Security of Blockchains with Strategic Entities

Abstract

Blockchains are typically thought to provide a simple interface --- a consensus layer on which rich applications can be built. While we have a good understanding of the conditions under which consensus can be securely realized, the interaction between the consensus layer and applications, turns out more complex --- recent works have shown that participants in the consensus protocols (block proposers or miners) can impact application’s security in subtle but serious ways. For example, in financial applications such as exchanges, proposers can profit off users by, e.g., placing their own transactions before users’ in so-called frontrunning attacks. Daian et al. (Flash boys 2.0, S\&P '20) coined the term Miner’s Extractable Value (MEV) to generalize such profits. Since 2020, a total of more than \$600M in MEV has been extracted, out of which more than \$120M was extracted in 2022. To make matters worse, not all MEV can be detected. In particular, bribery attacks -- where adversaries compensate proposers to violate protocol invariants or censor transactions -- represent a form of extractable value that is difficult to detect and prevent.These observations challenge the conventional abstraction that the consensus provides a neutral ordering service. Instead, consensus participants are economically motivated agents who can influence transaction inclusion and ordering which modifies the execution in ways that directly affects the security and fairness of the blockchain. This thesis argues that securing blockchain applications requires reasoning explicitly about the incentives and strategic behavior of consensus participants. To this end, this thesis develops a mechanism-design framework for analyzing and constructing blockchain protocols that remain secure even when proposers act rationally to maximize profit.Using this perspective, this thesis establishes fundamental limits on what security and efficiency guarantees can be achieved in adversarial blockchain environments and introduces new protocol mechanisms that align proposer incentives with application-level security. In particular, this thesis demonstrates how carefully designed economic constraints can increase the cost of censorship, prevent strategic manipulation in auction and transaction-ordering mechanisms, and ensure robust fee allocation even in the presence of execution uncertainty. These results show that protocol security cannot be achieved solely through cryptography or consensus correctness, but must also be enforced through incentive-compatible mechanism design.The central thesis of this work is that the security of blockchain applications fundamentally depends on the economic incentives of consensus participants, and that protocol-level mechanism design is necessary to ensure robust security in adversarial environments. By explicitly modeling proposers as strategic agents and designing incentive-compatible protocols, this thesis establishes principled foundations and practical mechanisms for building blockchain systems whose security holds even in the presence of rational, value-extracting participants.