The goal of this research is to create a methodology that evaluates the security posture of container images and helps improve container security. This was done by first searching for any guidelines or standards that focus on container images and security. After finding none, I decided to create an evaluative methodology. The methodology is composed of actions that users should take to evaluate the security of a container image. The methodology was created through in-depth research on container images and the build instructions used to create them and is referred to as the Security Evaluation Methodology for Container Images. The entire Methodology was reviewed by experts in containers, information technology, and security; updated based on their feedback; and then reviewed again for further feedback. Four of the most popular container images—nginx, redis, mbabineau/cfn-bootstrap, and google/cadvisor—were evaluated using the Methodology. The evaluation revealed security issues in each image and provided direction on how to resolve each issue. Based on the positive feedback of experts and the performance of the Methodology, I propose that the Methodology be used to evaluate all container images, as it provides valuable security insights about, and suggestions for, an image.

1. INTRODUCTION Lung cancer is the most common malignancy in the world and accounts for 1.09 million new cases with 972 000 deaths per year. In Lithuania too, lung cancer is the most commonly diagnosed cancer in men with nearly 1500 new cases each year. Lung cancer is five times more common in men than in women. At present, most patients who receive an initial diagnosis of lung cancer have advanced stage disease (stage IV – 30.2%), making cure with currently available therapies unlikely. The main prognostic information with regard to survival is associated with the biological characteristics of the primary tumour (histological subtype, aggressiveness, differentiation, etc.), the extent of spread to regional or distant lymph nodes or to the other structures, and the operability of the patient (age, function of residual lung, co-morbidity). Because the outcome is associated with the histological subtype and stage of the lung cancer at the diagnosis, there has been persistent interest in designing and testing various radiological methods for early detection of lung cancer. Chest radiography, computed tomography (CT), magnetic resonance imaging (MRI) can not only identify specific features in lung nodules, but add important information about the localisation, size and extent of the primary tumour (T), detect invasion of major mediastinal structures and chest wall, and locoregional (N) and distal spread of the tumour. Although theoretically features such as nodal shape... [to full text]