Container security has been one of the hottest topics in the industry as of late, but what does “container security” really mean. Does it mean ensuring your images don’t have vulnerabilities in them? It probably means something a little different depending on who you ask.
We think true container security involves deploying a much more comprehensive approach including securing the entire container stack from top to bottom. In this blog post, we’ll explain the three pillars we believe must be covered to ensure container security and give examples of ways to ensure your organization has covered its bases.
Securing the Host Operating System
Docker security starts at the host layer and is only as strong as this layer. If attackers are able to compromise the host OS, they could potentially compromise all processes on this OS, including Docker. For the vast majority of Docker users, the host operating systems in a Linux distribution, so best practices for secure OS infrastructure should follow. The host operating system should be kept patched and updated. One example of this is to minimize the attack surface of the host OS. For the most secure infrastructure, the base OS should be specifically designed to run Docker only, no other processes that could be compromised.
Securing Docker Images
Incorporating the appropriate mechanisms to conduct static analysis on your container images gives insight into any potential vulnerable OS and non-OS packages. As discussed in previous posts, Anchore gives you the ability to control whether or not you would like to promote non-compliant images into trusted registries through policy checks within a secure container build pipeline. This step is important because vulnerable images that make their way into production environments pose significant threats to those environments and are also costly. Within these images, focus on the security of the applications that will be running, although other checks are important, your application must be secure as well.
There’s more information in this article on the shifting left paradigm.
Securing Container Runtime
It is important to set up tooling to monitor the containers themselves that are running. If new vulnerabilities get published that are impactful to a particular container, the appropriate alerting mechanisms need to be in place to quickly stop and replace the vulnerable container. Put simply, container infrastructure should be immutable.
The first piece of securing the container runtime is securing the registries where the images reside. For organizations, it is considered best practice to only pull and run images from trusted container registries. As an added layer of security, only trusted and signed images should be promoted into production registries. Vulnerable, non-compliant images should not reside in container registries where images are staged for production deployments.
The Docker Engine hosts and runs built container images that are pulled from registries. Namespaces and Control Groups are two aspects of Docker Engine security that need to be taken into consideration.
Namespaces provide the first and most straightforward form of isolation: processes running within a container cannot see, and even less effect, processes running in another container, or in the host system.
- Namespaces should always be activated.
- Control Groups implement resource accounting and limiting.
- Resource limits should always be set for containers. This way a single container does not hog all resources and bring down the system.
Additionally, the Docker daemon itself should only be controlled by trusted users. Currently, root privileges are required to run Docker commands, and caution should be executed when making changes to the Docker group.
Sysdig Falco is an example of a tool that will continuously monitor and detect anomalous activity in applications and containers.
Containerized applications and environments present new security concerns, but fundamentally basic concepts for host and application security can play critical roles in establishing a stronger container security posture. When discussing container adoption, container image scanning, runtime monitoring, and host security are great places to start, however security best practices like scanning application source code for vulnerabilities/coding errors both in open source and propriety, along with adopting a DevSecOps culture both contribute to a more secure container environment.
Read more in part two of our Docker series.