Abschlussarbeiten

Nachstehend finden Sie eine Liste der offenen Themen in unserer Gruppe. Wenn Sie an einem dieser Themen interessiert sind, wenden Sie sich bitte direkt an die angegebene Kontaktperson. Wir haben aktuell keine anderen offenen Themen. Wenn Sie jedoch eine interessante Idee für eine Master- oder Bachelorarbeit im Bereich QPC haben, k?nnen Sie dieses Thema gerne einem oder allen von uns vorschlagen.

Offene Themen

• Title: Security Analysis of MAYO (Master thesis)
  Description: MAYO is a recently published signature scheme based on multivariate quadratic equations. Its performance and key sizes are compatible to lattice-based signature schemes, including NIST's finalists Dilithium and Falcon. An appealing task for such a young scheme is to investigate it regarding possible improvements in cryptanalysis. Like Rainbow and LUOV, MAYO is a modification of UOV. The past has shown that such modifications can lead to additional attack vectors, that might be exploited. Thus, a detailed security analysis is an important step in the development process of MAYO.
  Contact: Thomas Aulbach

• Title: Analyzing attacks against code-based signatures (Master thesis)
  Description: Three signature schemes based on codes were submitted to the NIST PQC standardization process. During the first round, all of these submissions have been broken. The goal of this thesis is identify design mistakes made in the development of these signature schemes by analyzing the attacks and looking for similarities and differences.
  Contact: Juliane Kr?mer (jointly supervised with Patrick Struck)

• Title: Studying several algorithms for optimizing code-based cryptography (Master thesis, in cooperation with the Thales Group)
  Description: Classic-McEliece is one of the oldest crypto schemes for asymmetric key exchange. From a security perspective, it is also one of the best algorithms, since there was no substantial decrease in its security level (as compared to other schemes, e.g., from lattice-based or multivariate cryptography). Besides these advantages, there are three disadvantages that are often reasons for avoiding the algorithm in practical use cases: 1) the public key size, 2) the time for computing the public key and 3) the need to store a big matrix on a processor until the computation of the public key is done. The main bottleneck for the last two disadvantages is the Gauss-Jordan algorithm for matrix inversion, which is mostly done in a na?ve manner. In this work, the main goal is to investigate the use of other algorithms as an alternative to the na?ve Gauss-Jordan, like Wiedemann’s algorithm, Strassen’s Algorithm or the Four Russian’s algorithm (or a combination). The algorithms should be investigated in regard for a sequential inversion (no need to hold the full matrix in a processor) with a low hardware footprint and in regard for parallelization with a larger hardware footprint but also for faster computation. After the investigation, the algorithm (or a combination of the mentioned algorithms) should be implemented on an FPGA.
  Contact: Juliane Kr?mer