Multi-Platform and Vault-Free Attestation of Confidential VMs

  • Davi Pontes UFCG
  • Fernando Silva UFCG
  • Anderson Melo UFCG
  • Asm Asadujjaman UFCG
  • Eduardo Falcão UFRN
  • Andrey Brito UFCG
  • Carlos Filho UFCG

Resumo


Confidential virtual machine (CVM) offers are increasing among most cloud providers and provide additional confidentiality and integrity features. Nevertheless, performing remote attestation of CVMs is not trivial. The integrity and confidentiality verifications become even more difficult due to the different types of Trusted Execution Environments (TEEs) supported by the public clouds and how their support is exposed to users. We propose an approach to attest different offers of CVMs, considering both different technologies and providers, in a homogeneous fashion. We leverage the open-source standard from Cloud Native Computing Foundation, the Secure Production Identity Framework for Everyone (SPIFFE), and its reference implementation, the SPIFFE Runtime Environment (SPIRE). We propose a new node attestor for Intel Trust Domain eXtensions (TDX) CVMs and propose combining it with other attestation strategies supported by SPIRE. This allows the creation and attestation of a multi-platform infrastructure comprising different CVM TEEs and multiple cloud providers. Finally, we propose a vault-free approach to provide cryptographic keys to CVMs directly on the SPIRE attestation process, avoiding the need to trust a critical key management service. This combination considerably simplifies the setup, operation, and usage of confidential cloud-native applications since microservices could be deployed on CVMs in public clouds, relying only on provider-agnostic, open-source software.
Palavras-chave: confidential computing, Intel TDX, AMD SEV-SNP, SPIRE
Publicado
26/11/2024
PONTES, Davi; SILVA, Fernando; MELO, Anderson; ASADUJJAMAN, Asm; FALCÃO, Eduardo; BRITO, Andrey; FILHO, Carlos. Multi-Platform and Vault-Free Attestation of Confidential VMs. In: LATIN-AMERICAN SYMPOSIUM ON DEPENDABLE COMPUTING (LADC), 13. , 2024, Recife/PE. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2024 . p. 241–251.