Positive Reputation Score for Bitcoin P2P Network

Abstract

Cryptocurrency relies on the underlying P2P networking to deliver the up-to-date block and transaction information to synchronize and agree on the processed financial transactions. To maintain healthy connectivity, Bitcoin has implemented a negative reputation scheme based on ban score, which tracks the peer node behavior and results in dropping the peer connection if exceeding a threshold. Previous research in cryptocurrency security showed that such ban score can be both ineffective (easy to bypass to un-ban itself) and vulnerable (against spoofing-based defamation). We design and build a novel positive reputation scheme that addresses such deficiencies of ban score by having the peer behavior only positively impact our score and by disincentivizing switching the network ID for a new reputation score. We provide a general framework for the positive reputation scheme to enable the use of different networking sensing and measurements as well as different aggregate weights for scoring. Then, we provide a concrete instance and build of the positive reputation score scheme focusing on the peer's behavior in relaying unique/new blocks and transactions. We implement, experiment, and analyze our scheme on an active Bitcoin node connected to the Mainnet to show how our scheme provides the reputation and informs the beneficialness of each of its peer connections (delivering new block and transaction information).