Security Analysis of the Development Stages of L2 Networks

The Ethereum community has been paying attention to the security development phase of Layer 2 (L2) networks. Recently, community members proposed a new standard to assess whether L2 networks have reached a high level of security. This standard requires L2 networks to operate on the Ethereum mainnet for more than six months and to have a total locked value (TVL) consistently exceeding $100 million, with at least $50 million in ETH and major stablecoins.

The co-founder of Ethereum delved into this, categorizing the security of L2 networks into three stages:

1. Phase 0: The Security Committee has full control and can easily override the results of the proof system.

2. Phase 1: Approval from more than 75% of the security committee members is required to override the system's operational results, and it must include members outside the main organization.

3. Stage 2: The Security Committee can only intervene in clear cases of error and can only choose from the proposals presented, without making arbitrary changes.

These three stages reflect the gradual decrease in the "voting share" of the security committee within the system.

The key issue is how to determine the best timing for transitioning from one stage to the next. This mainly depends on the level of trust in the proof system. The more confidence one has in the proof system, the more one should advance to a higher stage.

Through a simplified mathematical model, we can quantify the security at different stages. The model assumes that each member of the security committee has a 10% independent failure probability and considers the decision-making mechanisms at different stages. The results show that as the quality of the proof system improves, the optimal choice shifts from stage 0 to stage 1, and then to stage 2.

However, the reality is more complex than the model. There may be correlations between the members of the security committee, and the proof system itself may consist of multiple independent systems. These factors make Stage 1 and Stage 2 more attractive than the model predicts.

In theory, the existence of Phase 1 may not be necessary, and one could jump directly from Phase 0 to Phase 2. However, considering the response speed in emergencies, retaining Phase 1 may be more prudent. At the same time, entering Phase 2 too early may also pose risks, especially if it affects the strengthening work of the underlying proof system.

Ideally, data providers should demonstrate the audit and maturity metrics of the proof system while showing the development stage of the L2 network. This will help users and developers better assess the security and reliability of the L2 network.