camp0/aiengineAIEngine is a next generation interactive/programmable Python/Ruby/Java/Lua and Go NIDS (Network intrusion detection system).
AIEngine is a next generation interactive/programmable Python/Ruby/Java/Lua and Go network intrusion detection system engine with capabilities of learning without any human intervention, DNS domain classification, Spam detection, network collector, network forensics and many others.
AIEngine also helps network/security professionals to identify traffic and develop signatures for use them on NIDS, Firewalls, Traffic classifiers and so on.
The main functionalities of AIEngine are:
Check the docs folder for more information
To use AIEngine(reduce version) just execute the binary aiengine or use the python/ruby/java/lua binding.
luis@luis-xps:~/c++/aiengine/src$ ./aiengine -h
aiengine 2.1.0
Mandatory arguments:
-I [ --input ] arg Sets the network interface ,pcap file or
directory with pcap files.
Link Layer optional arguments:
-q [ --tag ] arg Selects the tag type of the ethernet layer (vlan,mpls).
TCP optional arguments:
-t [ --tcp-flows ] arg (=32768) Sets the number of TCP flows on the pool.
UDP optional arguments:
-u [ --udp-flows ] arg (=16384) Sets the number of UDP flows on the pool.
Domain optional arguments:
-D [ --domain-file ] arg Reads domain names from file.
-B [ --domain-protocol ] arg (=dns) Protocol to plug the domain-file (dns,
ssl, http).
-S [ --matched-domain ] Shows only the domains that matches.
Regex optional arguments:
-R [ --enable-signatures ] Enables the Signature engine.
-r [ --regex ] arg (=.*) Sets the regex for evaluate agains the flows.
-c [ --flow-class ] arg (=all) Uses tcp, udp or all for matches the signature
on the flows.
-m [ --matched-flows ] Shows the flows that matchs with the regex.
-M [ --matched-packet ] Shows the packet payload that matchs with
the regex.
-C [ --continue ] Continue evaluating the regex with the
next packets of the Flow.
-j [ --reject-flows ] Rejects the flows that matchs with the
regex.
-w [ --evidence ] Generates a pcap file with the matching
regex for forensic analysis.
Frequencies optional arguments:
-F [ --enable-frequencies ] Enables the Frequency engine.
-g [ --group-by ] arg (=dst-port) Groups frequencies by src-ip,dst-ip,src-por
t and dst-port.
-f [ --flow-type ] arg (=tcp) Uses tcp or udp flows.
-L [ --enable-learner ] Enables the Learner engine.
-k [ --key-learner ] arg (=80) Sets the key for the Learner engine.
-b [ --buffer-size ] arg (=64) Sets the size of the internal buffer for
generate the regex.
-Q [ --byte-quality ] arg (=80) Sets the minimum quality for the bytes of
the generated regex.
-y [ --enable-yara ] Generates a yara signature.
Optional arguments:
-n [ --stack ] arg (=lan) Sets the network stack (lan,mobile,lan6,virtual,
oflow).
-d [ --dumpflows ] Dump the flows to stdout.
-s [ --statistics ] arg (=0) Show statistics of the network stack (5 levels).
-T [ --timeout ] arg (=180) Sets the flows timeout.
-P [ --protocol ] arg Show statistics of a specific protocol of the
network stack.
-a [ --port ] arg (=0) Sets the HTTP listenting port.
-e [ --release ] Release the caches.
-l [ --release-cache ] arg Release a specific cache.
-p [ --pstatistics ] Show statistics of the process.
-o [ --summary ] Show protocol summmary statistics
(bytes,packets,% bytes,cache miss,memory).
-h [ --help ] Show help.
-v [ --version ] Show version string.
AIEngine supports six types of Network stacks depending on the network topology.
StackLan (lan) Local Area Network based on IPv4.
StackLanIPv6 (lan6) Local Area Network with IPv6 support.
StackMobile (mobile) Network Mobile (Gn interface) for IPv4.
StackVirtual (virtual) Stack for virtual/cloud environments with VxLan and GRE Transparent.
StackOpenFlow (oflow) Stack for openflow environments.
StackMobileIPv6 (mobile6) Network Mobile (Gn interface) for IPv6.
