbitrum’s sequencer.
Orbit will transform Arbitrum into a settlement layer similar to Ethereum, adding more value to its core chain and enhancing the long-term scalability of the entire Arbitrum ecosystem. Although the Arbitrum team acknowledges that the optimal L3 structure has not yet been found, this upgrade will allow dApps to enjoy the benefits of increased throughput and secure bridging provided by L3, as well as the network effects already existing in the Arbitrum ecosystem. Projects have already shown interest in this development, with decentralized gaming network Xai becoming the first representative of L3 on the Arbitrum platform.
zkSync and Hyperscaling
zkSync is an Ethereum L2 scaling solution developed by Matter Labs that leverages zk-rollup technology. The second version of zkSync, known as zkSync Era, was launched earlier this year and is Matter Labs’ zkEVM Rollup version. Since its deployment on the mainnet, zkSync Era has become one of the most widely used L2 and zkEVM solutions, with a total locked value (TVL) of over $625 million.
For more information on the zkSync Era, be sure to check out our previous report, “zkEVM World: Overview of zkSync.”
The zkSync Era is the second iteration of the zkSync product suite, and Hyperscaling embodies its ultimate goal: to process an unlimited number of transactions without compromising security or decentralization. To meet the growing demands of Web3, zkSync has designed a final solution centered around the concept of hyperchains. The zkSync architecture aims to connect a set of hyperchain networks with a central basechain.
What are Hyperchains?
Hyperchains are zkSync’s vision for L3, conceived as a trustless and customizable chain-based blockchain ecosystem. They are parallel-running, fractal-like zkEVM instances that can be created or deployed by any entity in a permissionless way. The implementation of zkSync Hyperchains takes a modular approach, offering developers a Hyperchains software development kit (SDK) framework that allows them to choose from various components to build their own blockchains or develop their own components.
Figure 8: Hyperchains are sovereign ZK chains on Ethereum, connected via customizable and fully trustless Hyperbridges.
The zkEVM engine plays a crucial role in maintaining network uniformity. In order to maintain trust and interoperability in this ecosystem, each Hyperchain must be supported by the zkEVM engine, the same engine that supports the main zkSync L2 platform. This allows Hyperchains to inherit their security directly, regardless of who deploys them.
What is the Basechain?
The Basechain is similar to zkSync Era. It can be seen as the main Hyperchain instance of zkSync Era, with the ability to settle its blocks directly on Ethereum L1. It not only serves as the main computing layer for general smart contracts, but also as the settlement layer for all other Hyperchains, including L3 and higher.
The zkSync Hyperchain aims to provide infinite scalability, covering the following aspects:
Security: Hyperchains are able to overcome typical vulnerabilities associated with non-native bridging, which often leads to hacker attacks. The interaction between fractal Hyperchains in zkSync L3 is done through native bridging, further enhancing security.
Performance: By implementing a Hyperchains architecture and achieving hyperscaling, the performance of L3 is improved.
Cost: Data costs are significantly reduced due to data availability solutions.
Ease of use: zkSync is expected to greatly improve the SDK, including introducing low-code and no-code solutions, making application development easier.
Composability: The system’s LLVM compiler supports Solidity and other modern programming languages, improving accessibility for developers using languages such as Rust, C++, and Swift.
Furthermore, the privacy, token economics, and data availability features of the superchain itself can be customized according to specific needs. zkSync’s permissionless solution provides developers with three different data availability options, allowing developers to choose between price, performance, and security trade-offs.
ZK Rollups: Suitable for applications that require Ethereum’s complete security functionality, primarily used for DeFi applications.
zkPorter: This option is suitable for applications that want to balance on-chain and off-chain data, and pursue economy, speed, and security, making it ideal for game developers.
Validium: Suitable for applications that seek slightly lower security than Ethereum but have the best performance.
We all know that scalability refers to the ability to handle a large number of transactions without compromising security or decentralization. While current scalability frameworks are relatively sufficient, they do not take into account the ever-growing needs of Web3, which will eventually be infinite. This is where the concept of hyperscaling comes into play. Hyperscaling refers to the ability to handle an infinite number of transactions without compromising security or decentralization. Remember the blockchain trilemma? That is, the network can only choose two out of scalability, security, and decentralization. Hyperscaling aims to eliminate this trade-off.
