Dış Kaynaklı Proje Süreçleri Yönetim Sistemi
Cryptanalysis of ciphers cannot be automated and they need to be constantly analyzed against known and new cryptanalysis techniques. Moreover, many cryptanalysis results are obtained theoretically and later some of them are shown to be wrong in practice due to oversight. Also, some theoretical attacks work better in practice due to the fact that the found weaknesses have better probabilities which can only be observed experimentally. The difference between theoretical and practical attacks are very clear in the differential-linear attacks on the ciphers ASCON and SERPENT.
Most theoretical attacks are not validated in practice due to required storage and computational power. However, experimentally performing the reduced versions of theoretical attacks can verify the correctness. Moreover, these experiments can be used to correct wrong attacks or improve correct attacks. Thus, experimentally verifying shortened versions of theoretically obtained cryptographic attacks is vital for assessing the security of ciphers.
Experimentally checking the correctness of verification sometimes requires computer aid and huge amount of computation. In this respect, we are going to use the huge parallel computational power of GPUs when CPUs are not sufficient. We are going to analyze mainly differential-linear attacks on symmetric encryption schemes in this respect. Our aims are to improve known attacks and experimentally obtain better distinguishers that cannot be observed by theoretical methods. Then we are going to focus on differential factors, which focus on the combination of the key addition and substitution layers, to verify, correct, and improve attacks.
Collaborating parties have proficiency in many cryptanalysis methods and they discovered their own cryptanalysis techniques. We expect to obtain new cryptanalysis results and get better understanding of theoretical and practical results.
In this two-year project, Cihangir Tezcan is expected to visit RUB for about a month every summer. Short-term visits by Gregor Leander at METU and Cihangir Tezcan at RUB and conference attendances throughout the year are also planned and expected. Moreover, a PhD student of RUB supervised by Gregor Leander, will visit METU for 6 weeks in the second year of this project. For the experiments, GPUs of Cyber Security Laboratory of Informatics Institute at METU are going to be used.
We expect that this collaboration will result in several publications at top international venues. Furthermore, this project will allow to turn the existing collaboration into a sustainable partnership. We expect future projects and student exchanges to be built upon the grounds laid here.
Besides that, the Turkish cryptographic community will benefit from the outputs of this project. There is no Turkish written source on cryptanalysis and during the project we are planning to create a Turkish-English dictionary for scientific terms about cryptography and cryptanalysis.