Data Security and Privacy Protection
Data privacy is critical to Alibaba’s business. Recently, laws and regulations have been published to enforce data privacy protection, such as the Data Security Laws published this year. Therefore, Privacy Preserving Computing (PPC) technologies are highly needed in Alibaba. We are building high-efficiency, low-cost, user-friendly and secure PPC systems. Our internal research team would like to collaborate with academic friends who have the same interests and vision for PPC. In this topic, we focus on Hardware-Software Co-design, particularly novel hardware architecture designs.
Hardware-Software Co-design for PPC contains two major categories, i.e., Trusted Execution Environment (TEE) based and cryptography based. TEE is already widely available in major general purpose computing platforms such as Intel SGX and ARM TrustZone. While TEE techniques keep evolving, TEE based PPC systems have been deployed in some applications. On the other hand, Cryptography base mechanisms, e.g., secure Multi-Party Computation (MPC), Fully Homomorphic Encryption (FHE), etc., provide higher level of data privacy protection, and are actively developed in both academy and industry.
• TEE/Cryptography based computer architecture to support the growing and evolving PPC requirements, including hardware enhancements and software toolchains.
• High-performance and scalable cryptography processor/accelerator, and the integration with PPC platforms/frameworks.
• Performance optimization methods with hardware-software co-design.
Related Research Topics
• Architecture for PPC processor/accelerator, including instruction set design, processing engine micro-architecture and software toolchain.
• Domain-specific hardware-software co-design for PPC performance optimization.
• ASIC design with advanced packaging techniques, such as 3D PIM.
• Workload characterization for privacy preserving computing.