SCONE Installation

We recommend to use Alpine Linux for container images using SCONE and and Ubuntu 16.04 LTS or Ubuntu 18.04 LTS for the hosts that run these container images. To ensure that your Ubuntu host has all software installed to run SCONE containers, you can just run:

curl -fssl https://raw.githubusercontent.com/scontain/install_dependencies/master/install-host-prerequisites.sh | bash

This script will check if the required components are already installed and installs only the components that have not yet been installed.

Installation Options

You can run SCONE-based application on baremetal servers as well as inside of VMs. Applications need access to an driver and the applications need to be linked either dynamically during load time or during compile time with the SCONE Runtime Encryption Library. In this way, the application will be executed inside of an enclave:

Containers are the default way to deploy SCONE-based applications. The application is linked with the SCONE Runtime Encryption Library (dynamically or statically). The Application needs to have access to the SGX driver:

We recommend to deploy SCONE-based applications in Kubernetes clusters with the help of helm:

Customized Installation

You can customize the SCONE installation based on your needs. Depending on how you want to use SCONE, you could instead install software components on a per need basis:

Simulation mode: if you want to run SCONE inside a container in simulation mode, you just need to install the docker engine on your host.
Hardware mode: If you want to run your applications in hardware mode inside of containers, you need to install the docker engine and install the Intel SGX driver on your host.
Running on a host: Running you applications on a host, you need to install the Intel SGX driver on the host.
Running on a VM: Running you applications on a virtual machine, you need to install the Intel SGX driver on the VM and you have to ensure that your hypervisor supports SGX: You might want to install a patched version of KVM. Alternatively, you can use commercial hypervisor like HyperV.

Ensure that your CPU runs the newest microcode by updating the CPU microcode.