Enabling in-network aggregation by diffusion units for urban scale M2M networks
- Authors
- Fan, Yao-Chung; Chen, Huan; Leu, Fang-Yie; You, Ilsun
- Issue Date
- 1-Sep-2017
- Publisher
- Academic Press
- Keywords
- M2M networks; Data aggregation; ECA (Event Condition Action) applications; 5G cellular networks
- Citation
- Journal of Network and Computer Applications, v.93, pp 215 - 227
- Pages
- 13
- Journal Title
- Journal of Network and Computer Applications
- Volume
- 93
- Start Page
- 215
- End Page
- 227
- URI
- https://scholarworks.bwise.kr/sch/handle/2021.sw.sch/7223
- DOI
- 10.1016/j.jnca.2017.05.002
- ISSN
- 1084-8045
1095-8592
- Abstract
- Machine-to-Machine (M2M) applications, which currently are growing rapidly, will be an important source of traffic on 5G cellular networks. A typical working model of M2M networks is a machine which is equipped with sensors or meters and directly delivers sensed data (e.g. temperature readings or inventory levels) to other machines that take actions based on the data. In such a working model, machines producing readings are referred to as sources, while those consuming the readings are called sinks. In this study, we consider a multiple-source multiple-sink scenario in M2M networks, in which a source may be demanded by multiple sinks and a sink may consume readings generated by several source nodes. However, how to efficiently process and reduce communication traffic in an urban-scale M2M network has been an engineering challenge. To solve the problem, in this paper, we propose a novel in-network aggregation scheme, called Diffusion Unit (DU), which encodes readings and aggregates data collected from its neighbors to effectively reduce transmitted data in a M2M network. We will show how existing distinct counting algorithms, like Linear Counting and FlajoletMartin Counting, are adapted to implement a DU and introduce a family of techniques to optimize a DU's space efficiency. Furthermore, a comprehensive performance evaluation on the proposed techniques and some existing state-of-the-art techniques is presented. The results demonstrate that our techniques significantly outperform existing ones.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - College of Engineering > Department of Information Security Engineering > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.