Paper Abstract and Keywords |
Presentation |
2016-03-04 14:30
An Autonomous DDoS Backscatter Detection System from Darknet Traffic Yuki Ukawa, Jun Kitazono, Seiichi Ozawa (Kobe Univ.), Tao Ban, Junji Nakazato (NICT), Jumpei Shimamura (clwit) ICSS2015-67 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
This paper proposes an autonomous DDoS backscatter detection system from UDP darknet traffic. To identify DDoS backscatter, we define 17 features and classify them using an L2-SVM. In addition, to adapt to emergence of new patterns of DDoS attacks, we utilize a one-class SVM to detect outliers and continuously update the L2-SVM classifier. In the experiments, we use a traffic data collected by darknet sensor of NICT for half a year, and show that the proposed system can detect DDoS backscatter with 0.90 in F-measure on average. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
network security / DDoS attacks / machine learning / Support Vector Machine / outlier detection / incremental learning / / |
Reference Info. |
IEICE Tech. Rep., vol. 115, no. 488, ICSS2015-67, pp. 123-128, March 2016. |
Paper # |
ICSS2015-67 |
Date of Issue |
2016-02-25 (ICSS) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
ICSS2015-67 |
|