DataAttest: A Framework to Attest Off-Chain Data Authenticity

In recent years, more and more applications are built on blockchains for boosting security, while keeping interaction with the off-chain world. These applications usually leverage blockchain oracles to access off-chain data. However, they are still threatened by the weakness of data authenticity as the external data source system may be attacked and provide forged and malicious data. In this paper, we propose DataAttest, a framework based on remote attestation to guarantee the authenticity of off-chain data. DataAttest first instruments the data source system for collecting runtime traces. Then, it measures the static code and runtime traces in TEE to construct a Data Authenticity Proof (DAP). Finally, blockchain oracles verify the DAP to guarantee that the integrity of the specific data source system is undamaged and the provided data is actually obtained from that system. In addition, DataAttest also adopts two optimization strategies to accelerate DAP verification and minimize the overhead brought by DAP construction. We implement DataAttest and evaluate it on a real-world system. The experimental results show that DataAttest can effectively check data authenticity and the overhead is relatively low.