Problem Description
Geographic decentralization is crucial for a censorship-resistant blockchain network, as it distributes validators across different regions, reducing risks from regulations. It also enhances the system's robustness by ensuring a globally distributed power structure and fairness — preventing any single region from having a disproportionate advantage. However, a fundamental tension exists: while a broader geographic distribution has many advantages, it also increases latency. Lower latency is important to validators because most of them participate in MEV-Boost auctions, where they can earn higher rewards by waiting longer for better bids — a well-known phenomenon called timing games in PBS. Validators are incentivized to co-locate to minimize latency and maximize rewards, similar to what happened in TradFi, where high-frequency traders adopt the same strategy — an effect famously described in Flash Boys.
Building on ongoing efforts at Flashbots Research, this project uses a simulation-based approach to analyze the current state of the Ethereum protocol under PBS and explore alternative proposals for the block-building process such as concurrent block proposers — examining their impact on latency and influence on incentives for co-location or geographic decentralization. Based on the simulation results, we further examine whether their utilities provide sufficient incentives to sustain geographic decentralization. If not, we analyze what the equilibrium for validator co-location would be and how quickly the system converges to it.
Motivation
The tension raises a question: Can we design a system where the utility gap between a random geographic topology and full validator co-location is minimized, thereby preserving network resilience without significant latency penalties? To answer this, we need to understand how a validator's geographical location influences latency and the resulting equilibrium.
Given the complexity of the scenario, a simulation-based approach may be well-suited as it allows for simulating complex interactions and approximating equilibrium when analytical solutions are infeasible.
References
- ‣
- ‣
- ‣
- Timing game (latency)
- ‣
- ‣
