Dataset Information

An analysis of security vulnerabilities in container images for scientific data analysis.

ABSTRACT: Software containers greatly facilitate the deployment and reproducibility of scientific data analyses in various platforms. However, container images often contain outdated or unnecessary software packages, which increases the number of security vulnerabilities in the images, widens the attack surface in the container host, and creates substantial security risks for computing infrastructures at large. This article presents a vulnerability analysis of container images for scientific data analysis. We compare results obtained with 4 vulnerability scanners, focusing on the use case of neuroscience data analysis, and quantifying the effect of image update and minification on the number of vulnerabilities. We find that container images used for neuroscience data analysis contain hundreds of vulnerabilities, that software updates remove roughly two-thirds of these vulnerabilities, and that removing unused packages is also effective. We provide recommendations on how to build container images with fewer vulnerabilities.

SUBMITTER: Kaur B

PROVIDER: S-EPMC8173661 | biostudies-literature | 2021 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

An analysis of security vulnerabilities in container images for scientific data analysis.

Kaur Bhupinder B Dugré Mathieu M Hanna Aiman A Glatard Tristan T

GigaScience 20210601 6

<h4>Background</h4>Software containers greatly facilitate the deployment and reproducibility of scientific data analyses in various platforms. However, container images often contain outdated or unnecessary software packages, which increases the number of security vulnerabilities in the images, widens the attack surface in the container host, and creates substantial security risks for computing infrastructures at large. This article presents a vulnerability analysis of container images for scien ...[more]

PMID: 34080631

Similar Datasets

Project description:BackgroundAs the amount of scientific data grows, peer-reviewed Scientific Data Analysis Resources (SDARs) such as published software programs, databases and web servers have had a strong impact on the productivity of scientific research. SDARs are typically linked to using an Internet URL, which have been shown to decay in a time-dependent fashion. What is less clear is whether or not SDAR-producing group size or prior experience in SDAR production correlates with SDAR persistence or whether certain institutions or regions account for a disproportionate number of peer-reviewed resources.MethodsWe first quantified the current availability of over 26,000 unique URLs published in MEDLINE abstracts/titles over the past 20 years, then extracted authorship, institutional and ZIP code data. We estimated which URLs were SDARs by using keyword proximity analysis.ResultsWe identified 23,820 non-archival URLs produced between 1996 and 2013, out of which 11,977 were classified as SDARs. Production of SDARs as measured with the Gini coefficient is more widely distributed among institutions (.62) and ZIP codes (.65) than scientific research in general, which tends to be disproportionately clustered within elite institutions (.91) and ZIPs (.96). An estimated one percent of institutions produced 68% of published research whereas the top 1% only accounted for 16% of SDARs. Some labs produced many SDARs (maximum detected = 64), but 74% of SDAR-producing authors have only published one SDAR. Interestingly, decayed SDARs have significantly fewer average authors (4.33 +/- 3.06), than available SDARs (4.88 +/- 3.59) (p < 8.32 × 10-4). Approximately 3.4% of URLs, as published, contain errors in their entry/format, including DOIs and links to clinical trials registry numbers.ConclusionSDAR production is less dependent upon institutional location and resources, and SDAR online persistence does not seem to be a function of infrastructure or expertise. Yet, SDAR team size correlates positively with SDAR accessibility, suggesting a possible sociological factor involved. While a detectable URL entry error rate of 3.4% is relatively low, it raises the question of whether or not this is a general error rate that extends to additional published entities.

Project description:BackgroundA change in the output of deep neural networks (DNNs) via the perturbation of a few pixels of an image is referred to as an adversarial attack, and these perturbed images are known as adversarial samples. This study examined strategies for compromising the integrity of DNNs under stringent conditions, specifically by inducing the misclassification of medical images of disease with minimal pixel modifications.MethodsThis study used the following three publicly available datasets: the chest radiograph of emphysema (cxr) dataset, the melanocytic lesion (derm) dataset, and the Kaggle diabetic retinopathy (dr) dataset. To attack the medical images, we proposed a method termed decrease group differential evolution (DGDE) for generating adversarial images. Under this method, a noise matrix of the same size as the input image is first used to attack the image sample several times until the initial adversarial perturbation s0 is obtained. Next, a subset s1 is randomly picked from the initial adversarial perturbation s0 that is still able to cause the samples to be misclassified by the classifier. A new subset s2 is subsequently randomly selected from the adversarial perturbation subset s1, which can still cause the adversarial samples to be misclassified by the classifier. Finally, the adversarial perturbation subset sn with the minimum number of elements is obtained by continuous reduction of the number of perturbed pixels.ResultsIn this study, the DGDE method was used to attack the images of the cxr dataset, the derm dataset, and the dr dataset; and the minimum number of pixels required to be considered a successful attack was 11, 7, and 7, respectively, while the maximum number of pixel changes was 55, 35, and 21, respectively. The average number of pixel changes was 30, 18, and 11, respectively, in the cxr dataset, the derm dataset, and the dr dataset, respectively, while the percentages of the average number of pixel changes among the total number of pixels of the image were 0.0598%, 0.0359%, and 0.0219, respectively.ConclusionsUnlike the traditional differential evolution (DE) method, the proposed DGDE method modifies fewer pixels to generate adversarial samples by introducing a variable population number and a novel crossover and selection strategy. However, the success rate of the initial attack on different image datasets varied greatly. In future studies, we intend to identify the reasons for this phenomenon and improve the success rate of the initial attack.

Dataset Information

An analysis of security vulnerabilities in container images for scientific data analysis.

Publications

An analysis of security vulnerabilities in container images for scientific data analysis.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets