• Please send your full manuscript to: jcm@etpub.com

Classification of Freenet Traffic Flow Based on Machine Learning

Seungwoon Lee 1, Seung-hun Shin 2, and Byeong-hee Roh 1
1. Dept. of Computer Engineering, Ajou University, Suwon 16499, Korea
2. Dasan University College, Ajou University, Suwon 16499, Korea
Abstract—An anonymous overlay network is a virtual and logical network which can assure privacy but is often misused as a crime. Therefore, it is necessary to find users operating the abnormal overlay network in managed network for network administrator. However, there is a lack of research on host detection using Freenet which is one of the popular anonymous overlay networks and all of previous methods require that at least one Freenet node be inserted into the network. In this paper, we propose classification of Freenet traffic flow based on machine learning. Through this, it is possible to identify the host operating the Freenet inside the network without joining Freenet. We also evaluate the performance of classification algorithms. Among them, Decision Tree is most effective with 94% of precision and 0.0029 sec of time spent.
 
Index Terms—Freenet, anonymous overlay network, traffic classification, network security

Cite: Seungwoon Lee, Seung-hun Shin, and Byeong-hee Roh, "Classification of Freenet Traffic Flow Based on Machine Learning," Journal of Communications, vol. 13, no. 11, pp. 654-660, 2018. Doi: 10.12720/jcm.13.11.654-660
Copyright © 2013-2017 Journal of Communications, All Rights Reserved
E-mail: jcm@etpub.com