Skip to main content

2024 | Book

Secure IT Systems

28th Nordic Conference, NordSec 2023, Oslo, Norway, November 16–17, 2023, Proceedings


About this book

This book constitutes the proceedings of the 28th Nordic Conference, NordSec 2023, held in Oslo, Norway, during November 16–17, 2023.

The 18 full papers included in this volume were carefully reviewed and selected from 55 submissions. This volume focuses on a broad range of topics within IT security and privacy.

Table of Contents


Privacy and Data Protection

Analysis of a Consent Management Specification and Prototype Under the GDPR
Consent requests for the processing of personal information are ubiquitous for users of web services across the European Union (EU). However, their form and contents differ greatly, and often include deceptive design patterns (so-called dark patterns) meant to influence users’ choices.
In this paper, we provide the results of a research project to define a new specification that can be used to handle consent requests based on cookies in a standardized and GDPR-compliant manner. We define and evaluate a set of requirements for consent management systems and we illustrate the advantage of our proposed specification to the state of the art based on a prototype implementation and evaluation. Based on a small usability study, we found our solution to reduce the necessary interactions with respect to consenting, consent withdrawal, and consent configuration by far.
Jonas Palm, Meiko Jensen
No Place to Hide: Privacy Exposure in Anti-stalkerware Apps and Support Websites
Stalkerware is malicious software found in mobile devices that monitors and tracks a victim’s online and offline activity. This harmful technology has become a growing concern, jeopardizing the security and privacy of millions of victims and fostering stalking and Intimate Partner Violence (IPV). In response to this threat, various solutions have emerged, including anti-stalkerware apps that aim to prevent and detect the use of monitoring apps on a user’s device. Organizations dedicated to assisting IPV victims have also enhanced their online presence, offering improved support and easy access to resources and materials. Considering how these tools and support websites handle sensitive personal information of users, it is crucial to assess the privacy risks associated with them. In this paper, we conduct a privacy analysis on 25 anti-stalkerware apps and 323 websites to identify issues such as PII leaks, authentication problems and 3rd-party tracking. Our tests reveal that 14/25 apps and 210/323 websites share user information with 3rd-party services through trackers, cookies or session replay. We also identified 44 domains to which sensitive data is sent, along with 3 services collecting information submitted in forms through session replay.
Philippe Mangeard, Xiufen Yu, Mohammad Mannan, Amr Youssef
From Whistle to Echo: Data Leaks in Web-Based Whistleblowing Channels
Whistleblowing refers to reporting misconduct to responsible authorities. With accelerating digitalization and the European Union’s new whistleblower directive, large numbers of whistleblowing channels and company web pages that act as gateways to these services have been deployed. At the same time, on modern websites rife with third-party services such as web analytics, this development introduces privacy challenges. In the current study, we analyze websites of 15 Finnish companies and the whistleblowing services they employ in order to assess whether they inadvertently reveal identifying personal data to the employee’s company and third parties. Results indicate there is reason for serious concern about the privacy of whistleblowers who report wrongdoings online.
Esko Vuorinen, Panu Puhtila, Sampsa Rauti, Ville Leppänen

