Contributions published at Communication Systems Group (Burkhard Stiller)

null

null

Businesses were moving during the past decades to-ward full digital models, which made companies face new threatsand cyberattacks affecting their services and, consequently, theirprofits. To avoid negative impacts, companies’ investments incybersecurity are increasing considerably. However, Small andMedium-sized Enterprises (SMEs) operate on small budgets,minimal technical expertise, and few personnel to address cy-bersecurity threats. In order to address such challenges, it isessential to promote novel approaches that can intuitively presentcybersecurity-related technical information.This paper introduces SecBot, a cybersecurity-driven conver-sational agent (i.e., chatbot) for the support of cybersecurityplanning and management. SecBot applies concepts of neuralnetworks and Natural Language Processing (NLP), to interactand extract information from a conversation. SecBot can(a)identify cyberattacks based on related symptoms,(b)indicatesolutions and configurations according to business demands,and(c)provide insightful information for the decision on cy-bersecurity investments and risks. A formal description hadbeen developed to describe states, transitions, a language, anda Proof-of-Concept (PoC) implementation. A case study and aperformance evaluation were conducted to provide evidence ofthe proposed solution’s feasibility and accuracy

null

Voting is one of the cornerstones of any democracy. With Electronic Voting (E-Voting), the Swiss confederation wants to advance Switzerland's voting and electoral system to the digital age. One of the main targets of its e-government initiative is the development of a Remote E-Voting (REV) system, with which voters can easily vote from their own devices via the internet. However, this electronic channel, which is proposed as an additional option to in-person and postal voting, poses unique challenges. Digitizing the human right of ballot secrecy to REV systems is not straightforward. Indeed, enabling the verifiability required in order for any voter to check that their vote was counted, is in direct opposition to privacy. Blockchains (BC) promise to bring many of the properties we look for in a voting system: transparency, resilience, tamper-proof, and decentralization. Switzerland, with its federalistic structure, provides a perfect political analogy for implementing a REV system, where trust is distributed among multiple authorities. Canton and municipalities can collaborate to establish a network running the voting software, removing the need to trust any single entity with the correctness of the vote { it would provide transparency to voters and resilience against voting suppression. This thesis conducts a security analysis of the Provotum REV system in order to assess risks and threats to the current design. Based on the results of the security audit, and improved design for a fully decentralized voting system is proposed. The new architecture brings an advanced notion of voter privacy to BC-based voting: Receipt-Freeness - a measure against vote selling. By introducing a new authority called Randomizer, the voter can not prove to a vote buyer how she voted while maintaining the ability to verify the correctness of her vote. Finally, the new scheme is analyzed proving that Provotum 3.0 achieves Receipt-Freeness while maintaining Ballot-Secrecy and verifiability. The evaluation includes a scalability analysis, showing any cryptographic material can be generated and verified in linear time, proving that the system can scale to nation-wide elections and votes.

The goal of this thesis is to implement TinyIPFIX, which is an application layer protocol to send data in a Wireless Sensor Network (WSN). Devices used in a WSN are often battery powered, therefore, it is important that the protocol used to send data is very efficient. The main principle of TinyIPFIX is to send data and metadata in separate messages, such that the metadata messages can be sent less often than the data messages, which decreases overhead and saves energy. The implementation is performed for the Espressif ESP32-WROOM-32D module, which becomes popular, while it is cheap and more powerful than older devices. The implementation is provided using the MicroPython programming language, which implies that the code is more readable and simple in comparison to a lower level programming language. TinyIPFIX has already been demonstrated on older devices using a low level language; the goal of this thesis is to bring TinyIPFIX to the next generation of devices and provide it as extensible software.

Caused by digitization and the Covid-19 pandemic courses are transformed to remote classes. In usual exams the presence of students is verified and validated by the teacher, who checks on every student with a list of students and with the UZH-Card. Digitally this is way more difficult. In this thesis a prototype is developed to tackle this problem. In a simplified version it models a way to check the identity of a student in a online class. Therefore the information on the UZH-card is read and compared to a stored list of student information. Also the face of the student in front of the computer is compared to stored values. The matching information from both comparisons are again compared to see if the person in front of the camera is the person, who they claim to be (the person on the UZH card). Obviously the prototype is simplified with dummy databases and access to the own webcam instead of other webcams. Furthermore, it is specified, that this prototype is only available to certain people with a teaching function at the University, because the system works with personal data, which should not be accessible to everyone. In the end there is an evaluation on the precision and two test cases are presented.

null

null

null

Long Range (LoRa) defines a popular modulation scheme based on the chirp spread spectrum technique. It is used in Low Power Wide Area Networks (LP-WANs) for the Internet-of-Things (IoT). Thus, this work here designs, specifies, implements, and evaluates a Cloud Radio Access Network (C-RAN) architecture for LoRa networks, while using (a) Software Defined Radios (SDR) to receive/send radio signals and (b) Docker to virtualize the setup. (c) A software modulator is developed to emit signals on the downlink targeting regular LoRa end-device receivers, such as Semtech SX1276 chips. Finally, the network, processing, and cost requirements of the C-RAN implemented are evaluated.

null

null

null

null

null

null

null