Graphene - a Library OS for Unmodified Applications
Open-Source community project driven by a core team of contributors.
A few words about Graphene
Applications programmed for one system often do not work on another. Graphene bridges this gap by hoisting application-facing code from the operating system (OS) kernel into a userspace library. Graphene uses a platform adaptation layer (PAL) that is easy to implement on a new host system. As long as a system implements the PAL interface, all of POSIX/Linux will follow.
Graphene is a library OS, similar to a unikernel. Compared to running a complete guest OS in a virtual machine (VM), Graphene is much lighter weight. Work is ongoing to integrate Graphene with Docker containers.
A particular use case for Graphene is Intel® Software Guard Extensions (Intel® SGX), where applications do not work out-of-the-box. Graphene solves this problem, with the added security benefits. Graphene can serve as a compatibility layer on other platforms.
Intel SGX integration made simple
Regular integration of Intel SGX
Integration of Intel SGX with Graphene
Applications can benefit from confidentiality and integrity guarantees of Intel SGX, but developers need to be very skilled for effective partitioning and code modification for Intel SGX environment.
Graphene runs unmodified applications inside Intel
SGX. It supports dynamically loaded libraries, runtime linking, multi-process abstractions, and file authentication. For additional security, Graphene performs cryptographic and semantic checks at untrusted host interface. Developers provide a manifest file to configure the application environment and isolation policies, Graphene automatically does the rest.
The commitment behind Graphene
Graphene started as a research project at Stony Brook University, led by Chia-Che Tsai and Don Porter. Over time, scientists at other universities and labs have contributed to Graphene to accelerate their research on emerging hardware platforms.
In 2015, Intel Labs recognized the potential for Graphene to be an open-source compatibility layer for Intel SGX, and has contributed to Graphene development since.
Golem and Invisible Things Lab (ITL) have identified similarly opportunity for Graphene to play a huge role in the decentralized ecosystem, where data integrity, confidentiality, and security are cornerstones to the robust development of infrastructure and applications. Driving Graphene and ensuring its usability is part of Golem's commitment.
Today, there is a strong team of developers and researchers from these companies working together with the founders of the project (now faculty at UNC and Texas A&M) to make sure it meets the highest quality standards with the easiness of integration. Graphene has a growing user and contributor community. It has the potential to become a standard in the Intel SGX world and can be adopted by a broad variety of use cases in a diverse technological landscape.
Past and future plans
2011
Graphene development starts in OSCAR LAB at Stony Brook University
arrow_right
2013
arrow_right
First paper is published at EuroSys and first public release
2015
Graphene for Intel SGX development starts in Intel Research Lab
arrow_right
Graphene for Intel SGX public release
2017
arrow_right
ITL/Golem get involved in the project
Graphene for Intel SGX paper is published at USENIX ATC
ITL/Golem work to deploy Graphene for Intel SGX in Golem Network and add more features
2018
arrow_right
First working Graphene integration - demo with Golem
The Graphene working group is established
2019
arrow_right
Building contributors' community
Planning stable release (v1.0) and Docker integration
Creating protected network & filesystem, supporting static binaries, adding Go and Java runtimes
2020
Supporting Windows as a host and other features