Paper Abstract and Keywords |
Presentation |
2015-03-02 11:40
Hierarchical Traffic Engineering Based on Model Predictive Control Tatsuya Otoshi, Yuichi Ohsita, Masayuki Murata (Osaka Univ.), Yousuke Takahashi, Keisuke Ishibashi, Kohei Shiomoto (NTT), Tomoaki Hashimoto (Osaka Univ.) IN2014-136 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Traffic engineering with traffic prediction is one approach to accommodate time-varying traffic stably.
In this approach, the routes are calculated so as to avoid congestion based on the predicted traffic.
To achieve the robustness against the prediction error, we proposed a traffic engineering method called Model Predictive Traffic Engineering (MP-TE) which is based on the idea of Model Predictive Control (MPC).
However, this method has difficulty in scalability.
In this method, a central control server repeatedly collects the whole traffic information to correct the traffic prediction and recalculate the whole routes at each time slot.
In accordance, load on the central control server significantly increases as the network becomes larger.
To solve the scalability problem, we propose a prediction-based hierarchical traffic engineering method in this paper.
In this method, we divide the network in multiple areas and aggregate the topology in each area as an upper layer topology.
Control servers are deployed in each area and upper layer to calculate the routes inner area and inter area, respectively.
By reducing the scale of the topology which one control server manages, the load on each control server is kept low even when the scale of network become large.
Through the simulation, we show that the proposal method can reduce the calculation time while the achieved performance is close to that of the centralized control. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Model Predictive Control / Hierarchical Routing / Traffic Engineering / Traffic Prediction / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 114, no. 478, IN2014-136, pp. 91-96, March 2015. |
Paper # |
IN2014-136 |
Date of Issue |
2015-02-23 (IN) |
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 |
IN2014-136 |
|