Ant Group Announces Open‑Source Privacy Computing Framework “YinYu 1.0” and the First Domestic Financial‑Grade TEE Solution “HyperEnclave”

On July 7, 2023, the World Artificial Intelligence Conference’s “Data Elements and Privacy Computing Summit” opened, and Ant Group released two privacy‑computing open‑source products: the “YinYu Open‑Source Framework 1.0” and the first domestic financial‑grade TEE solution “HyperEnclave”. These technologies are key components of Ant’s six‑year “Trusted Privacy Computing YinYu Technology Stack”, enabling industrial‑grade privacy computing and previously earning the “Eight Treasures of the Museum” award at the 2022 World AI Conference.

The YinYu Open‑Source Framework constitutes the compute‑engine layer of the stack, with HyperEnclave serving as its trusted base. The stack is entirely self‑developed, holds over a thousand patents, and integrates the YinYu framework, the Occlum TEE operating system, HyperEnclave and other leading privacy‑computing technologies, incubating new techniques such as Trusted Encrypted Computation (TECC). The open‑source release seeks a leap‑frog improvement in usability and universality, helping AI applications become safer and more trustworthy while unlocking data value.

Deep AI applications raise higher demands on data, algorithms, and compute, as well as on security, privacy, and ethics. Ensuring data safety, privacy protection, and solid AI ethics and security is a prerequisite for AI to deliver real value. Privacy computing, as a crucial technical support, is gaining new momentum, yet ease‑of‑use and universality remain industry bottlenecks.

YinYu 1.0 achieves leap‑frog improvements in open‑source scope, performance, and ease‑of‑use. The framework, first open‑sourced in July of the previous year, was upgraded to version 1.0 after one year. The new version offers three major advantages: (1) an open‑source Kuscia privacy‑computing task orchestration framework that resolves integration issues such as port merging and API access, supporting inter‑connection or embedded deployment with third‑party systems; (2) added support for the SS‑LR open algorithm protocol, enabling full‑stack black‑box/white‑box interoperability; (3) a “plug‑and‑play” lightweight deployment package that further lowers the entry barrier.

At the forum Ant also open‑sourced the financial‑grade TEE solution “HyperEnclave”. TEE is one of the privacy‑computing routes, regarded as the most effective method for secure cloud data and privacy‑preserving computation. HyperEnclave’s advantages are security and compatibility: it supports major domestic and foreign CPU platforms, provides a unified TEE abstraction, and its core code has undergone formal verification. It hosts the root of trust in a trusted authority, meets domestic self‑development requirements, and already has commercial‑scale deployment experience. The open‑source release is expected to bring a more transparent, trustworthy, unified, and generic domestic TEE solution to the industry.

IEEE officially released the industry’s first “Trusted Execution Environment Security” international standard IEEE2952‑2023 “Standard for Secure Computing Based on Trusted Execution Environment”. Ant Group, China Academy of Information and Communications Technology, Shanghai Jiao Tong University, InsightTech, and Chongli Online participated in the ceremony. Led by Ant, the standard defines a technical framework for secure computing systems based on TEE, covering isolation, confidentiality, compatibility, performance, availability, and security, and specifies use cases and scenarios, providing effective guidance for the industry.

Additionally, China Academy of ICT, China Telecom, China Mobile, China Unicom, Ant Group and InsightTech jointly released the “Privacy Computing Cross‑Platform Interoperability Open Protocol Part 2: SS‑LR”, building a more open, transparent, and secure privacy‑computing inter‑connect platform to protect data security.

Historical opportunity and technological change are pushing data intelligence to unprecedented heights while intensifying data‑security challenges. Data flow moving toward encrypted states is the future trend. Under data‑encryption requirements, privacy‑computing methods, platform frameworks, and technical standards face new transformations. Ant will continue to expand the openness and breadth of privacy‑computing core products, working with the industry to build a data‑security moat for the AI era.