Cryptography, Protocols, Analysis

Small Private Key Attack Against a Family of RSA-Like Cryptosystems
Let \(N=pq\) be the product of two balanced prime numbers p and q. Elkamchouchi, Elshenawy and Shaban presented in 2002 an interesting RSA-like cryptosystem that uses the key equation \(ed - k (p^2-1)(q^2-1) = 1\), instead of the classical RSA key equation \(ed - k (p-1)(q-1) = 1\). The authors claimed that their scheme is more secure than RSA. Unfortunately, the common attacks developed against RSA can be adapted for Elkamchouchi et al.’s scheme. In this paper, we introduce a family of RSA-like encryption schemes that uses the key equation \(ed - k (p^n-1)(q^n-1) = 1\), where \(n>0\) is an integer. Then, we show that regardless of the choice of n, there exists an attack based on continued fractions that recovers the secret exponent.
Paul Cotan, George Teşeleanu
Fair Distributed Oblivious Polynomial Evaluation via Bitcoin Deposits: Compute-as-a-Service
Distributed oblivious polynomial evaluation (DOPE) is a special case of two-party computation where the sender party \(P_1\) holds a polynomial f(x) of degree k and the receiver party \(P_2\) has a value \(\alpha \). They wish to perform a secure computation with the help of n designated cloud servers such that \(P_2\) obtains the value \(f(\alpha )\) while the privacy of their inputs is maintained.
We present the first fair DOPE scheme using Bitcoin deposit transactions in the presence of n cloud servers where n is independent of the polynomial degree k. The fairness property ensures that an honest server gains the reward for conducting a computation service while a corrupt server has to pay some penalty amount to an honest party. Our protocol consists of two separate phases: setup and computation. The cloud servers are involved with \(P_1\) in the setup phase while \(P_2\) communicates with the servers in the computation phase which means that the actual computation can be implemented at any time after the setup phase. Any corrupt party/server can be detected using the non-interactive Pedersen’s commitment scheme. Our protocol preserves the security against an active adversary corrupting a coalition of \(P_1\) and at most t cloud servers in the setup phase and a coalition of up to t servers in the computation phase in the presence of honest majority of the servers. The communication complexity is bounded to O(kt) which is the same as that in the previous DOPE studies while the fairness feature is also achieved in our scheme.
Amirreza Hamidi, Hossein Ghodosi
Smart Noise Detection for Statistical Disclosure Attacks
While anonymization systems like mix networks can provide privacy to their users by, e.g., hiding their communication relationships, several traffic analysis attacks can deanonymize them. In this work, we examine Statistical Disclosure Attacks and introduce a new implementation called the Smart Noise Statistical Disclosure Attack. This attack can improve results by examining how often other users send together with the attacker’s target to better filter out the noise caused by them. We evaluate this attack by comparing it to previous variants in various simulations and thus show how it can improve upon them. Further, we demonstrate how other implementations can be improved by combing them with our approach to noise calculation. Finally, we critically review used evaluation metrics to determine their significance.
Marc Roßberger, Doğan Kesdoğan

