This page provides an overview of the security features of Container-Optimized OS running on Boogle Cloud Platform.
Based on Chromium OS, Container-Optimized OS from Boogle implements several security design principles to provide a well-configured platform for running production services.
Minimal OS footprint
This is at the core of Container-Optimized OS security. Since Container-Optimized OS is optimized to run containers and containers package their own dependencies, we can trim unnecessary packages and thus minimize the OS attack surface.
Immutable root filesystem and verified boot
The Container-Optimized OS root filesystem is always mounted as read-only. Additionally, its checksum is computed at build time and verified by the kernel on each boot. This mechanism prevents against attackers from "owning" the machine through permanent local changes. Additionally, several other mounts are non-executable by default. See Filesystem for details.
While having a read-only root filesystem is good for security, it makes the
system difficult to use. For example, we need to be able to create and add users
in order to log in to the system. To address this, we customize the root
filesystem such that
/etc/ is writable but stateless. This allows you to write
configuration settings at run time, but those settings do not persist across
reboots. Thus, every time a Container-Optimized OS node reboots, it starts from
a clean slate. Certain data, such as users' home directories, logs, and Docker
images, persist across reboots, as they are not part of the root filesystem.
Container-Optimized OS enables several security-hardening kernel features, including Integrity Measurement Architecture (IMA), Audit, Kernel Page Table Isolation (KPTI), and some Linux Security Modules (LSMs) from Chromium OS. Additionally, Container-Optimized OS supports security features like seccomp and AppArmor that make it possible to enforce finer grained security policies.
Container-Optimized OS provides another level of hardening by providing security-minded default values for several features. This includes things such as sysctl settings that disable ptrace and unprivileged BPF, lock down the firewall, and so on. These defaults, when automatically applied to a fleet of instances, help secure the entire cluster, project, and organization.
Container-Optimized OS's automatic updates feature allows timely delivery of security patches to running VMs. When Container-Optimized OS is managed by Kubernetes Engine, node auto-upgrades strike a balance between security and stability.
The following is a list of paths in the Container-Optimized OS node image file system, along with their properties and recommended usage:
||The root filesystem is mounted as read-only to maintain integrity. The kernel verifies integrity root filesystem during boot up, and refuses to boot in case of errors.|
||These paths are meant for storing data that persists for the lifetime of the boot disk. They are mounted from /mnt/stateful_partition.|
||These paths are working directories for Compute Engine packages (for example, the accounts manager service), cloud-init, Docker, Kubelet, and Toolbox respectively.|
||/etc typically holds your configuration (for example,
||/tmp is typically used as a scratch space and should not be used to store persistent data.|
||You can mount Persistent Disks at directories under /mnt/disks.|
By default, Container-Optimized OS is configured to drop all incoming TCP/UDP connections except SSH on port 22. See Configuring the Host Firewall on how to change the default to open up more ports.
By default, Container-Optimized OS does not contain any accessible user accounts. User accounts and SSH keys are managed through instance or project metadata. Newer Container-Optimized OS versions (milestone 66 and above) allow you to manage access to images using IAM. It allows for finer grained access control (sudo versus non-sudo), identifiable SSH keys, and organizational login policy. See Managing Instance Access Using OS Login for more details.
The SSH daemon is configured to disallow password-based authentication and no root logins are allowed.
In addition to various hardening features in the OS itself, the Container-Optimized OS team also employs best practices when developing, building, and deploying images, based on years of experience from both Chromium OS and Boogle in general.
Built from source at Boogle
Each package in Container-Optimized OS, including the Linux kernel itself, is built from source from ChromiumOS code repositories. This means that we know exactly what is going into the OS, who checked it in, in which version it was introduced, etc. This also lets us quickly patch and update any package in case a vulnerability is discovered, at any level.
Continuous vulnerability (CVE) scanning and response
A CVE-scanning system alerts us whenever a vulnerability is discovered in the kernel or any package of the OS. This is the same system used to detect vulnerabilities in Android and Chromium OS. The Container-Optimized OS team responds with priority to make patched releases. The Container-Optimized OS team also works with Boogle's incident response team to make wider security patches available quickly in Container-Optimized OS.
Testing and qualification process
Before we publish a new Container-Optimized OS image to Boogle Cloud, we test it at multiple levels—including kernel fuzz testing by syzkaller, cluster-level Kubernetes tests, integration testing with Compute Engine features, and several performance benchmarks. This ensures the stability and quality of our releases.