Utility for extraction of subset of KDD '99 features [1] from realtime network traffic or .pcap file This utility is a part of our project at University of Bergen.
Some feature might not be calculated exactly same way as in KDD, because there was no documentation explaining the details of KDD implementation found. Algorithms are based on some articles [2][3] and observation of values in KDD dataset.
Features in KDD should be the same as features introduced by Lee & Stolfo in their work [2].
- Current version is not 100% guarenteed to be perfect in sense that some features might be calculated bit different algorighms than KDD '99 dataset a Lee & Stolfo used. Hovewer, it is suitable for educational purposes.
- Compiled & tested in following environments:
- Windows 7 x64, MSCV 2015 (14), WinPcap 4.1.3
- Windows 7 x64, MSCV 2013 (12), WinPcap 4.1.3
- Ubuntu 12.04 x64, gcc 4.6.3, libpcap 4.2
- Subset of KDD '99 features [1]
- Content features (columns 10-22 of KDD) are not included
- Optional extra features - IP addresses, ports, timestamp of last packet (option
-e)
- Sniffer
- Network traffic sniffer & frame parser
- IP reassembler
- Only IP header "summaries"
- Payload not reassembled (content features not extracted, it is not needed)
- Connection/Conversation reconstructor
- Reconstructs conversations
- Computes intrinsic features (columns 1-9 of KDD)
- Statistical engine
- Computes derived features (columns 23-41 of KDD)
- Create a folder to temporal build files
mkdir build-files - Enter in the folder and compile the cache
cd build-files
cmake -DCMAKE_BUILD_TYPE=Debug -G "CodeBlocks - Unix Makefiles" .. - Exit the folder of build cache and compile the project
cd ..
cmake --build ./build-files --target kdd99extractor -- -j 4 - Path to compiled project is:
build-files/src/kdd99extractor
- TODOs (e.g. IP checksum checking not implemented)
- Known/possible problems, bugs & limitations
[1] KDD Cup 1999 Data, http://kdd.ics.uci.edu/databases/kddcup99/kddcup99.html