Fast-SNAP stands for Fast SNort using the Automata Processor (AP). This repository contains code for parsing the Snort rules and converting them into binary images which can be loaded on the AP for Deep Packet Inspection.

The implementation uses the Python API provided by the APSDK. The code was tested using APSDK 1.7.34 and Python 2.7.12.

The Snort rules can be downloaded from the website. They can then be converted for execution on the AP by executing the following:

python fastsnap.py <path to directory containing .rules files> -c

The above is the most conservative mode of operation. Flags can be used to enable support for multiple independent patterns in a rule, negated patterns, patterns with backreferences, etc. The following can be executed for the full usage information:

python fastsnap.py --help

• Roy, Indranil, Ankit Srivastava, Matt Grimm, Marziyeh Nourian, Michela Becchi, and Srinivas Aluru. "Evaluating High Performance Pattern Matching on the Automata Processor." IEEE Transactions on Computers (2019).
• Roy, Indranil, Ankit Srivastava, Marziyeh Nourian, Michela Becchi, and Srinivas Aluru. "High Performance Pattern Matching using the Automata Processor." In Parallel and Distributed Processing Symposium, 2016 IEEE International, pp. 1123-1132. IEEE, 2016.