Revisiting Rollbacks on Smart Contracts in TEE-protected Private Blockchains

Abstract

Blockchain technology offers decentralized security but fails to ensure data confidentiality due to its inherent data replication across all network nodes. To address these confidentiality challenges, integrating blockchains with Trusted Execution Environments (TEEs), such as Intel SGX, offers a viable solution. This approach, by encrypting all data outside the SGX enclave and making them unrecognizable to untrusted network nodes, ensures secure processing of data and computations within TEEs. Fabric Private Chaincode (FPC), an enhancement of Hyperledger Fabric, demonstrates this integration by securing smart contracts in enclaves, thereby enhancing confidentiality. However, FPC's reliance on states stored on the blockchain introduces vulnerabilities, especially to rollback attacks. This work provides a detailed analysis of rollback attacks in FPC, evaluates existing protection mechanisms, and proposes a solution: a Merkle Tree method implemented in an FPC application named SecretKeeper. Through experimental validation, this solution shows significant security enhancements against rollback attacks within FPC contexts.