A hyperscaling blockchain system involves multiple different ZK chains (or superchains) running in parallel, with block proofs being aggregated and settled on L1. Theoretically, this could even be an infinite number of superchains, representing the entire system. The entire process is related to an idea called fractal scaling, which was first introduced by StarkWare. Fractal scaling is based on the concept of a multi-layered network, where applications specific to L3 are recursively built on top of L2.
However, hyperscaling takes it one step further by introducing super bridges, native bridges that connect each L3 application chain to each other, making transfers between superchains not consume resources on third-party chains, and further ensuring that the underlying chain does not become a centralized, scalable bottleneck, thus maintaining the principle of parallel hyperscaling. Figures 9 and 10 respectively show the comparison between fractal scaling and the zkSync hyper-scaling vision.
Figure 9: Without super bridges, fractal scaling may cause the basic chain to become the main intersection of most transfers over time, which may hinder scalability.
Figure 10: With super bridges, transfers between super chains become as simple and efficient as regular transfers, just like hyperlinks allow seamless navigation from one webpage to another with just one click, without extra navigation through each layer.
StarkWare and Fractal Scaling
StarkWare, the company behind the universal Ethereum L2 ZK Rollup project Starknet, can be considered as the first major organization to start discussing fractal scaling and multi-layer architecture. In their blog post “Fractal Scaling: From L2 to L3” from December 2021, StarkWare published their views on L3, including use cases and potential benefits.
In short, StarkWare’s multi-layer network argument argues that while L2 is a general-purpose extension, L3 should be used for custom extensions. This argument is not dissimilar to what we have previously discussed in some L2 projects. In fact, StarkWare’s initial concept likely inspired other L2 projects to start exploring the next stage of Ethereum scaling evolution.
Figure 11: A visual representation of StarkWare’s initial multi-layer architecture framework proposed.
Although StarkWare has not made any further announcements on the above argument since its initial publication, we still see some developments happening in their ecosystem. Slush further advances the concept of fractal scaling and is committed to building an SDK for zkVM L3 on top of Starknet. While the specifics of this project are very technical and beyond the scope of this report, it is worth keeping a close eye on Slush’s progress on their roadmap and who they can attract into their zkVM L3 world.
Additionally, in August of last year, StarkWare announced the launch of recursive STARKs, an innovative solution that provides new scalability for blockchain systems. These STARKs can bundle multiple transaction proofs into a single proof, enhancing the ability of L2 extensions. The technology relies on Cairo, a proof-generating programming language that supports complex computational tasks, and SHARP, which can aggregate transactions from multiple applications into a single STARK proof. This development not only greatly increases the number of transactions that can be written to Ethereum with a single proof, but also opens up new horizons for super-scaling and advanced applications.
The implementation of recursive STARKs is aimed at optimizing costs, delays, and computing resources to facilitate the development of L3 solutions on the public Starknet network. As StarkWare continues to refine the recursive verifier, further improvements in performance and cost-effectiveness can be expected. Undoubtedly, recursive STARKs represent an important step forward for Starknet in the L3 network space.
Based on these noteworthy developments, it will be interesting to observe the ongoing progress of Starknet and their regularly updated strategic approaches to advancing Ethereum’s scalability.
Polygon 2.0 is the latest announcement from Polygon regarding its L2 development, released in early June. The vision of Polygon 2.0 is to create the “internet of value layer”.
Before delving deeper, let’s first understand some background on Polygon. Polygon is a platform dedicated to enhancing Ethereum scalability, achieved through various solutions. Their flagship product is the Polygon PoS sidechain, which is EVM-compatible. Polygon’s L2 solution processes approximately 2-3 million transactions per day from 300-400k active addresses. Polygon has also ventured into the application chain space, launching their own solution called Supernets, which enables developers to build customizable application chains. Several companies, including gaming giant Nexon, have announced adoption of Supernets, including Nexon’s new game in MapleStory Universe. Additionally, Polygon’s zkEVM is their equivalent ZK-rollup solution, which went live on the mainnet in late March last year and currently has over 177k unique addresses and processes 200-500k transactions per day.
