Snort3

The next generation of the Snort intrusion prevention and detection system.

Snort++

Snort 3 is the next generation Snort IPS (Intrusion Prevention System). This file will show you what Snort++ has to offer and guide you through the steps from download to demo. If you are unfamiliar with Snort you should take a look at the Snort documentation first. We will cover the following topics:

OVERVIEW

This version of Snort++ includes new features as well as all Snort 2.X features and bug fixes for the base version of Snort except as indicated below:

Project = Snort++
Binary = snort
Version = 3.0.0 (Build 250) from 2.9.11

Here are some key features of Snort++:

  • Support multiple packet processing threads
  • Use a shared configuration and attribute table
  • Use a simple, scriptable configuration
  • Make key components pluggable
  • Autodetect services for portless configuration
  • Support sticky buffers in rules
  • Autogenerate reference documentation
  • Provide better cross platform support
  • Facilitate component testing
  • Use a shared network map

Additional features on the roadmap include:

  • Support pipelining of packet processing
  • Support hardware offload and data plane integration
  • Support proxy mode
  • Windows support

DEPENDENCIES

If you already build Snort, you may have everything you need. If not, grab the latest:

Additional packages provide optional features. Check the manual for more.

DOWNLOAD

There is a source tarball available in the Downloads section on snort.org:

snort-3.0.0-a3.tar.gz

You can also get the code with:

git clone https://github.com/snort3/snort3.git

There are separate extras packages for cmake that provide additional features and demonstrate how to build plugins. The source for extras is in snort3_extra.git repo.

BUILD SNORT

Follow these steps:

  1. Set up source directory:

  • If you are using a github clone:

    cd snort3/

  • Otherwise, do this:

    tar zxf snort-tarball
cd snort-3.0.0*

  1. Setup install path:

    export my_path=/path/to/snorty

  2. Compile and install:

  • To build with cmake and make, run configure_cmake.sh. It will automatically create and populate a new subdirectory named 'build'.

    ./configure_cmake.sh --prefix=$my_path
cd build
make -j $(nproc) install

Note:

  • If you can do src/snort -V you built successfully.
  • If you are familiar with cmake, you can run cmake/ccmake instead of configure_cmake.sh.
  • cmake --help will list any available generators, such as Xcode. Feel free to use one, however help with those will be provided separately.

RUN SNORT

Here are some examples. If you are using Talos rules and/or configs, you should first set any needed variables at the top of snort.lua and snort_defaults.lua.

  • Snort++ provides lots of help from the command line, including:

    $my_path/bin/snort --help
$my_path/bin/snort --help-module suppress
$my_path/bin/snort --help-config | grep thread

  • Examine and dump a pcap. In the following, replace a.pcap with your favorite:

    $my_path/bin/snort -r a.pcap
$my_path/bin/snort -L dump -d -e -q -r a.pcap

  • Verify a config, with or w/o rules:

    $my_path/bin/snort -c $my_path/etc/snort/snort.lua
$my_path/bin/snort -c $my_path/etc/snort/snort.lua -R $my_path/etc/snort/sample.rules

  • Run IDS mode. In the following, replace pcaps/ with a path to a directory with one or more *.pcap files:

    $my_path/bin/snort -c $my_path/etc/snort/snort.lua -R $my_path/etc/snort/sample.rules \
    -r a.pcap -A alert_test -n 100000

  • Let's suppress 1:2123. We could edit the conf or just do this:

    $my_path/bin/snort -c $my_path/etc/snort/snort.lua -R $my_path/etc/snort/sample.rules \
    -r a.pcap -A alert_test -n 100000 --lua "suppress = { { gid = 1, sid = 2123 } }"

  • Go whole hog on a directory with multiple packet threads:

    $my_path/bin/snort -c $my_path/etc/snort/snort.lua -R $my_path/etc/snort/sample.rules \
    --pcap-filter \*.pcap --pcap-dir pcaps/ -A alert_fast --max-packet-threads 8

Additional examples are given in doc/usage.txt.

FUZZ TESTING

Snort includes libFuzzer fuzz targets for testing various components. Building and running fuzzers requires Clang.

  • Build fuzzers with AddressSanitizer:

    CC=clang CXX=clang++ ./configure_cmake.sh --enable-fuzz-sanitizer --enable-fuzzers \
    --enable-address-sanitizer
cd build
make fuzz -j $(nproc)

  • If you need to specify DAQ include and library paths:

    CC=clang CXX=clang++ ./configure_cmake.sh --enable-fuzz-sanitizer --enable-fuzzers \
    --enable-address-sanitizer --with-daq-includes=/path/to/daq/include --with-daq-libraries=/path/to/daq/lib
cd build
make fuzz -j $(nproc)

  • Run a fuzz target by providing a corpus directory. Each target is built under build/fuzz/:

    mkdir file_olefile_fuzz_corpus
./fuzz/file_olefile_fuzz ./file_olefile_fuzz_corpus

  • If DAQ libraries are not on the default library path, set LD_LIBRARY_PATH:

    LD_LIBRARY_PATH=/path/to/daq/lib ./fuzz/file_olefile_fuzz ./file_olefile_fuzz_corpus

  • Press Ctrl+C to stop the fuzzer. List saved corpus files with:

    ls ./file_olefile_fuzz_corpus

  • Build with coverage instrumentation instead of AddressSanitizer to measure which code paths the fuzzer exercises:

    CC=clang CXX=clang++ ./configure_cmake.sh --enable-fuzz-sanitizer --enable-fuzzers \
    --enable-fuzz-coverage
cd build
make fuzz -j $(nproc)

  • Run the coverage build against an existing corpus (using -runs=0 to replay without generating new inputs):

    LLVM_PROFILE_FILE="file_olefile_fuzz.profraw" ./fuzz/file_olefile_fuzz \
    file_olefile_fuzz_corpus -runs=0 -detect_leaks=0

  • Generate an HTML coverage report:

    llvm-profdata merge -sparse file_olefile_fuzz.profraw -o file_olefile_fuzz.profdata
llvm-cov show -format=html -instr-profile=file_olefile_fuzz.profdata \
    ./fuzz/file_olefile_fuzz -output-dir=file_olefile_fuzz_cov_html

    Open file_olefile_fuzz_cov_html/index.html in a browser to view the report.

DOCUMENTATION

Take a look at the manual, parts of which are generated by the code so it stays up to date:

$my_path/share/doc/snort/snort_manual.pdf
$my_path/share/doc/snort/snort_manual.html
$my_path/share/doc/snort/snort_manual/index.html

It does not yet have much on the how and why, but it does have all the currently available configuration, etc. Some key changes to rules:

  • you must use comma separated content sub options like this: content:"foo", nocase;
  • buffer selectors must appear before the content and remain in effect until changed
  • pcre buffer selectors were deleted
  • check the manual for more on Snort++ vs Snort
  • check the manual reference section to understand how parameters are defined, etc.

It also covers new features not demonstrated here:

  • snort2lua, a tool to convert Snort 2.X conf and rules to the new form
  • a new HTTP inspector
  • a binder, for mapping configuration to traffic
  • a wizard for port-independent configuration
  • improved rule parsing - arbitrary whitespace, C style comments, #begin/#end comments
  • local and remote command line shell

SQUEAL

o")~

We hope you are as excited about Snort++ as we are. Let us know what you think on the snort-users list. In the meantime, we'll keep our snout to the grindstone.