Towards Verifiable Trust Proof for Trusted Confidential Virtual Machines
Jingkai Mao, Xiaolin Chang, Lin Li, Haoran Zhu, Junchao Fan · IEEE Transactions on Network Science and Engineering · 2025
Trusted-Execution-Environment (TEE)-based Confidential Virtual Machines (CVMs) deployed in public clouds currently lack verifiable trust proof of their running state, undermining remote users' confidence in using them. While Virtual Trusted Platform Module (vTPM) techniques can be leveraged to generate such proof for CVMs, existing vTPM-based approaches suffer from three key weaknesses: lack of a well-defined root-of-trust, lack of vTPM protection, and lack of vTPM's trust proof. Consequently, it is hard for these approaches to generate verifiable trust proof to represent the states of remote users' CVMs. This paper proposes T3CVM, an approach that leverages vTPM to generate verifiable trust proof for remote users' AMD SEVbased CVMs. T3CVM comprises three components and a trusted boot scheme to overcome the weaknesses of the existing state-ofthe- art approaches and then increase user's confidence in the trust proof. Our approach allows trusted computing-based applications to run seamlessly in the trusted CVM. Additionally, our approach can be extended to the CVMs based on other VM-based TEE technologies, such as Intel TDX. We formally analyze T3CVM's security and evaluate its performance through a prototype implementation.