Analyzing Infrastructure as Code to Prevent Intra-update Sniping Vulnerabilities

Abstract

AbstractInfrastructure as Code is a new approach to computing infrastructure management that allows users to leverage tools such as version control, automatic deployments, and program analysis for infrastructure configurations. This approach allows for faster and more homogeneous configuration of a complete infrastructure. Infrastructure as Code languages, such as CloudFormation or TerraForm, use a declarative model so that users only need to describe the desired state of the infrastructure. However, in practice, these languages are not processed atomically. During an upgrade, the infrastructure goes through a series of intermediate states. We identify a security vulnerability that occurs during an upgrade even when the initial and final states of the infrastructure are secure, and we show that those vulnerability are possible in Amazon’s AWS and Google Cloud. We call such attacks intra-update sniping vulnerabilities. In order to mitigate this shortcoming, we present a technique that detects such vulnerabilities and pinpoints the root causes of insecure deployment migrations. We implement this technique in a tool, Häyhä, that uses dataflow graph analysis. We evaluate our tool on a set of open-source CloudFormation templates and find that it is scalable and could be used as part of a deployment workflow.