Kubernetes cybersecurity
Master Thesis
Author
Μορφωνιός, Ιωάννης
Morfonios, Ioannis
Date
2023-04Advisor
Ξενάκης, ΧρήστοςXenakis, Christos
View/ Open
Keywords
Kubernetes ; Security ; Virtualization ; ContainersAbstract
Kubernetes is a widely used container orchestration tool that has greatly benefited the fast-paced development lifecycle. Its ability to manage thousands of containers and some of its key features, such as container lifecycle management, auto-healing, and auto-scaling, have made it a top choice for managing demanding workloads such as large scale web applications. However, just like any other software tool, Kubernetes has its own set of security weaknesses as well. Many vulnerabilities that affect its components have surfaced in the past, but a large percentage of successful security breaches in Kubernetes environments are not actually attributed to security flaws in the platform itself. As a matter of fact, the most common security threats that Kubernetes faces are created by misconfigurations. Due to the complexity of Kubernetes and the inexperience of many administrators, securing a Kubernetes cluster and its workloads is still a challenge for many companies. In this thesis, we will discuss the deployment and configuration of a Kubernetes cluster, as well as the subsequent evaluation of its security posture with the use of the kube-hunter and Kubescape vulnerability scanning tools. The goal is to evaluate many aspects of the cluster's security by using several scanning techniques, such as internal and external scanning, YAML file scanning, inspection of its components for vulnerabilities, and even estimate the overall security risk. To make the configuration more realistic, real misconfiguration scenarios will be introduced to the cluster, and some sample applications will be deployed as well. Afterward, some of the discovered security flaws will be exploited to demonstrate the amount of damage a malicious actor could cause to the cluster and its workloads. Finally, to effectively strengthen the cluster, we will analyze and mitigate any discovered vulnerabilities that are actively exposing it at risk, while ignoring any false positive warnings.