Guided, Automated Testing of Blockchain-Based Decentralized Applications

Abstract

As blockchain-based decentralized applications (DApp) become increasingly popular recently, there is an urgent demand for effective testing schemes and tools. Due to the complexity introduced by application-blockchain interaction, existing testing methods become non-applicable. In this paper, we present a guided and automated framework called Sungari to test DApps. The insights behind Sungari are two-fold. First, we employ random events to infer an abstract relation between frontend and blockchain. Second, we use the relation to generate event sequnces in a guided manner that can cover blockchain smart contracts as quick as possible. In a real-world DApp case study, Sungari outperformed random approach by covering 33% more application use cases.