Paper Abstract and Keywords |
Presentation |
2018-03-02 15:20
Method for Generating a Data Set to Detect Cyber Attacks for Autonomous and Distributed Internet Security Infrastructure Yusei Katsura, Hiroyuki Kimiyama (Tokyo Denki Univ.), Akihiro Nakao (Tokyo Univ.), Naoki Yonezaki, Tomoaki Tsutsumi, Kaoru Sano (Tokyo Denki Univ.), Takeshi Okamoto, Mitsuru Maruyama (Kanagawa Institutes of Technology), Hiroshi Kobayashi (Tokyo Denki Univ.) NS2017-238 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
We proposed "Autonomous and distributed Internet security (AIS) infrastructure" that enables to protect our resources on the Internet by cooperating with all ISPs. One of AIS functions is that Multi-Layer Binding Router, which is core device of AIS infrastructure, discards attack packets detected and requested to discard by end nodes. Attack packets detection proceeds rapidly and correctly since this function works effectively. We developed a program to extract statistical attribution per TCP session from traffic data without sampling. We experimented in extracting statistical attribution from actual traffic data, and classifying attack packets and normal packets by machine learning methods with that statistical attribution. We demonstrated that this developed program is effective to detect attack packets and some issues to resolve from these results. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Cyber attack / Attack detection / Statistical attribution / Machine learning / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 117, no. 459, NS2017-238, pp. 397-401, March 2018. |
Paper # |
NS2017-238 |
Date of Issue |
2018-02-22 (NS) |
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 |
NS2017-238 |
|