Výkonové testování intrusního detekčního systému s využitím technologie CUDA

Abstract

The thesis deals with the performance testing of the Intrusion Detection System Suricata over CUDA technology. In the theoretical part, the performance possibilities of the technologies using graphic processors are introduced, in this case those from NVIDIA company. These possibilities are compared with the calculation capabilities of the computer processor. The advantages of using CUDA technology are then presented in one of the practical fields of application, such as ensuring a secure run of the computer network by either the intrusion detection or the prevention system. The practical part of the thesis then focuses on the performance testing itself, with a purpose to analyze the calculation performance of CUDA technology in comparison to the common processor in the environment of intrusion detection systems. The advantages of this technology are presented by the speed of information processing and also by the decrease of calculation performance of the computer processor, which is transfered to the graphic processor and its cores. In conclusion, according to the testing results, there is a comparison showing the areas where the benefit of CUDA technology is most significant.