AIEngine is a python/ruby/java/lua module also that allows to be more flexible in terms of integration with other systems and functionalities. The main objects that the python module provide export are the following ones.
BitcoinInfo
Cache
CoAPInfo
DCERCPInfo
DHCPInfo
DHCPv6Info
DNSInfo
DTLSInfo
DatabaseAdaptor
DomainName
DomainNameManager
Flow
FlowManager
Frequencies
FrequencyGroup
HTTPInfo
HTTPUriSet
IMAPInfo
IPAbstractSet
IPRadixTree
IPSet
IPSetManager
LearnerEngine
MQTTInfo
NetbiosInfo
NetworkStack
StackLan
StackLanIPv6
StackMobile
StackMobileIPv6
StackOpenFlow
StackVirtual
POPInfo
PacketDispatcher
PacketFrequencies
QuicInfo
Regex
RegexManager
SIPInfo
SMBInfo
SMTPInfo
SSDPInfo
SSHInfo
SSLInfo
TCPInfo
For a complete description of the class methods in Python
import pyaiengine
help(pyaiengine)
Check the directory examples in order to have usefull use cases, and check /docs for documentation
You should have installed pcre-devel, libpcap-devel and boost-devel in your system as minimun.
$ git clone https://bitbucket.com/camp0/aiengine
$ ./autogen.sh
$ ./configure
$ make
The system provides the following enable/disable functionalities depending your requirements.
This options only can be enable/disable on compilation time on the configure script.
For compile the Python library is also recomended boost-python3-devel or boost-python-devel and python-devel.
The first option for compile the library is using O3 compile optimization, this will generate a small library
$ git clone https://bitbucket.com/camp0/aiengine
$ ./autogen.sh
$ ./configure
$ cd src
$ make python
$ python pyai_test.py
The second option will compile the library by using the standard pythonic way by using setup.py, this will generate a bigger library size if compare with the previous one.
$ git clone https://bitbucket.com/camp0/aiengine
$ ./autogen.sh
$ ./configure
$ cd src
$ python setup.py build_ext -i
$ python pyai_test.py
$ python3.6 setup.py build_ext -i
$ python3.6 pyai_test.py
The python lib contains all the functionality that the engine provides.
For compile on Ruby swig and ruby-devel are needed.
$ git clone https://bitbucket.com/camp0/aiengine
$ ./autogen.sh
$ ./configure
$ cd src
$ make ruby
$ ruby ruai_tests.rb
The ruby lib dont provides all the functionality of the engine.
For compile on Java the java devel package is needed.
$ git clone https://bitbucket.com/camp0/aiengine
$ ./autogen.sh
$ ./configure
$ cd src
$ make java
$ java -cp ".:/usr/share/java/junit.jar:/usr/share/java/hamcrest/core.jar:./buildjava" org.junit.runner.JUnitCore JunitTestSuite
The java lib provides some of the functionality of the engine, with some exceptions.
For compile on Lua swig the lua-devel package is needed.
$ git clone https://bitbucket.com/camp0/aiengine
$ ./autogen.sh
$ ./configure
$ cd src
$ make lua
$ lua luai_tests.lua
The lua lib provides the majority of the functionality of the engine.
For compile on Go swig.
$ git clone https://bitbucket.com/camp0/aiengine
$ ./autogen.sh
$ ./configure
$ cd src
$ make go
$ ./goai_test
The Go lib have some functional restrictions due to the nature of the language.
You can find a full description of different use cases with the most common NIDS on http://aiengine.readthedocs.io/en/latest/aiengine.html#performance-with-other-engines
AIEngine is develop with c++11/14/17 standard and is under the terms of GPLv2.
Check out the AIEngine source with
$ git clone https://bitbucket.com/camp0/aiengine
If you are interested in a specific functionality, promote the project or just need some help just drop me an email. Contributions are always helpful.
AIEngine have been develop using test driven development. So in order to maintain the same life cycle, the new functionatly should have unit test on the directory created of the new functionality and for integrate with all the system, later integrate with the main tests.cc file on the /src directory
If you have a use case that with the provided examples is not cover please feel free to contact me.
If you like the project, you can buy me beers with bitcoin :)
3DG7FhkFUuKq7VhfXdP9QpsUzfWPXu49ng