This article presents an end‑to‑end privacy‑preserving machine‑learning solution for AI and big‑data workloads built on Intel SGX, the open‑source TEE OS Occlum, and BigDL PPML, detailing its architecture, key features, deployment steps, and real‑world applications.
This presentation explores the background and current state of privacy computing, its relevance to big data and AI, discusses SGX and LibOS technologies, introduces the BigDL PPML solution for secure Spark/Flink workloads, and reviews real-world applications and future outlook.
The article explains how confidential computing, built on trusted execution environments like Intel SGX, addresses data‑in‑use security, outlines the technical hurdles developers face, and showcases Ant Group's open‑source SOFAEnclave components—Occlum, HyperEnclave, and KubeTEE—highlighting Rust’s pivotal contribution.
This article introduces confidential computing with Intel SGX, explains why SGX development is difficult, and demonstrates how Ant Group's open‑source TEE OS Occlum dramatically lowers the barrier by using three simple commands to build, package, and run a Hello World program inside an enclave.
The article chronicles Tian Hongliang’s evolution from a Rust‑loving coder who excelled in Ant Group’s internal coding competition to a leading researcher in confidential computing, detailing his work on Intel SGX, the open‑source Occlum project, and the team’s recruitment drive for security engineers.
