e-space
Manchester Metropolitan University's Research Repository

    Hardware Acceleration of Network Intrusion Detection System Using FPGA

    Hashmi, Adeel (2011) Hardware Acceleration of Network Intrusion Detection System Using FPGA. Doctoral thesis (PhD), Manchester Metropolitan University.

    [img]
    Preview

    Available under License Creative Commons Attribution Non-commercial Share Alike.

    Download (2MB) | Preview

    Abstract

    This thesis presents new algorithms and hardware designs for Signature-based Network Intrusion Detection System (SB-NIDS) optimisation exploiting a hybrid hardwaresoftware co-designed embedded processing platform. The work describe concentrates on optimisation of a complete SB-NIDS Snort application software on a FPGA based hardware-software target rather than on the implementation of a single functional unit for hardware acceleration. Pattern Matching Hardware Accelerator (PMHA) based on Bloom filter was designed to optimise SB-NIDS performance for execution on a Xilinx MicroBlaze soft-core processor. The Bloom filter approach enables the potentially large number of network intrusion attack patterns to be efficiently represented and searched primarily using accesses to FPGA on-chip memory. The thesis demonstrates, the viability of hybrid hardware-software co-designed approach for SB-NIDS. Future work is required to investigate the effects of later generation FPGA technology and multi-core processors in order to clearly prove the benefits over conventional processor platforms for SB-NIDS. The strengths and weaknesses of the hardware accelerators and algorithms are analysed, and experimental results are examined to determine the effectiveness of the implementation. Experimental results confirm that the PMHA is capable of performing network packet analysis for gigabit rate network traffic. Experimental test results indicate that our SB-NIDS prototype implementation on relatively low clock rate embedded processing platform performance is approximately 1.7 times better than Snort executing on a general purpose processor on PC when comparing processor cycles rather than wall clock time.

    Impact and Reach

    Statistics

    Activity Overview
    6 month trend
    142Downloads
    6 month trend
    119Hits

    Additional statistics for this dataset are available via IRStats2.

    Repository staff only

    Edit record Edit record