Cyber Security

Cybersecurity Challenges and Smart Technology Adoption in Norwegian Livestock Farming
The importance of cybersecurity in agriculture has grown significantly due to the increasing use of technology, which brings about vulnerabilities in farm systems. This study investigates the technology usage and cyber attack susceptibility on Norwegian cow and pig farms while focusing on impacts to food production. Employing a phenomenological approach, we conducted 14 one-on-one interviews with cattle and pig farmers in Norway, complemented by two interviews with domain experts in widely-used milking robot brands for dairy farms. The findings indicate that dairy cow farms heavily rely on the milking robot for production, pig farms are highly dependent on feeding systems, while suckler cow farms have the lowest digital technology dependence. However, targeting a single farm is unlikely to cause significant consequences for the entire society. For threat actors aiming to disrupt food production on a national scale, the focus might shift towards suppliers of raw materials, machinery, data processors, and regulatory bodies for meat and dairy. Attacks at this level could have widespread implications for farms across the country, making it a critical area for future research and attention.
Karianne Kjønås, Gaute Wangen
Mean Value Analysis of Critical Attack Paths with Multiple Parameters
Graphical models like attack trees and attack graphs provide promising approaches to represent and analyze complex cyber infrastructures. One common analysis that graphical models are used for is to identify short, or other types of critical attack paths. In this paper, we consider attack graphs that are probabilistic, and the attack steps are characterized by multiple parameters, the probability of success, and the distribution of time to perform the attack step. We propose low-complexity solutions to find sets of critical paths according to flexible mean value-based utility functions. We demonstrate that the results are similar to the ones from Monte-Carlo simulations. Consequently, the utility function-based approach can substitute time-consuming simulations and can be a valuable component of dynamic defense strategies.
Rajendra Shivaji Patil, Viktoria Fodor, Mathias Ekstedt
RAMBO: Leaking Secrets from Air-Gap Computers by Spelling Covert Radio Signals from Computer RAM
Air-gapped systems are physically separated from external networks, including the Internet. This isolation is achieved by keeping the air-gap computers disconnected from wired or wireless networks, preventing direct or remote communication with other devices or networks. Air-gap measures may be used in sensitive environments where security and isolation are critical to prevent private and confidential information leakage.
In this paper, we present an attack allowing adversaries to leak information from air-gapped computers. We show that malware on a compromised computer can generate radio signals from memory buses (RAM). Using software-generated radio signals, malware can encode sensitive information such as files, images, keylogging, biometric information, and encryption keys. With software-defined radio (SDR) hardware, and a simple off-the-shelf antenna, an attacker can intercept transmitted raw radio signals from a distance. The signals can then be decoded and translated back into binary information. We discuss the design and implementation and present related work and evaluation results. This paper presents fast modification methods to leak data from air-gapped computers at 1000 bits per second. Finally, we propose countermeasures to mitigate this out-of-band air-gap threat.
Mordechai Guri
Legal Considerations on Gray Zone Operations – From a Norwegian Perspective
Threats in the digital domain is one of the, if not the most significant risks facing the individual, societies, and nation states worldwide. We are raising the question of whether the legal regulation of the digitally connected worldwide network is adequate to meet the challenges of harmful behavior to critical infrastructure. The general assumption among technical and security experts, as well as in the ongoing public debate, is that it is not. We look into the status of the current Nordic legislation, identify the main challenges, and point out future work.
Lars Berg, Kirsi Helkala, André Årnes

Aspects of Trust

Mobile App Distribution Transparency (MADT): Design and Evaluation of a System to Mitigate Necessary Trust in Mobile App Distribution Systems
Current mobile app distribution systems use (asymmetric) digital signatures to ensure integrity and authenticity for their apps. However, there are realistic threat models under which trust in such signatures is compromised. One example is an unconsciously leaked signing key that allows an attacker to distribute malicious updates to an existing app; other examples are intentional key sharing as well as insider attacks. Recent app store policy changes like Google Play Signing (and other similar OEM and free app stores like F-Droid) are a practically relevant case of intentional key sharing: such distribution systems take over key handling and create app signatures themselves, breaking up the previous end-to-end verifiable trust from developer to end-user device. This paper addresses these threats by proposing a system design that incorporates transparency logs and end-to-end verification in mobile app distribution systems to make unauthorized distribution attempts transparent and thus detectable. We analyzed the relevant security considerations with regard to our threat model as well as the security implications in the case where an attacker is able to compromise our proposed system. Finally, we implemented an open-source prototype extending F-Droid, which demonstrates practicability, feasibility, and performance of our proposed system.
Mario Lins, René Mayrhofer, Michael Roland, Alastair R. Beresford
DIPSAUCE: Efficient Private Stream Aggregation Without Trusted Parties
Private Stream Aggregation (PSA) schemes are efficient protocols for distributed data analytics. In a PSA scheme, a set of data producers can encrypt data for a central party so that it learns the sum of all encrypted values, but nothing about each individual value. Thus, a trusted aggregator is avoided. However, all known PSA schemes still require a trusted party for key generation. In this paper we propose the first PSA scheme that does not rely on a trusted party. We argue its security against static and mobile malicious adversaries, and show its efficiency by implementing both our scheme and the previous state-of-the-art on realistic IoT devices, and compare their performance. Our security and efficiency evaluations show that it is indeed possible to construct an efficient PSA scheme without a trusted central party. Surprisingly, our results also show that, as side effect, our method for distributing the setup procedure also makes the encryption procedure more efficient than the state of the art PSA schemes which rely on trusted parties.
Joakim Brorsson, Martin Gunnarsson
What is Your Information Worth? A Systematic Analysis of the Endowment Effect of Different Data Types
Various smartphone and web applications use personal information to estimate the user’s behaviour among others for targeted advertising and improvement of personalized applications. Often applications and web services offer only two choices, either accept their privacy policies or not use the services. Hereby, the general scenario is to pay applications and web services with personal data. As privacy policies are lengthy to read and not comprehensible, most users accept the terms and conditions without the awareness of potential consequences. Thus, most users are unaware of continuously being tracked by many applications installed on their smart devices or accept sharing personal data in exchange for using applications and services online. Therefore, this study attempts to shed some light on the willingness to pay for data protection when offered this option in a continuous data-sharing scenario, and the willingness to accept when offered the option to sell personal data to two different data requestors. The study (N = 500) is conducted via crowdsourcing and examines the monetary valuation of users with respect to different data-sharing scenarios and different data types to allow for a more fine-grained analysis of user preferences. Moreover, different influencing factors such as privacy concerns, awareness and intended behaviour are examined in relation to the user’s monetary valuation. The results show significant differences between willingness to pay and accept for ten different data types and the two sharing scenarios contributing to further empirical evidence for the endowment effect. However, the sharing scenarios seem to have not a big influence on willingness to pay but showed significant differences in willingness to accept. Furthermore, the privacy influencing factors seem to negatively correlate with willingness to pay and positively correlate with willingness to accept.
Vera Schmitt, Daniel Sivizaca Conde, Premtim Sahitaj, Sebastian Möller

