Purdue University Graduate School
Browse

Network Intrusion Detection Leveraging Targeted Regular Expressions On FPGAs

thesis
posted on 2025-07-23, 13:04 authored by Jack Tyler GardelJack Tyler Gardel
<p dir="ltr"><u>Abstract</u></p><p dir="ltr">Current software-based IDS solutions for protecting networks do not scale well to meet the higher linerates of larger enterprise networks such as those at universities. Much of the bottleneck has to do with regular expressions which take the most amount of time to process. While hardware accelerators such as GRAPEFRUIT and HARE exist, they are either too slow or not dense enough to support large rulesets. Pigasus IDS uses an FPGA-based approach to filtering out packets before regex matching but suffers from trace-dependent performance. We present TRex; an FPGA-based IDS that has a filter similar to Pigasus and a custom regular-expression accelerator for this application. TRex only checks regular expressions that are relevant to each packet, allowing it to process packets in parallel on separate rules. Exploiting parallelism in this way allows TRex to support 2x throughput, lessening the demand for software to take over.</p>

History

Degree Type

  • Master of Science in Electrical and Computer Engineering

Department

  • Electrical and Computer Engineering

Campus location

  • West Lafayette

Advisor/Supervisor/Committee Chair

Sanjay Rao

Advisor/Supervisor/Committee co-chair

Vishal Shrivastav

Additional Committee Member 2

Mark Johnson

Usage metrics

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC