Towards Stronger Blockchains: Security Against Front-Running Attacks
Anshuman Misra, Ajay D. Kshemkalyani
TL;DR
This paper addresses front-running in blockchains by showing it constitutes a violation of causal ordering and proposing a causality-preserving total order, termed a strong blockchain. It develops a modular protocol that couples Byzantine-tolerant broadcast, threshold cryptography, and round-based synchronization to enforce that only safe transactions with completed causal history enter consensus, achieving a causally consistent total order. The authors formalize front-running as a causal violation, prove correctness and fairness, and establish practical bounds such as a message complexity of $O(n^2)$ and a latency bound of $(\21) rounds$ under synchronous permissioned settings with at most $t-1$ Byzantine among $n$ miners. The approach preserves application semantics, enabling classic centralized applications to be securely implemented on blockchains, and offers a path toward stronger security guarantees without overhauling existing consensus mechanisms.
Abstract
Blockchains add transactions to a distributed shared ledger by arriving at consensus on sets of transactions contained in blocks. This provides a total ordering on a set of global transactions. However, total ordering is not enough to satisfy application semantics under the Byzantine fault model. This is due to the fact that malicious miners and clients can collaborate to add their own transactions ahead of correct clients' transactions in order to gain application level and financial advantages. These attacks fall under the umbrella of front-running attacks. Therefore, total ordering is not strong enough to preserve application semantics. In this paper, we propose causality preserving total order as a solution to this problem. The resulting Blockchains will be stronger than traditional consensus based blockchains and will provide enhanced security ensuring correct application semantics in a Byzantine setting.