Towards H-SDN Traffic Analytic Through Visual Analytics and Machine Learning


Chiang, Tin Tze and Tan, Saw Chin and Lee, Ching Kwang and Yusoff, Zulfadzli and Kaspin, Rizaludin (2019) Towards H-SDN Traffic Analytic Through Visual Analytics and Machine Learning. In: Security, Privacy, and Anonymity in Computation, Communication, and Storage. Springer, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 117-132. ISBN 978-3-030-24906-9

[img] Text
278.pdf - Published Version
Restricted to Repository staff only

Download (423kB)


With new networking paradigm emerged through Software-Defined Networking (SDN) offering various networking advantages over the traditional paradigm, organizations are attracted to migration their legacy networks to SDN networks. However, it is both challenging and impractical for organizations to migrate from traditional network architecture to full SDN architecture overnight. Therefore, the migration plan is performed in stages, resulting in a new type of network termed hybrid SDN (H-SDN). Effective migration and traffic scheduling in H-SDN environment are the two areas of challenges organizations face. Various solutions have been proposed in the literatures to address these two challenges. Differing from the approaches taken in the literatures, this work utilizes visual analytic and machine learning to address the two challenges. In both full SDN and H-SDN environment, literatures showed that data analytics applications have been successfully developed for various purposes as network security, traffic monitoring and traffic engineering. The success of data analytic applications is highly dependent on prior data analysis from both automated processing and human analysis. However, with the increasing volume of traffic data and the complex networking environment in both SDN and H-SDN networks, the need for both visual analytic and machine learning in inevitable for effective data analysis of network problems. Hence, the objectives of this article are three-folds: Firstly, to identify the limitations of the existing migration plan and traffic scheduling in H-SDN, followed by highlighting the challenges of the existing research works on SDN analytics in various network applications, and lastly, to propose the future research directions of SDN migration and H-SDN traffic scheduling through visual analytics and machine learning. Finally, this article presents the proposed framework termed VA-hSDN, a framework that utilizes visual analytics with machine learning to meet the challenges in SDN migration and traffic scheduling.

Item Type: Book Section
Uncontrolled Keywords: Software defined networking
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5101-6720 Telecommunication. Including telegraphy, telephone, radio, radar, television
Divisions: Faculty of Computing and Informatics (FCI)
Faculty of Engineering (FOE)
Depositing User: Ms Suzilawati Abu Samah
Date Deposited: 17 Sep 2021 03:51
Last Modified: 17 Sep 2021 03:51


Downloads per month over past year

View ItemEdit (login required)