A Survey on AI-Empowered Softwarized Industrial IoT Networks

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This work analyzes the development and challenges of AI-empowered softwarized industrial IoT networks from the perspectives of related works, keyword trends, and classification statistics and comparisons. It also explores future research directions. Overall, the paper is well-written. To further enhance the current manuscript, the following recommendations are suggested for consideration:

1. In Section 2, it is recommended to separate the introduction of related work into a subsection. Meanwhile, the publication where the paper is published can also be considered important parameters.

2. The relevance in Tables 1 and 2 was derived subjectively based on the three major themes of the survey (AI, IIoT, and network softwarization). Does this imply that the relevance represents the number of matching themes?

3. The vicinity around the center of Figure 2 has densely packed values, resulting in significant data occlusion. Consider adopting a more suitable visualization method or adjusting the scale range and intervals.

4. From Figure 4, it is observed that between 2018 and 2024, there are hundreds of works across seven categories. However, only a portion of these works is analyzed in Section 4, with no indication provided in the text regarding the criteria for selecting these methods.

5. The calculation of contributions in Section 4 appears subjective, with no explicit calculation criteria provided. It might be beneficial to design a clear quantitative metric based on parameters such as citation count and implementation effectiveness, weighted accordingly.

6. The manuscript suggests organizing the content into the following structure: existing research and its classification, key technologies, applications, conclusions, and prospects.

7. The descriptions of the AI+IIOT application system structure, AI/ML, SDN/NFV, MEC, and other vital technologies are unclear. Ensure a comprehensive illustration of these technologies so readers can understand their roles and implementation methods.

Author Response

Please see the attachment. Thank you.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The word 'chapter' in the paper should be removed. There are typos in the paper that needs corrections

Abstract: The second last sentence is confusing and needs elaboration.

section 2.1: Rephrase first sentence.

Page 13: Define "OPC UA"

section 4.7: Rephrase line 712-713.

section 4.8. Line 764, remove the repetition of 'artificial neural network'

Conclusion: I suggest re-write it in a better way.

 

Comments on the Quality of English Language

The paper is easy to understand, there are however some typos and a few confusing statements that need correction.

Author Response

Please see the attachment. Thank you.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This survey reviews related works in AI-empowered softwarized Industrial IoT Networks, with a particular focus on the specific role of network softwarization, adding context about the trends, categorizing works in several types, and comparing them based on their contribution to the advance in the state of the art.

Some questions and comments are listed below:

1. In Section 4, why do the authors highlight reconfiguration solutions? The importance of reconfiguration should be introduced clearly. What is the role that the reconfiguration plays in SDN?

2. The authors just review the existing works one by one. The relations and differences among them should be logically presented to provide readers with more insights.

3. The logic between different subsections should be given before introducing them. It is confusing for the reviewer to “Reconfiguration solutions leveraging AI in software IIoT networks” and “AI in software IIoT networks with energy-related objectives”. For example, why survey energy-related objectives instead of application-specific metrics?

4. As a promising technique in 6G, more existing works on digital twins and the relations between DT and network virtualization should be summarized.  Section 4.4 at this time is not comprehensive:

a.C. Zhou, J. Gao, M. Li, N. Cheng, X. Shen, and W. Zhuang, "Digital Twin-Based 3D Map Management for Edge-Assisted Device Pose Tracking in Mobile AR," in IEEE Internet of Things Journal

b.Yue Yin, Pai Zheng, Chengxi Li, Lihui Wang, A state-of-the-art survey on Augmented Reality-assisted Digital Twin for futuristic human-centric industry transformation, Robotics and Computer-Integrated Manufacturing

5. Existing surveyed works on security are a little bit broad, and should be classified into some specific categories. In addition, some important topics on cybersecurity are not based on AI, which should also be compared to show the advances of AI-based security solutions.

a.N. Redini et al., "Diane: Identifying Fuzzing Triggers in Apps to Generate Under-constrained Inputs for IoT Devices," 2021 IEEE Symposium on Security and Privacy (SP), San Francisco, CA, USA, 2021

b.Xiaoyue Ma, Qiang Zeng, Haotian Chi, and Lannan Luo. No More Companion Apps Hacking but One Dongle: Hub-Based Blackbox Fuzzing of IoT Firmware. In Proceedings of the 21st Annual International Conference on Mobile Systems, Applications and Services (MobiSys '23), 2023

6. The authors adopt ‘star’ to evaluate the contributions of these works. While this approach can help readers to differentiate existing works, fairness may be an issue. What is the specific standard when the authors evaluate the contributions of these works? Can the adopted standard avoid some subjective factors in the evaluation?

 

 

Comments on the Quality of English Language

Some typos should be double-checked.

Author Response

Please see the attachment. Thank you.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

1. 
   
   • Paper provides a comprehensive review of existing literature related to AI and IoT in the industrial sector, focusing specifically on softwarization technologies such as Software-Defined Networking (SDN) and Network Functions Virtualization (NFV).
   • It identifies gaps in integrated practical implementations and the alignment with current standards, which are crucial for advancing industrial IoT applications.
2. 
   
   • Works are categorized based on their contributions to the field, especially regarding AI and IoT's roles in enhancing industrial network functionalities.
   • A comparative analysis is conducted against current architectural standards and practical implementations, highlighting areas that require further research and development.
3. 
   
   • The paper concludes with a detailed discussion on future challenges and potential research directions. This includes addressing the desynchronization between ongoing research and existing standards, and the need for practical implementations of the surveyed theoretical models.
4. 
   
   • The methodology involves a detailed search and review process focusing on publications from the last decade, emphasizing the significance of AI and IoT in smart industrial networks.
   • Trends and shifts in the domain are analyzed, showing a growing emphasis on cloud continuum and AI applications, with a noted decline in older technologies like SDN due to their maturation and integration into mainstream applications.

Comments on the Quality of English Language

1. Some sentences are quite long and complex, making them potentially difficult to follow. Breaking these into shorter, more concise sentences could enhance readability and ensure the content is accessible to a broader audience.

2. Ensuring consistent use of key terms and phrases related to the technologies discussed (such as AI, IoT, SDN, NFV) could help in maintaining clarity and avoiding confusion, especially for readers who may not be deeply familiar with the subject matter.

3. While appropriate for a technical audience, the paper could benefit from occasional simplifications or explanations of very technical terms and concepts, especially when they are introduced for the first time. This could make the paper more accessible to interdisciplinary readers who may be experts in related fields but not in the specific technologies discussed.

4. The paper frequently uses passive voice constructions, which are common in academic writing but can sometimes obscure the action or lead to vagueness. Wherever clarity and engagement can be enhanced, switching to active voice might be beneficial.

5. Attention to minor punctuation details and grammatical structures can further polish the presentation. For instance, ensuring that lists and series within the sentences are punctuated consistently would help maintain a professional tone.

6. Improving transitions between sections and paragraphs to create a smoother flow of ideas could enhance the reader's ability to follow complex arguments and the paper's overall narrative structure.

Author Response

Please see the attachment. Thank you.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

No more comments and suggestions