Dual-IoTID: A Session-Based Dual IoT Device Identification Model

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript entitled "Dual-IoTID: A Session-based Dual IoT Device Identification Model" presents an intriguing perspective on IoT devices.

 Several clarifications and suggestions are provided for the manuscript:

 

 Clarifications:

 

In the Introduction, the phrase "...As of 2023, there are 16.7 billion..." should be updated to reflect the current year, 2024.

Section 4.3, "Algorithm selection," lacks references to the algorithms used.

The letter "X" is missing when mentioning eXtreme Gradient Boosting (GB) on line 357 of the manuscript.

 

Suggestions:

 

Consider including a section detailing the implementation characteristics of your algorithm. Highlight the algorithm's limitations in IoT device implementation.

Discuss the specific characteristics that machine learning algorithms should possess to accommodate devices with low computational capacity.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

In this article, authors propose a dual-machine-learning-based IoT device identification algorithm, Dual-IoTID, which identifies devices based on the payload of IoT device sessions. Compared to previous works that focused on identifying IoT devices by extracting features from the header of relatively easily tampered packets, their method focuses on extracting the payload content and features of IoT device sessions to identify IoT devices. The proposal is interesting and the results are competitive. However, I have some comments: 

 

1. In the introduction, I suggest explaining more about the reason for identifying IoT devices using IA or other algorithms and whether there are some advantages with respect to classic methods.

 

1. On page 8, step 2. Frequent Item Extraction,  I suggest adding a figure, or pseudocode, or diagram to describe this step in more detail.

2. On page 8, step 3. Initial Classification, The pseudocode or block diagram could be improved to explain this step.

3. In subsection 4.1 Dataset, the authors employed a dataset with 23 IoT devices in the experiments. If the number of devices is increased, what will the impact be on the proposed method?

4. In Table 2, is there any relation between the number of sessions for each IoT device and the accuracy or precision of the identification?

5. In subsection 4.1 Dataset, What are the criteria for using a 70:30 ratio for the experiments? What happens if the ratio is changed?

6. In subsection 5.1 Experimental results, Why did the authors select these conditions for experiments?

7. In Table 6, If the number of non-IoT devices is increased, what will impact the obtained results?

8. In Table 7, the authors show the accuracy values in the comparison, but is it possible to obtain and compare the time values used to perform the identification?

9. Please check the proper reference format according to the Journal.

10. Please check the English language and the article structure according to the Journal.

Comments on the Quality of English Language

1. Please check the English language and the article structure according to the Journal.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

In this paper, the authors proposed a method for IoT device identification based on network communication payload. The topic of this research work is very interesting and worth investigating. The authors very nicely explain the main concepts needed to understand the work, highlight the drawbacks of available works, and perfectly summarize the main contributions of the work. However, I have the following concerns that I would like to ask the authors to address them to improve the work.  

1- In Section 3.1, the definition of the steps for the proposed method can be summarized in an itemize list for the sake of clarity for the readers. 

2- In Section 3.1.1, it would be interesting for the authors to know the main reasons and rationale behind splitting the traffic based on flows and sessions 

3- The evaluation section demands a major revision to provide a comparison between the proposed method and other available SoAs. 

4- The authors should expand the part regarding the possible future directions to be explored, because the current one is very short and quite generic. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors, after reviewing the new version of the article, in my opinion all comments have been answered.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors addressed all of my concerns and the paper quality has been improved.