Node-based Sampling P2P Bot Detection

Chunyong Yin, Ruxia Sun, Lei Yang, Darius Iko


The concept of using node-based sampling for the treatment of packet capture mechanism based on Libpcap of network-based detecting Peer-to-Peer botnet process was tested, and its effect on the time window of feature extracting and sampling time interval was explored. Node-based sampling treatment resulted in significant increase in the detection performance due to node profile of the novel behaviors to the detected computer in Peer-to-Peer bot detection, and the degradation of false positive. At relatively right time window (e.g., about 180s), precision was completely maximized, while the false positive decreased by 10% to 15%. The detection rate can be significantly increased due to the false positive degradation. A new performance index called Comprehensive Evaluation Index is proposed for more clearly represent the effectiveness. Sampling can reduce morn than 60% input raw packet traces and achieve a high detection rate (about 99%) and a low false positive rates (0-2%).



Full Text:



  • There are currently no refbacks.

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

shopify stats IJEECS visitor statistics