Polygon 2.0 is the latest member of their L2 product suite, aimed at unifying these platforms and creating a seamless interface for users. Conceptually, Polygon 2.0 is a group of L2 chains driven by ZK technology, utilizing a unique cross-chain coordination protocol. The solution aims to achieve seamless interoperability upgrades between Polygon zkEVM, PoS, and Supernets. The network has the ability to accommodate virtually unlimited chains and ensures secure and instantaneous cross-chain interaction without additional security or trust requirements. The Polygon team hopes that this solution will provide infinite scalability and a unified liquidity pool, similar to the vastness and accessibility of the internet.
Interoperability: Polygon 2.0 aims to enable easy flow and interaction between multiple blockchains by using a breakthrough cross-chain coordination protocol (specific details have not been disclosed). These interactions should feel like interacting with a unified, large-scale chain thanks to cost-effective and fast ZK proofs. This theoretically could unlock unprecedented levels of coordination and collaboration between different Polygon chains.
Security: By adopting zero-knowledge technologies in combination with existing PoS mechanisms, Polygon aims to increase security and privacy. They hope to fuse the advantages of both technologies in the upcoming Polygon 2.0. They recently announced that the transition of Polygon PoS to zkEVM validium is proof of this.
Scalability: The Polygon team is designing Polygon 2.0 to support “practically infinite chains,” aiming to achieve unprecedented levels of capacity and scalability through this method. Polygon co-founder Sandeep Nailwal has even expressed a vision of “unlimited and ever-growing block space.” The application of zero-knowledge technology will be crucial to achieving this vision, as ZK has the potential to significantly improve scalability.
Latest Updates and Roadmap
Governance will be an important part of the Polygon 2.0 transition. Polygon Labs President Ryan Wyatt emphasized that sustainable decentralized governance is the cornerstone of the Polygon 2.0 strategy. To this end, the latest update planned for Polygon 2.0 is to upgrade Polygon PoS to zkEVM validium through an initial governance proposal, described as the “first decentralized L2 secured by ZK proofs.”
The Polygon PoS sidechain is currently protected by its own set of validating nodes, rather than by ZK proofs, but the vision for Polygon 2.0 is for every Polygon chain to be a ZK L2. The solution proposed by the Polygon team is to upgrade the PoS sidechain to zkEVM validium. Validium is an extension solution that uses ZK proofs to ensure transaction validity but stores transaction data off-chain. This is in contrast to rollup solutions like Polygon zkEVM that store data on-chain. Publishing transaction data is expensive and can limit throughput. In this context, validium can be seen as a low-cost, high-throughput cousin of rollups. The main trade-off is that validium must ensure the availability of transaction data outside of Ethereum, while Polygon PoS can support data availability guarantees through their set of 100+ validating nodes.
The user and developer experience will not be affected by this change, and Polygon PoS will continue to operate as it has in the past, only with increased security due to the addition of ZK proofs. After this upgrade, both Polygon PoS and Polygon zkEVM will continue to exist as two public networks in the Polygon ecosystem. Additionally, both networks will use zkEVM technology, with one as a rollup and the other as a validium.
The Polygon team believes that these two networks will complement each other. Specifically:
Polygon zkEVM: Has the highest level of security, but has slightly higher fees and limited throughput. Best suited for high-value applications with the highest security priorities, such as DeFi.
Upgraded Polygon PoS (zkEVM validium): Has very high scalability and low fees, but slightly lower security compared to Polygon zkEVM. Best suited for applications that require high transaction volume and very low transaction fees, such as games and social applications.
If the preliminary proposal is supported, it will become a formal governance proposal and be discussed at governance meetings and other forums. If consensus is reached, the Polygon team plans to launch the zkEVM validium mainnet by the end of Q1 2024. The Polygon team also announced a planned timeline for more details on Polygon 2.0. There will be a series of announcements.
Figure 12: Polygon 2.0 announcement timeline.
Given Polygon’s performance in the PoS sidechain space and the successful deployment of their zkEVM solution, this new vision has significant potential and may offer interesting enhancements to the wider L2 ecosystem. We look forward to monitoring its development with the release of more details on Polygon 2.0.
5. Market Comparison
The L2 space is known for its vibrancy and competitiveness, providing us with a fascinating area of analysis. Each network brings unique perspectives in their own way, with notable differences despite some commonalities. Building on our previous discussions of L2 and their evolving visions, Figure 13 provides a preliminary understanding of this emerging market.
