unified2: pure-python parser for IDS (think Snort) unified2 binary log format

Module allows to process IDS logs in binary "unified2" format into python objects.

It does not resolve rule ids and is not meant to be a replacement for barnyard2 or Snort itself in that role.

In fact, I've found that snort-produced text logs are enough for my purposes, so this module isn't really maintained or updated (at least by me), so if you really find it useful, I encourage you to fork the module, and contact me to transfer pypi package ownership to you.

Main purpose is to extract a packet data from the log, associated with some particular triggered (and resolved/logged separately via other means, e.g. alert_syslog or alert_csv snort modules) rule, so I haven't paid much attention to processing event metadata.

Module doesn't have C components and doesn't use ctypes, so should be fairly portable to non-cPython language implementations.

Format

Format definition is derived from Snort headers (src/sfutil/Unified2_common.h) via pyclibrary module and are cached in unified2/_format.py file.

Newer definitions (say, if new data types were added) can be generated by running the same script on the Snort's Unified2_common.h:

bzr branch lp:pyclibrary
cd pyclibrary
python .../unified2/_format.py .../snort-2.X.Y.Z/src/sfutil/Unified2_common.h

Installation

It's a regular package for Python 2.7 (not 3.X).

Using pip is the best way:

% pip install unified2

If you don't have it, use:

% easy_install pip
% pip install unified2

Alternatively (see also):

% curl https://raw.github.com/pypa/pip/master/contrib/get-pip.py | python
% pip install unified2

Or, if you absolutely must:

% easy_install unified2

But, you really shouldn't do that.

Current-git version can be installed like this:

% pip install 'git+https://github.com/mk-fg/unified2.git#egg=unified2'

Usage

Simple example:

import unified2.parser
for ev, ev_tail in unified2.parser.parse('/var/log/snort/snort.u2.1337060186'):
	print 'Event:', ev
	if ev_tail: print 'Event tail:', ev_tail

Event object here is a dict of metadata and a "tail", which can either be a blob or a similar recursively-parsed tuple of metadata-dict and "tail" (e.g. for UNIFIED2_EXTRA_DATA).

unified2.parser.Parser interface is best illustrated by the unified2.parser.read function:

parser, buff_agg = Parser(), ''
while True:
	buff = parser.read(src)
	if not buff: break # EOF
	buff_agg += buff
	while True:
		buff_agg, ev = parser.process(buff_agg)
		if ev is None: break
		yield ev

Idea here is that Parser.read method should be called with a stream (e.g. a file object), returning however many bytes parser needs to get the next parseable chunk of data (one packet, in case of u2 log) or whatever can be read at the moment, empty string is usually an indication of EOF or maybe non-blocking read return.

Parser.process then should be called with accumulated (by Parser.read calls) buffer, returning the first packet that can be parsed from there (or None, if buffer isn't large enough) and remaining (non-parsed) buffer data.

Related stuff

• http://manual.snort.org/node21.html#SECTION00369000000000000000
• http://www.securixlive.com/barnyard2/docs/unified.php
• https://github.com/firnsy/barnyard2
• http://blog.snort.org/2010/12/working-with-unified-and-unified2-files.html
• https://github.com/mephux/unified2
• https://code.launchpad.net/~luke-campagnola/pyclibrary
• https://github.com/mk-fg/bordercamp-irc-bot