Ethical Hacking News
A recent study has revealed a fundamental flaw in attested TLS, the security protocol underlying confidential computing. The vulnerability, known as relay attacks, allows malicious servers to intercept and encrypt client traffic, compromising the integrity of sensitive data. As Europe's sovereign cloud ambitions continue to grow, this finding highlights the need for greater transparency and accountability within the industry.
The security protocol underlying confidential computing, attested TLS, may be fundamentally flawed due to vulnerability to relay attacks. Three out of seven proposed binding mechanisms achieve level one, while the rest fail even this baseline. Post-handshake attestation may not be achievable within the current architecture of intra-handshake attestation. The Confidential Computing Consortium (CCC) Special Interest Group failed to publish research findings due to lack of transparency and accountability.
Confidential computing has been touted as a key component of Europe's sovereign cloud ambitions, promising to provide verified protection for sensitive data. However, a recent study has revealed that the security protocol underlying this technology may be fundamentally flawed.
The protocol in question is attested TLS, which aims to prove cryptographic trust in the system by generating evidence during the TLS handshake itself. This evidence is then used to vouch for the authenticity of the server and its software. However, independent research has shown that this protocol is vulnerable to relay attacks, where a client can be tricked into encrypting its traffic to a malicious server.
The researchers, led by Muhammad Usama Sardar, used a formal verification tool called ProVerif to analyze attested TLS protocols. They discovered that three out of seven proposed binding mechanisms achieve level one, while the rest fail even this baseline. Level one ties evidence only to the very first key exchange in the handshake, while level three requires binding the evidence to the application traffic key itself.
Sardar's team also found that post-handshake attestation, which waits until after data starts flowing, may not be achievable within the current architecture of intra-handshake attestation. This is because by the time the key exists to bind against, the evidence has already been sent, unless the TLS protocol itself is significantly changed.
The researchers' findings have prompted institutional responses from various organizations, including the IETF's Secure Evidence and Attestation Transport (SEAT) working group, which has incorporated formal verification into its charter. However, Intel and Google continue to market confidential computing as proof of sovereign, verified protection, despite acknowledging the vulnerability in their responses.
The delay in publishing the research findings highlights a more significant issue: the lack of transparency and accountability within the industry. Sardar's team had to publish the evidence themselves due to the inaction of the Confidential Computing Consortium (CCC) Special Interest Group, which governs the adopted proof-of-concept project.
This raises questions about the regulatory environment surrounding confidential computing and the ability of vendors to market their products as sovereign solutions without proper guarantees. As Europe's digital sovereignty push continues, it is essential that policymakers and industry leaders prioritize transparency, accountability, and security in their efforts to protect sensitive data.
Related Information:
https://www.ethicalhackingnews.com/articles/The-Untrustworthy-Handshake-A-Critical-Review-of-Confidential-Computings-Security-Mechanism-ehn.shtml
https://www.theregister.com/security/2026/07/04/confidential-computings-trust-mechanism-is-broken-the-fix-may-not-exist/5266056
Published: Wed Jul 8 10:21:37 2026 by llama3.2 3B Q4_K_M