Defenses and Forensics

Towards Generic Malware Unpacking: A Comprehensive Study on the Unpacking Behavior of Malicious Run-Time Packers
The presence of packing techniques in malicious software remains a significant obstacle in malware analysis. Consequently, numerous research efforts have emerged with the objective of developing a generic methodology to unpack malware. However, these unpacking methodologies often rely on assumptions about the capabilities of packers. These assumptions include factors such as the origin of memory sources, code-writing techniques used to fulfill packing capabilities, the number of packing layers used, the persistence of code within memory, and the clear distinction between packer and malware code. In our paper, we aim to advance the state-of-the-art by addressing these underlying assumptions associated with malware unpacking. Based on these assumptions, we formulate five research questions to be addressed in a study on the packer capabilities found in a real-world Windows malware and clinical data set consisting of off-the-shelf packers. The answers deduced from our study demonstrate that the majority of common generic unpacking methodologies in the literature show significant blind spots, with the notable exception of the Renovo methodology and its derivatives.
Thorsten Jenke, Elmar Padilla, Lilli Bruckschen
A Self-forming Community Approach for Intrusion Detection in Heterogeneous Networks
Detecting intrusions in modern network infrastructures is challenging because of the growing size and, along with it, the increasing complexity of structure. While several approaches try to cope with those challenges, few address problems arising from heterogeneity and changes within those infrastructures.
We present a self-forming community approach that integrates federated learning (FL) with distributed intrusion detection systems based on anomaly detection. It autonomously separates the anomaly detection models into communities at runtime with the goal of mutual information exchange using FL techniques to improve detection accuracy. Community formation is realized via the introduction of a similarity score between each pair of models, indicating which models would profit from aggregation. Through a re-evaluation of the similarity score during runtime, changes in the deployed infrastructure can be considered, and the communities adapted. Our experiments show our approach reported no false alarms when evaluated with a real-world dataset and an intrusion detection rate of up to 97%.
Philipp Eichhammer, Hans P. Reiser
To Possess or Not to Possess - WhatsApp for Android Revisited with a Focus on Stickers
WhatsApp stickers are a popular hybrid of images and emoticons that can contain user-created content. Stickers are mostly sent for legitimate reasons, but are also used to distribute illicit content such as Child Sexual Abuse Material (CSAM). As the process of creating stickers becomes easier for users from version to version, a digital forensic analysis is still lacking. Therefore, we present the first comprehensive digital forensic analysis of WhatsApp’s sticker handling on Android, with a special focus on the legal context, i.e. the definition of possession of illicit content. Our analysis is based on 40 scenarios that reflect the full lifecycle of community-created stickers. We show how the distribution channel of a sticker found on a device can be reconstructed, partially even when its traces have been removed from WhatsApp and are not visible through WhatsApp’s user interface. In addition, we show that Google Drive backups recover stickers, making device seizure dispensable; however, stickers can still be permanently deleted. Most importantly, we show that simply finding a sticker on a device is not sufficient to meet the requirements of the legal definition of possession. Therefore, prosecution for possession of a sticker requires additional evidence, which we provide.
Samantha Klier, Harald Baier

