Protecting Wasm Workloads with Confidential Computing and TEE...
summary
Shows a working pipeline that runs a WASM inference binary (WASI) inside hardware TEEs with attested mutual TLS: manager spins confidential VMs (AMD SEV‑SNP and Intel TDX support), agent loads a WASM algorithm, verifies checksum, performs remote attestation, and returns the result (demo OCR inference correctly outputs digit 4). Uses Burn AI for model build, Enarx and Ultraviolet’s CocoSE AI for orchestration, and is validated in the EU ELASTIC project (13 partners, €5M).
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:25Confidential computing basics
- 3:32TEE hardware limits
- 5:19WebAssembly in TEEs
- 6:00Elastic EU project overview
- 7:05Enarx for Wasm in TEEs
- 8:30Coco SE AI platform
- 10:18Coco SE AI architecture
- 14:29Attested TLS and attestation
- 15:53Burn AI for confidential ML
- 16:16Demo setup and OCR
- 21:03Computation assets linking
- 23:10Wasm inference workload
- 24:58Demo results and security
- 25:21Future research directions
- 27:47Closing and Q&A
related talks
36mBringing honest computing to the WASM world
30mA WASM Runtime for FaaS Protected by TEE
24mStrictly Confidential: Use Cases & Demos of AI Workflows in TEEs
42mSponsored Session: Confidential Computing - New Capabilities for New Workloads- 42mSponsored Session: Confidential Computing - New Capabilities for New Workloads
40mConfidential Computing Across Multiple Clouds and Platforms: Project Enarx
37mIt's All About Trust: Why Confidential Computing Matters
38mEnarx for Confidential Computing