A Linux-compatible Library OS for Multi-Process Applications
Graphene Library OS is a project which provides lightweight guest OSes with support for Linux multi-process applications. Graphene can run applications in an isolated environment with virtualization benefits such as guest customization, platform independence, and migration, which is comparable to other virtual machines.
Graphene Library OS supports native, unmodified Linux applications on any platform. Currently, Graphene Library OS is successfully ported to Linux, FreeBSD and Intel SGX enclaves upon Linux platforms.
With the Intel:registered: SGX support, Graphene Library OS can secure a critical application in a hardware encrypted memory region. Graphene Library OS can protect applications against a malicious system stack with minimal porting effort.
Graphene Library OS is a work published in Eurosys 2014. For more information. see the paper: Tsai, et al, "Cooperation and Security Isolation of Library OSes for Multi-Process Applications", Eurosys 2014.
Graphene Library OS is consist of five parts:
Graphene Library OS currently only works on x86_64 architecture.
Graphene Library OS is tested to be compiling and running on Ubuntu 14.04/16.04 (both server and desktop version), along with Linux kernel 3.5/3.14/4.4. We recommand to build and install Graphene with the same host platform. Other distributions of 64-bit Linux can potentially, but the result is not guaranteed. If you find Graphene not working on other distributions, please contact us with a detailed bug report.
The following packages are required for building Graphene: (can be installed with 'apt-get install')
The following packages are also required for building Graphene for SGX (can be installed with 'apt-get install'):
To build the system, simply run the following commands in the root of the source tree:
git submodule update --init make
Each part of Graphene can be built separately in the subdirectories.
To build Graphene library OS with debug symbols, run "make DEBUG=1" instead of "make".
** Note: this step is optional. **
** Note: for building with Intel:registered: SGX support, skip this step, go to section 2.2 **
** Disclaimer: this feature is experimental and may contain bugs. Please do no use in production system before further assessment.
To enable sandboxing, a customized Linux kernel is needed. Note that this feature is optional and completely unnecessary for running on SGX. To build the Graphene Linux kernel, do the following steps:
cd Pal/linux-3.19 make menuconfig make make install (Add Graphene kernel as a boot option by commands like "update-grub") (reboot and choose the Graphene kernel)
Please note that the building process may pause before building the Linux kernel, because it requires you to provide a sensible configuration file (.config). The Graphene kernel requires the following options to be enabled in the configuration:
For more details about the building and installation, see the Graphene github Wiki page: https://github.com/oscarlab/graphene/wiki.
(1) Generating signing keys A 3072-bit RSA private key (PEM format) is required for signing the enclaves. If you don't have a private key, create it with the following command:
openssl genrsa -3 -out enclave-key.pem 3072
You could either put the generated enclave key to the default path, 'host/Linux-SGX/signer/enclave-key.pem', or specify the key through environment variable 'SGX_ENCLAVE_KEY' when building Graphene with SGX support.
After signing the enclaves, users may ship the application files with the built Graphene Library OS, along with a SGX-specific manifest (.manifest.sgx files) and the signatures, to the SGX-enabled hosts.
(2) Installing Intel SGX SDK and driver The Intel SGX Linux SDK is required for running Graphene Library OS. Download and install from the official Intel github repositories:
A Linux driver must be installed before runing Graphene Library OS in enclaves. Simply run the following command to build the driver:
** Please make sure the GCC version is either 4 or 5 **
cd Pal/src/host/Linux-SGX/sgx-driver make (The console will be prompted to ask for the path of Intel SGX driver code) sudo ./load.sh
__2.1.2 Building Graphene-SGX
To build Graphene Library OS with Intel SGX support, in the root directory of Graphene repo, run following command:
To build with debug symbols, run the command:
make SGX=1 DEBUG=1
Using "make SGX=1" in the test or regression directory will automatically generate the enclave signatures (.sig files).
__2.1.3 Run Built-in Examples in Graphene-SGX
There are a few built-in examples under LibOS/shim/test/. The "native" folder includes a rich set of C programs and "apps" folder includes a few tested applications, such as GCC, Python, and Apache.
(1) Build and run a Hello World program with Graphene on SGX
go to LibOS/shim/test/native, build the enclaves via command:
The command will build enclaves for all the programs in the folder
Generate the token from aesmd service, via command:
Run Hello World program with Graphene on SGX:
SGX=1 ./pal_loader helloworld or ./pal_loader SGX helloworld
(2) Build and run python helloworld script in Graphene on SGX
go to LibOS/shim/test/apps/python, build the enclave:
Run python helloworld with Graphene-SGX via:
SGX=1 ./python.manifest.sgx scripts/helloworld.py
Graphene library OS uses PAL (libpal.so) as a loader to bootstrap an application in the library OS. To start Graphene, PAL (libpal.so) will have to be run as an executable, with the name of the program, and a "manifest file" given from the command line. Graphene provides three options for specifying the programs and manifest files:
option 1: (automatic manifest)
[PATH TO Runtime]/pal_loader [PROGRAM] [ARGUMENTS]... (Manifest file: "[PROGRAM].manifest" or "manifest")
option 2: (given manifest)
[PATH TO Runtime]/pal_loader [MANIFEST] [ARGUMENTS]...
option 3: (manifest as a script)
[PATH TO MANIFEST]/[MANIFEST] [ARGUMENTS]... (Manifest must have "#![PATH_TO_PAL]/libpal.so" as the first line)
Using "libpal.so" as loader to start Graphene will not attach the applications to the Graphene reference monitor. Tha applications will have better performance, but no strong security isolation. To attach the applications to the Graphene reference monitor, Graphene must be started with the PAL reference monitor loader (libpal_sec.so). Graphene provides three options for specifying the programs and manifest files to the loader:
option 4: (automatic manifest - with reference monitor)
SEC=1 [PATH TO Runtime]/pal_loader [PROGRAM] [ARGUMENTS]... (Manifest file: "[PROGRAM].manifest" or "manifest")
option 5: (given manifest - with reference monitor)
SEC=1 [PATH TO Pal/src]/pal_loader [MANIFEST] [ARGUMENTS]...
option 6: (manifest as a script - with reference monitor)
SEC=1 [PATH TO MANIFEST]/[MANIFEST] [ARGUMENTS]... (Manifest must have "#![PATH TO Pal/src]/pal_sec" as the first line)
Although manifest files are optional for Graphene, running an application usually requires some minimal configuration in its manifest file. A sensible manifest file will include paths to the library OS and GNU library C, environment variables such as LD_LIBRARY_PATH, file systems to be mounted, and isolation rules to be enforced in the reference monitor.
Here is an example of manifest files:
loader.preload = file:LibOS/shim/src/libsysdb.so loader.env.LDL_LIBRAY_PATH = /lib fs.mount.glibc.type = chroot fs.mount.glibc.path = /lib fs.mount.glibc.uri = file:LibOS/build
More examples can be found in the test directories (LibOS/shim/test). We have also tested several commercial applications such as GCC, Bash and Apache, and the manifest files that bootstrap them in Graphene are provided in the individual directories.
For more information and the detail of the manifest syntax, see the Graphene github Wiki page: https://github.com/oscarlab/graphene/wiki.
For any questions or bug reports, please contact us:
or post an issue on our github repository: