Collaborative Research: SaTC: TTP: Medium: Toward Complete, User-Friendly, and Trustworthy Confidential Computing with Gramine

点击翻译按钮获取中文摘要

This project aims to mature Gramine, previously known as Graphene or Graphene-SGX, from a successful, open-source (LGPL v3.0) research prototype into a robust, easy-to-use, and trustworthy building block for confidential computing applications. Confidential computing protects code and data in use, building upon recent hardware trusted execution environments (TEEs), such as Intel's SGX enclaves. Confidential computing is essential for cloud computing applications that use sensitive data, such as health applications, where one must balance the economic benefits of cloud computing with regulatory compliance or other security concerns. Gramine is a "lift-and-shift" framework for running unmodified applications in Intel SGX. The project addresses various barriers to adopting Gramine in production settings, including challenges in compatibility, usability, and security of Gramine. The project's novelties are creating a robust, general-purpose, open-source, Linux-compatibility layer that can easily migrate legacy application code from one platform to another---here, emerging confidential computing hardware. The project's broader significance and importance is to accelerate the study of confidential computing and other emerging computational platforms. Gramine is already a building block for over 100 academic papers and several commercial product prototypes. Gramine is publicly available at https://github.com/gramineproject/gramine.The project focuses on three aspects of Gramine development. First, the project expands the set of system interfaces and applications that work on Gramine, with the goal of supporting 90% of the applications installed on a representative Debian/Ubuntu system. Second, the project improves the Gramine user experience, by addressing deployment issues, simplifying configuration and policy decisions, better integrating with other software frameworks, and expanding the set of supported TEEs. Third, the project improves the trustworthiness of Gramine's code base with advanced testing and analysis, as well as rewriting critical code in Rust programming language.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.