Use case deep dive: Secure and privacy-preserving cyber threat intelligence exchange
Talkat Confidential Computing Mini Summit at Open Source Summit Europe 2023 · on Monday, 2023-09-18·
summary
Claims a practical stack: federated learning with homomorphic encryption for per-organization confidentiality plus confidential computing to protect aggregation integrity. Implementation trades scalability for correctness, building a custom secure aggregation scheme for tens of organizations instead of cloud-scale protocols. Handles petabytes by pushing preprocessing to data owners, uses VM-based confidential enclaves for I/O performance, and shows the approach raises trust and reduces raw data sharing — effective but brittle, requiring fine-tuning and policy attestation.
AI-written summary · about
chapters (11)
- 0:00Cyber threat intelligence sharing intro
- 1:01Challenges in CTI sharing
- 3:54Why CTI sharing is complex
- 9:11Current workarounds limits
- 10:57Input/output privacy in CTI
- 13:24Confidential computing for CTI
- 15:51Trust relationships in CTI
- 16:56Stacking tech for CTI sharing
- 20:51Future of blind trust in CTI
- 21:55Privacy as moving target
- 23:39Q&A
related talks
19mTransform CyberSec Governance with GenAI
24mSafeguarding Sensitive Data Access At Scale with Privacy-Enhancing Technologies
15mKeynote: Trustworthy AI with Confidential Computing
23mConfidential computing: from niche to mainstream
40mConfidential Computing and Privacy-Enhancing Technologies - The Landscape
30mSharing without Sharing: the Case for Data Clean Rooms Powered by Confidential Computing
25mImproving Safety with Privacy: An Exploration of Privacy...
39mData Privacy Engineering to power Responsible Enterprise AI