Containerized applications are becoming more popular with each passing year. A reason for this rise in popularity could be the pivotal role they play in Continuous Delivery by enabling fast and automated deployment of software services.
Security still remains a major concern mainly because of the way container images are being used. In the world of VMs, infra/security team used to validate the OS images and installed packages for vulnerabilities. But with the adoption of containers, developers are building their own container images. Images are rarely built from scratch. They are typically built on some base image, which is itself built on top of other base images. When a developer builds a container image, he typically grabs a base image and other layers from public third party sources. These images and libraries may contain obsolete or vulnerable packages, thereby putting your infrastructure at risk. An added complexity is that many existing vulnerability-scanning tools may not work with containers, nor do they support container delivery workflows including registries and CI/CD pipelines. In addition, you can’t simply scan for vulnerabilities – you must scan, manage vulnerability fixes and enforce vulnerability-based policies.
The Container Security Problem
The table below shows the number of vulnerabilities found in the images available on dockerhub. Note that (as of April 2016) the worst offending community images contained almost 1,800 vulnerabilities! Official images were much better, but still contained 392 vulnerabilities in the worst case.
If we look at the distribution of vulnerability severities, we see that pretty much all of them are high severity, for both official and community images. What we’re not told is the underlying distribution of vulnerability severities in the CVE database, so this could simply be a reflection of that distribution.
Over 80% of the latest versions of official images contained at least one high severity vulnerability!
- There are so many docker images readily available on dockerhub - are you sure the ones you are using are safe?
- Do you know where your containers come from?
- Are your developers downloading container images and libraries from unknown and potentially harmful sources?
- Do the containers use third party library code that is obsolete or vulnerable?
In this blog post, I will explain some of the solutions available which can help with these challenges. Solutions like ‘Docker scanning services‘, ‘Twistlock Trust’ and an open-source solution ‘Clair‘ from Coreos.com which can help in scanning and fixing vulnerability problems making your container images secure.
Clair is an open source project for the static analysis of vulnerabilities in application containers. It works as an API that analyzes every container layer to find known vulnerabilities using existing package managers such as Debian (dpkg), Ubuntu (dpkg), CentOS (rpm). It also can be used from the command line. It provides a list of vulnerabilities that threaten a container, and can notify users when new vulnerabilities that affect existing containers become known. In regular intervals, Clair ingests vulnerability metadata from a configured set of sources and stores it in the database. Clients use the Clair API to index their container images; this parses a list of installed source packages and stores them in the database. Clients use the Clair API to query the database; correlating data in real time, rather than a cached result that needs re-scanning.
Clair identifies security issues that developers introduce in their container images. The vanilla process for using Clair is as follows:
- A developer programmatically submits their container image to Clair
- Clair analyzes the image, looking for security vulnerabilities
- Clair returns a detailed report of security vulnerabilities present in the image
- Developer acts based on the report
How to use Clair
Docker is required to follow along with this demonstration. Once Docker is installed, use the Dockerfile below to create an Ubuntu image that contains a version of SSL that is susceptible to Heartbleed attacks.
Build the image using below command:
After creating the insecure Docker image, the next step is to download and install Clair from here. The installation choice used for this demonstration was the Docker Compose solution. Once Clair is installed, it can be used via querying its API or through the clairctl command line tool. Submit the insecure Docker image created above to Clair for analysis and it will catch the Heartbleed vulnerability.
You can view the detailed report here.
Docker Security Scanning
Docker Cloud and Docker Hub can scan images in private repositories to verify that they are free from known security vulnerabilities or exposures, and report the results of the scan for each image tag. Docker Security Scanning is available as an add-on to Docker hosted private repositories on both Docker Cloud and Docker Hub.
Security scanning is enabled on a per-repository basis and is only available for private repositories. Scans run each time a build pushes a new image to your private repository. They also run when you add a new image or tag. The scan traverses each layer of the image, identifies the software components in each layer, and indexes the SHA of each component.
The scan compares the SHA of each component against the Common Vulnerabilities and Exposures (CVE®) database. The CVE is a “dictionary” of known information security vulnerabilities. When the CVE database is updated, the service reviews the indexed components for any that match the new vulnerability. If the new vulnerability is detected in an image, the service sends an email alert to the maintainers of the image.
A single component can contain multiple vulnerabilities or exposures and Docker Security Scanning reports on each one. You can click an individual vulnerability report from the scan results and navigate to the specific CVE report data to learn more about it.
Twistlock is a rule-based access control policy system for Docker and Kubernetes containers. Twistlock is able to be fully integrated within Docker, with out-of-the-box security policies that are ready to use.
Security policies can set the conditions for users to, say, create new containers but not delete them; or, they can launch containers but aren’t allowed to push code to them. Twistlock features the same policy management rules as those on Kubernetes, wherein a user can modify management policies but cannot delete them.
Twistlock also handles image scanning. Users can scan an entire container image, including any packaged Docker application. Twistlock has done its due-diligence in this area, correlating with Red Hat and Mirantis to ensure no container is left vulnerable while a scan is running.
Twistlock also deals with image scanning of containers within the registries themselves. In runtime environments, Twistlock features a Docker proxy running on the same server with an application’s other containers. This is essentially traffic filtering, whereupon the application container calling the Docker daemon is then re-routed through Twistlock. This approach enforces access control, allowing for safer configuration where no containers are set to run as root. It’s also able to SSH into an instance, for example. In order to delve into these layers of security, Twistlock enforces the policy at runtime.
When new code is written in images, it is then integrated into the Twistlock API to push an event, whereupon the new image is deposited into the registry along with its unique IDs. It is then pulled out by Twistlock and scanned to ensure it complies with the set security policies in place. Twistlock deposits the scan result into the CI process so that developers can view the result for debugging purposes.
Integrating these vulnerability scanning tools into your CI/CD Pipeline:
These tools becomes more interesting paired with a CI server like Jenkins, TravisCI, etc. Given proper configuration, process becomes:
- A developer submits application code to source control
- Source control triggers a Jenkins build
- Jenkins builds the software containers necessary for the application
- Jenkins submits the container images to vulnerability scanning tool
- Tool identifies security vulnerabilities in the container
- Jenkins receives the security report, identifies a high vulnerability in the report, and stops the build
There are many solutions like ‘Docker scanning services’, ‘Twistlock Trust’, ‘Clair‘, etc to secure your containers. It’s critical for organizations to adopt such tools in their CI/CD pipelines. But this itself is not going to make containers secure. There are lot of guidelines available in the CIS Benchmark for containers like tuning kernel parameters, setting proper network configurations for inter-container connectivity, securing access to host level directories and others. I will cover these items in the next set of blogs. Stay tuned!