One Ring -3 To Secure Them All: Computing with Hardware Enclaves
summary
Trusted execution environments like Intel SGX, ARM TrustZone, Apple Secure Enclave and AWS Nitro Enclaves run sensitive code inside a protected area so the OS and VM cannot read secrets. Trade offs reported include measurable throughput loss from marshalling and context switches, vulnerability to microcode or speculative execution fixes that require reboots and tokens for attestation, and vendor gating for production keys. Demo code seals and unseals a random value with a sealing/unsealing demo and runs in simulation mode on GitHub.
AI-written summary · about
chapters (15)
- 0:00Start
- 1:07Trusted Execution Environment
- 1:27How do you solve secure execution for critically sensitive data?
- 6:01It represents a change in the Threat Model
- 8:16First Impressions
- 8:33Ergonomics
- 9:03Broad application: arm
- 9:44Most depth: intel
- 10:42Considerations
- 12:16The Threat Model is changed, but threats still exist
- 12:39Microcode issues are difficult to fix
- 14:35Some designs rely on protection rings vs true separation
- 15:15Some implementations
- 17:33Hardware Enclaves are a key component in advancing the design of secure software
- 19:25Secure design should consider tactical use of hardware to solve difficult trust problems
related talks
37mIt's All About Trust: Why Confidential Computing Matters
36mAttacking & Defending Kubernetes TEE Enclaves in Critical Infrastructure
42mSponsored Session: Confidential Computing - New Capabilities for New Workloads
41mPanel: A New Root Of Trust: Changing Computing With TEE
35mVeracruz: Privacy-preserving mutliparty computation using trusted hardware- 42mSponsored Session: Confidential Computing - New Capabilities for New Workloads
28mSanctum: Minimal Hardware Extensions for Strong Software Isolation
40mKeeping secrets in your opponent's hands