While all protocols have a grand goal narrative similar to that of Ethereum co-founders, Optimism stands out with its Superchain theory. There is a certain degree of overlap between zkSync and Optimism, considering the open-source zkEVM rollup SDK it proposes, which allows developers to flexibly exchange modules, similar to OP Stack Hacks. However, Optimism is in stark contrast to zkSync and Arbitrum, which present somewhat divergent visions in execution strategy, though with slight consistency in L3.
Figure 13: A visual representation of the relationship between the next generation Layer2 rollups and the Ethereum mainnet. Note that Polygon and Starknet are not included as we await further development from these projects.
So far, Starknet and Polygon have delivered unique strategies in shaping their own scaling theoretical approaches. Starknet echoes the approach of zkSync and Arbitrum to some extent, betting on a multi-layer structure with fractal scaling to L3. On the other hand, Polygon has chosen a more comprehensive model, merging its suite of L2 products together, including the ability for dApps to establish their own chains within the super network.
These different L2 visions are centered around custom and application-specific chains. However, specific details are still evolving, particularly for Polygon, which we eagerly await to reveal more about their Polygon 2.0 concept in the coming days and weeks. To further conceptualize these comparative highlights, Figure 14 captures some subtle differences between the networks.
Figure 14: A snapshot of Layer2 comparisons and their place in the next stage of Ethereum scaling.
It can be seen from the table above that the separation of the permissioned L2 development and permissionless L3 development is an interesting distinction for Arbitrum. Arbitrum Orbit provides a permissionless platform for creating L3 chains, but still requires DAO approval to use Arbitrum’s IP to develop L2. This semi-closed structure, while possibly limiting the interoperability of the Arbitrum Orbit chain with other L2 chains on Ethereum, effectively mitigates the risk of competing chains diluting Arbitrum L2 user activity. This is distinct from some other L2 designs. For example, Optimism and their OP Stack adopts a fully open-source framework in the Optimism code, enabling developers to deploy L2s using a shared sequencer set. While Optimism’s model does provide greater flexibility, it may inadvertently open the door for generic Rollups developed using the OP Stack to cannibalize Optimism’s liquidity and users.
However, the reality is that the growing network of chains on the superchain corresponds directly to the increasing revenue of Optimism, a significant portion of which is expected to flow to a wider Optimism ecosystem. This revenue increase is primarily attributed to Optimism’s role as a sequencer and fees from other OP chains. In contrast, Arbitrum Orbit shows a unique value accrual approach, requiring L3s built on Arbitrum to pay fees to its sequencer.
Regardless of the underlying technology, ensuring the creation of an essentially sustainable system is key for L2s aimed at expanding and unlocking network effects. In the complexity of open-source or closed-source, licensed or unlicensed L2s and L3s, several aspects should also be considered to determine how a project accumulates value in their respective ecosystems.
Software customizability: The convenience of copying code and building on top of L2s will be a key differentiating factor. Projects built on top of L2s may seek specific modifications and may invest resources in funding these customizations.
Transaction fees and sequencers: Sequencers play an important role in maintaining and operating expenses related to cloud services, and should be fairly compensated through transaction fees charged to users.
Role of governance tokens: With the trend towards decentralization, the role of governance tokens for L2s becomes more important. Most Rollups may require token staking to join the shared sequencer network.
As far as scalability, efficiency, and usability are concerned, the L2 ecosystem has made tremendous progress. Now discussions about the “next stage” of L2 evolution are taking place, leading to different approaches and visions for the future path. Vitalik Buterin himself has expressed his own views on this discussion, thinking about potential routes, but also pointing out the many possibilities for moving forward.
The value of L3s, Superchains, and hyperscalability is obvious, but standing on the threshold of these advancements, there is clearly no single, definitive vision or answer. Although there are differences in opinions, they share a common goal on the basic goal: infinite scalability. The ultimate goal is to create a Web3 world that seamlessly scales like Web2, providing the best playground for developers and users.
In fact, for this topic, we are still in the early stages, and projects have not strictly defined and executed their methods. Therefore, the views and definitions presented today are just a snapshot of a dynamic and evolving space. We look forward to witnessing how the next stage of Ethereum’s scalability unfolds and impacts L2s and the wider blockchain ecosystem.