|
カワイ ユキコ
KAWAI YUKIKO
河合 由起子 所属 京都産業大学 情報理工学部 情報理工学科 職種 教授 |
|
| 言語種別 | 英語 |
| 発行・発表の年月 | 2016 |
| 形態種別 | 研究論文 |
| 査読 | 査読あり |
| 標題 | Scalable pub/sub system using openflow control |
| 執筆形態 | その他 |
| 掲載誌名 | Journal of Information Processing |
| 出版社・発行元 | Information Processing Society of Japan |
| 巻・号・頁 | 24(4),pp.635-646 |
| 著者・共著者 | Toyokazu Akiyama,Yuuichi Teranishi,Ryohei Banno,Katsuyoshi Iida,Yukiko Kawai |
| 概要 | Pub/Sub communication model becomes a basis of various applications, e.g., IoT/M2M, SNS. These application domains require new properties of the Pub/Sub infrastructure, for example, supporting a large number of devices in a widely distributed manner. In order to meet the demands, we proposed Scalable Pub/Sub System using OpenFlow Controller, which we call SDN Aware Pub/Sub (SAPS). SAPS utilizes the both Application Layer Multicast (ALM) and OpenFlow based multicast (OFM). A simulation was done for evaluating the hybrid architecture in traffic and transmission delay reduction. The result shows that in the tree topology, even if it has only 100 subscribers, OFM can reduce inter-cluster traffic 71.6% with 16 clusters compared to ALM-LA. It can also reduce maximum intercluster hops 87.5%. On the other hand, with only 100 subscribers, almost all of switches are involved in the OFM tree construction and consume flow table space for the topic. It indicates that our hybrid approach is effective in Pub/Sub optimization considering the resource limitation of OpenFlow switches. |
| DOI | 10.2197/ipsjjip.24.635 |
| ISSN | 1882-6652 |
| DBLP ID | journals/jip/AkiyamaTBIK16 |
| PermalinkURL | http://dblp.uni-trier.de/db/journals/jip/jip24.html#journals/jip/AkiyamaTBIK16 |