kumoh national institute of technology
Networked Systems Lab.

Amaizu Gabriel, Jae-Min Lee and Dong-Seong Kim "SDNAD: A Software Defined Network Anomaly Detection Model using Enhanced Convolutional Neural Network", IET Communications, 2020 (R)
By : Gabriel
Date : 2020-04-16
Views : 215

SDNAD: A Software Defined Network Anomaly Detection Model using Enhanced Artificial Neural Network

Thank you for submitting your paper to IET Communications. The peer review process is now complete. I regret to inform you that the conclusion is that the paper is not suitable for publication in the journal.

The comments from the reviewers are included below. I appreciate that this decision will be a disappointment but would like to thank you for your interest in IET Communications.

Yours sincerely



Dr Liuqing Yang
Editor-in-Chief
IET Communications



Associate Editor Comments:

Comments to the author

I regret to inform you that the work has several issues as the two reviewers mentioned and can not be published in IET Communication. I suggest submitting it to a related conference.



Reviewer Comments:

Comments to the Author

#1 Submitted by: Reviewer 1
I recommend the authors submit this manuscript to a Conference or Workshop.


There is no novelty in Anomaly Detection Model based on SDN.
The authors just used SDN in text. There is no scientific and technical aspects on SDN-based Anomaly Detection Model.


Why authors used ANN to evaluate Anomaly Detection Model? What is enhanced method that gives novelty on classic ANN?
There is no comparison with powerful methods in evaluation results.


#2 Submitted by: Reviewer 2
After careful reading the paper, I did not see any novelty of the paper, the presentation of the paper requires to be updated and corporate the reason behind using SDN with AI method.

Why ANN considered? SDN is the real-time manner, how do you model it on the OF and rule distribution? It is not clear from the paper since there is not clear motivation and contribution presentations, it needs heavy work to be done in the paper.

The dataset used is unclear and there is not clear character explanation on it.

The training and ted phases are unclear for me to follow in the paper. Hence, the paper can not be published in the journal and perhaps the conference version is a replacement for the updated version of the paper.


#3 Submitted by: Reviewer 3
The authors must have to tackle the following issues present in their article so that it can be considered for publication.


1. Provide an industrial application in the Introduction section in order to further motivate this work.
2. In Table 1, the name or type of threats must be specified.
3. The authors should discuss on more related works. Some relevant studies can be added.
4. Authors can add a nice table of abbreviations and notations for easy follow with the paper.
5. There is no symmetry between the citations and figures in the entire section of performance evaluation. Authors should revise it.
6. Also, make sure that your paper is free of grammatical, spelling and other common errors.