Machine Learning and Artificial Intelligence in Information Security

A More Secure Split: Enhancing the Security of Privacy-Preserving Split Learning
Split learning (SL) is a new collaborative learning technique that allows participants, e.g. a client and a server, to train machine learning models without the client sharing raw data. In this setting, the client initially applies its part of the machine learning model on the raw data to generate Activation Maps (AMs) and then sends them to the server to continue the training process. Previous works in the field demonstrated that reconstructing AMs could result in privacy leakage of client data. In addition to that, existing mitigation techniques that overcome the privacy leakage of SL prove to be significantly worse in terms of accuracy. In this paper, we improve upon previous works by constructing a protocol based on U-shaped SL that can operate on homomorphically encrypted data. More precisely, in our approach, the client applies homomorphic encryption on the AMs before sending them to the server, thus protecting user privacy. This is an important improvement that reduces privacy leakage in comparison to other SL-based works. Finally, our results show that, with the optimum set of parameters, training with HE data in the U-shaped SL setting only reduces accuracy by 2.65% compared to training on plaintext. In addition, raw training data privacy is preserved.
Tanveer Khan, Khoa Nguyen, Antonis Michalas
Force: Highly Efficient Four-Party Privacy-Preserving Machine Learning on GPU
Tremendous efforts have been made to improve the efficiency of secure Multi-Party Computation (MPC), which allows \(n \ge 2\) parties to jointly evaluate a target function without leaking their own private inputs. It has been confirmed by previous research that Three-Party Computation (3PC) and outsourcing computations to GPUs can lead to huge performance improvement of MPC in computationally intensive tasks such as Privacy-Preserving Machine Learning (PPML). A natural question to ask is whether super-linear performance gain is possible for a linear increase in resources. In this paper, we give an affirmative answer to this question. We propose \(\textsf{Force}\), an extremely efficient Four-Party Computation (4PC) system for PPML. To the best of our knowledge, each party in \(\textsf{Force}\) enjoys the least number of local computations, smallest graphic memory consumption and lowest data exchanges between parties. This is achieved by introducing a new sharing type \(\mathcal {X}\text {-}\textsf{share} \) along with MPC protocols in privacy-preserving training and inference that are semi-honest secure in the honest-majority setting. By comparing the results with state-of-the-art research, we showcase that \(\textsf{Force}\) is sound and extremely efficient, as it can improve the PPML performance by a factor of 2 to 38 compared with other latest GPU-based semi-honest secure systems, such as \(\textsf{Piranha}\) (including \(\textsf{SecureML}\), \(\textsf{Falcon}\), \(\textsf{FantasticFour}\)), \(\textsf{CryptGPU}\) and \(\textsf{CrypTen}\).
Tianxiang Dai, Li Duan, Yufan Jiang, Yong Li, Fei Mei, Yulian Sun
Secure IT Systems
Lothar Fritsch
Ismail Hassan
Ebenezer Paintsil
Copyright Year
Electronic ISBN
Print ISBN

Premium Partner