Are L2s Parasites on ETH: The Evolution of Ethereum’s Scaling Journey

Recently, Vitalik Buterin has stepped up his efforts to “push” the growth of Layer 2 (L2) solutions. On September 6, the Ethereum co-founder announced that he would no longer invest in L2 projects. Following this, on September 12, he posted another decision: any L2 that fails to reach Stage 1 by the end of the year will be “kicked out” of the rollup category. While these statements seemed firm, on September 17, Vitalik’s name unexpectedly appeared on the investor list of RISE Chain — a project labeled as an Ethereum L2. Despite various explanations, this incident suggests uncertainty from both Vitalik and Ethereum regarding the L2/rollup path.

The Love-Hate Trilogy: From Embracing to Doubting L2s

Ethereum’s most notable move on its roadmap was The Merge, shifting from a Proof of Work (PoW) to a Proof of Stake (PoS) mechanism, leading to 0.2% savings on global electricity consumption. Subsequently, Ethereum adopted a rollup-centric scaling strategy, leading to the booming L2 ecosystem.

Phase 1: Rollup-Centric Declaration

Simply put, rollups excel in generality and flexibility. A rollup can run a complete Ethereum Virtual Machine (EVM), allowing existing Ethereum applications to migrate to rollups with minimal code changes. This compatibility makes rollups an appealing scaling solution, enhancing performance while maintaining alignment with the current system.

Rollups work by bundling multiple processed transactions into batches on the L2 chain and executing them before submitting the results to the Layer 1 (L1) blockchain. This method significantly reduces the data requirements and transaction costs on L1 while preserving data integrity and security. However, the rollup scaling route comes with challenges and trade-offs, including differences between rollup solutions and issues related to sequencing and fraud-proof mechanisms.

Phase 2: Separating L2 from Rollups

The technical roadmap of L2s and rollups has been redefined by Vitalik, triggered by the rise of Celestia. In simple terms, Celestia embodies modular concepts, breaking down blockchain into modules for settlement, consensus, data availability (DA), and execution. Celestia specifically handles the DA layer. While, rollups bundle L2 transactions and confirm them collectively on L1 to save gas fees. Celestia extends this approach by allowing L2 transactions to be published on its network, bypassing Ethereum’s high data storage costs.

Thus, the protection of Ethereum’s central role led Vitalik to grow wary of the L2 concept. Moving forward, rollups are considered L2s, but not all L2s will qualify as rollups.

In Vitalik’s view, the key difference between rollups and L2s lies in their goals and implementation. Rollups are specific L2 scaling solutions that aim to compress transaction data and submit it to the Ethereum mainnet (L1) to improve throughput and lower fees. Rollups come in two main types: ZK-Rollups and Optimistic Rollups, each varying in security, flexibility, and latency.

In contrast, L2 is a broader concept encompassing scaling solutions built outside the main blockchain, focusing on enhancing performance and scalability. L2 solutions include rollups, state channels, Plasma, and more. Simply put, solutions that use Ethereum as the DA layer qualifies as rollups. Anything outside this definition may fall under other L2 categories.

Phase 3: Further Defining Rollup Stages

Vitalik has proposed three stages to classify rollups based on their dependency on human intervention: Stage 0, Stage 1, and Stage 2.

• Stage 0: The initial stage where all transactions must be processed on-chain. Rollup projects at this stage rely entirely on centralized auxiliary systems for transaction validation and bundling.
• Stage 1: Rollups begin reducing their dependence on centralized systems, partially automating processes through smart contracts. Although some manual intervention remains, this stage marks a move toward decentralization.
• Stage 2: The ultimate goal, where rollups are fully governed by smart contracts and adopt a permissionless fraud-proof system, achieving a “training wheels off” state. At this stage, all operations occur in a decentralized environment without the need for human intervention.

Vitalik believes that only rollups reaching at least Stage 1 are worth recognition and discussion, as they demonstrate potential progress toward greater security and decentralization. According to L2BEAT data, as mentioned earlier, only four rollups — Arbitrum One, OP Mainnet, dYdX v3, ZKsync Lite, and two reaching Stage 2 (DeGate V1 and Fuel v1) — currently meet these criteria. While L2s are thriving, only a handful truly meet the standards of security and decentralization.

Post-Rollup: Sharding Still a Distant Goal

With Ethereum’s shift to PoS, EIP-4844 has become a critical component of the Cancun upgrade. It aims to address Ethereum’s scalability issues by creating space for data “blobs,” allowing L2 solutions to use this blob space for settlement instead of calldata, thereby reducing L2 costs.

After EIP-4844, Ethereum’s expansion roadmap will focus on several key areas:

• Sharding: While off-chain scaling solutions like rollups and validiums can expand Ethereum in the short term, sharding remains the fundamental long-term strategy. Sharding will split Ethereum L1 into multiple shard chains with shared security, significantly increasing overall throughput.
• Storage Solutions: As data throughput and storage pressures grow, EIP-4844 has standardized the remove of historical data, introducing challenges like the need for centralized facilities to resettle historical data. Hence, the community is exploring more decentralized storage solutions, such as the EthStorage incentivized modular storage network.

The Ethereum community is actively promoting various scaling solutions to address network congestion. Rollup technology holds immense potential and advantages, and Ethereum will likely continue optimizing its scaling strategies to maintain decentralization and security while improving transaction speed and throughput.