Emer: Reputation-Based Event Consumer for Event-Driven Decentralized Systems

Yepeng Ding, Lingfeng Zhang, Junwei Yu, Hiroyuki Sato, Tohru Kondo, Kouji Nishimura

Abstract: Event-driven decentralized systems trigger off-chain functions upon consuming events emitted by decentralized applications deployed on blockchains. However, ensuring dependable, cost-efficient, and flexible event consumption is challenging due to the consensus mechanisms inherent in blockchains. Meanwhile, the need for dependable event consumption has become increasingly critical with the rise of decentralized finance. In this paper, we propose Emer, a reputation-based event consumer that adopts a distributed architecture and leverages a novel asynchronous reputation mechanism to mitigate the risk of missing or inconsistent events. Besides, Emer reduces the communication cost through reputation-based optimization and proves scalable as an infrastructure. Additionally, Emer offers trust condition customization to enable prioritizing sensitivity of timeliness or correctness for specific use cases. Furthermore, we support our claims with simulation analysis and demonstrate its applicability through a real-world scenario in decentralized finance.