Paper Abstract and Keywords |
Presentation |
2014-06-27 14:45
Aggregating Flows for Traffic Engineering Noriaki Kamiyama (Osaka Univ./NTT), Yousuke Takahashi, Keisuke Ishibashi, Kohei Shiomoto (NTT), Tatsuya Otoshi, Yuichi Ohsita, Masayuki Murata (Osaka Univ.) NS2014-44 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
SDN enables us to control the routes with finer granularity of individual flows, and we can expect an improvement of the effect of balancing the link load. However, the increase of states which need to be managed in each node is a problem in large-scale networks. To solve this problem, aggregating multiple flows into a macro flow and assigning the route for each macro flow seems effective. When constructing macro flows as the target of TE, it is desirable to suppress the variation of traffic rate in each macro flow to improve the stability of routes. Therefore, in this paper, we propose two approaches generating macro flows: the one is based on a greedy algorithm minimizing the traffic variation of macro flows, and the other is an approach clustering flows into some groups and optimizing the ratio of traffic of each group which is aggregated into each macro flow. Through the evaluation using the traffic demand matrixes during 48 hours in the Internet2, we show that the proposed methods can reduce the number of TE targets by about 1/50 $sim$ 1/400 while avoiding degradation of the effect of balancing the link load of TE. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
SDN / traffic engineering / granularity / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 114, no. 107, NS2014-44, pp. 29-34, June 2014. |
Paper # |
NS2014-44 |
Date of Issue |
2014-06-19 (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 |
NS2014-44 |
|