Architektury a softwarové vybavení pro vysokorychlostní zpracování síťového provozu

Architectures and software for high-speed network traffic processing

software

required - licence fee

Cabal Jakub, Ing. (CESNET)
Matoušek Jiří, Ing., Ph.D. (CESNET)
Špinler Martin, Ing. (CESNET)
Košař Vlastimil, Ing., Ph.D. (DCSY FIT BUT)
Fukač Tomáš, Ing. (DCSY FIT BUT)
Gurka Oliver (FIT BUT)
Šišmiš Lukáš, Ing. (CESNET)
Martínek Tomáš, doc. Ing., Ph.D. (DCSY FIT BUT)
Kořenek Jan, doc. Ing., Ph.D. (CESNET)

FPGA; encrypted traffic analysis; connection tracking table; machine learning

It is a software module for FPGA technology that provides efficient feature extraction from network traffic to enable encrypted traffic analysis and classification using machine learning techniques. The core of the software module is a connection tracking table, which allows the control of packet processing at the network flow level and aggregates statistics for IPFIX monitoring and classifiers created by machine learning techniques. Due to the well-designed architecture, the set of statistics can be changed easily. The module utilizes the network traffic characteristics and optimizes communication with the external DRAMs to achieve high speed. The module has been designed and implemented to be integrated into probes with FPGA SmartNICs and can be used for networks with up to 400Gbps speed.

Flow-based Encrypted Traffic Analysis (VJ02010024)

Accelerated Network Technologies Research Group (VZ ANT)

CESNET National Research and Education Network (CESNET)
Department of Computer Systems FIT BUT (DCSY FIT BUT)