The meeting focused on reviewing the PQC threat model document and discussing potential new topics for the Device Security working group. Srikanth provided an update on device security topics, highlighting DRM governance and suspicious devices as areas for further exploration. The group then reviewed the PQC threat model, with a focus on the primary threat of forging secure boot and firmware signatures with PQC, and the impact on long-lived device identities and attestation. There was extensive discussion on how to define and communicate the severity of these threats, with suggestions to provide context or use classification buckets rather than a simple severity scale. Glenn, Filip, Yasser, Lars, and others contributed to the discussion, emphasizing the need for a framework that is clear for both technical and non-technical audiences. The conversation ended with a reminder to review and contribute to the documents ahead of the next call in two weeks.
- Srikanth to update the PQC threat model document to clarify that the private key can be derived from the public key in the context of a CRQC, and tag Mark.
- Srikanth to consider and potentially add chipset manufacturers as impacted parties in the threat model document, and clarify the meaning of “impacted” in the context.
- Srikanth to develop a proposal for defining severity in the threat model document, including a rationale, to be reviewed at the next meeting.
- All participants: Review the device security topics document on Google Drive, vote on existing topics, and/or add new topics for consideration by the working group.
- Antitrust Statement and WG Discussion
Device Security and DRM Governance
- Srikanth provided updates on device security topics, including DRM governance and suspicious devices, and encouraged participants to review and vote on potential topics in the shared document.
- He explained the goals of DRM governance, which involves using existing DRM channels to carry policy information for media use cases like AI-based enhancements, and outlined the intent behind sharing intelligence on suspicious devices to combat piracy and fraud.
- Srikanth also discussed the PQC threat model document, encouraging participants to review, contribute, and provide feedback to help shape industry expectations for PQC.
PQC Threat Model Updates
- The team discussed updates to a PQC (Post-Quantum Cryptography) threat model document, which is being created as an extension to an existing OTT threat model document rather than augmenting it directly.
- The primary threat identified was forging secure boot and firmware signatures using PQC, where an attacker with access to a public key could potentially recover the private key and run arbitrary software on devices.
- The team agreed this was a crucial issue given that many devices will remain in service until the 2030s with hard-coded cryptographic roots.
- They also discussed a secondary threat regarding compromise of long-lived device identities and attestation, particularly affecting service providers, content owners, and DRM systems.
- Mark suggested clarifying the language around deriving private keys from public keys to better communicate the severity of the threat.
Certificate Key Impact Documentation Strategy
- The team discussed how to define and document the impact of certificate and attestation key changes across different stakeholders.
- Glenn suggested separating chipset manufacturers from device manufacturers as distinct impacted parties, while noting that technical teams should focus on providing threat assessments rather than business analysis.
- The group agreed to provide examples of technical threat assessments that business executives could adapt for their specific environments, rather than creating a unified business-focused document, acknowledging that engineers should stay within their technical expertise.
Severity Levels Implementation Discussion
- The team discussed how to define and implement severity levels for impacted sections, with Yasser suggesting the need to differentiate between device revocation and content quality settings.
- Philip expressed concerns about being prescriptive about severity interpretations across different businesses, while Glenn proposed using classification buckets with contextual explanations rather than definitive severity rankings.
- The group agreed that severity levels should include contextual information about why an issue is severe, with Srikanth confirming they would incorporate this feedback into the definition.
Security Vulnerability Severity Framework
- The team discussed developing a severity framework for security vulnerabilities in media streaming services, with participants debating whether to use existing frameworks like CVSS or create a simpler, more tailored approach.
- Brian emphasized the need for technical teams building streaming services to understand security implications, while Roger raised concerns about CVSS being too complex and service-dependent for their needs.
- The group agreed to pursue a simpler framework that would distinguish between hardware-bound signatures (high severity) and software-updatable certificates (lower severity), with Srikanth committing to draft a definition for severity and present it at the next meeting in two weeks.
Next Meeting: Tuesday, September 29, 2026 8am PDT