Priority-Aware Resource Management for Adaptive Service Function Chaining in Real-Time Intelligent IoT Services

Round 1

Reviewer 1 Report

In this manuscript, a priority-aware resource management for adaptive service function chaining is provided by modeling the configuration of real-time Internet of Things (IoT) service requests associated with various mission-critical applications. The primary features that impact the optimization of configuration times and resource utilization are considered in the modeling. The performance evaluation of the proposed priority-aware policy with multi-agent deep Q-networks for optimal resource management is conducted. Overall, the present manuscript is in good organization, and could be accepted after minor revisions.

1. Please correct the grammar error in the sentence 'In IoT networks, the configuration of adaptive service function chaining (SFC), a set of virtual network functions (VNFs) orders for a particular service to execute in network function virtualization (NFV)-enabled environment, is emerged to ensure the optimality and elasticity of resource expenditure.'

2. In the lines 221 and 222, Equations (1) and (2) are not presented as the formula of equations; so as Equation (3) in the lines 236 and 237. They need to be corrected.

3. The symbol of execution delays in Line 246 needs to be corrected.

4. Uncertainty is an intrinsic attribute to physical systems, such as the measurement error of the end-to-end executing delay of requested services. How the proposed scheme performs and whether its advantages over other existing reference schemes remain in this situation? Detailed analysis and discussion are suggested to be included.

5. Too many abbreviations are included in the text. A list of abbreviations is suggested to be presented in front or at end of the main text.

Author Response

Dear Honorable Reviewer, 

We are thankful for the constructive comments to improve our manuscript’s quality. Please kindly see the attachment for our responses. 

Best Regards,

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper presents an SFC model to configure requests for IoT applications. The proposed model uses reinforcement learning and has shown promising results. Some sections are well written and clear, as the related works, in section 2. The results are clear, and the performance of the proposed system is relevant. Some suggestions are addressed by the authors.

- The results are not mentioned in the abstract. Please, rewrite inserting the results obtained in Section 5.

- Please, increase the quality of Figure 1. Moreover, Figures 1, 2, and 4 are presented before being mentioned in the text. 

-Line 53:” Experience- and data-driven (…)”. It appears that is something missing after “experience”.

- Lines 71-75. Insert references to support the challenges of IoT services.

- I think that sections 1.1 and 1.2 are not necessary. These sections can be inserted into the text of the introduction itself.

- Define TOSCA in text.

- Improve the legend of Table 1.

- Use a fixed number for decimal points on results.

- It would be interesting to evaluate the results presented in Figure 5 in terms of repeatability. Instead of analyzing just one point on the graphs, error dispersions for each of the metrics could be considered.

- The main question of section 5 is the lack of discussion of the results obtained with the models present in the literature. The obtained results must be compared with those already developed by other authors to provide greater solidity to the PA-DQNs-RM.

- The authors can address in which IoT-applied cases the PA-DQNs-RM can be targeted. Would all types of IoT devices benefit from it?

Author Response

Dear Honorable Reviewer,

We are truly thankful for reviewing our manuscript and providing us the significant comments and suggestions for augmenting manuscript quality. Please see the attachment for the responses.

Best Regards, 

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper "Priority-Aware Resource Management for Adaptive Service Function Chaining in Real-Time Intelligent IoT Services" presents a novel way for IoT applications to prioritise traffic. The introduction does not give a great deal of detail, but section 2 gives the necessary background and references.

The paper seems to report on moderate improvement on the responsiveness of the network. From the graphs and text, the improvements all seem marginal, which is disappointing from my point of view. My main points of difficulty with the manuscript are the following:

1) The use of language is convoluted and confusing. My feeling is that the clarity could be improved a lot, thereby making the paper more relevant.

2) The conclusions are not very clear. The conclusion should summarise the results, in the sense of, what gains have been made on the responsiveness of the network?

3) It is unclear what the role of optimisation and self-learning is in this paper. There are no graphs showing the results of this optimisation.

In summary, I find this paper a little underwhelming and some of the language is confusing. However, the results are fine and assuming that the authors take the above issues into account, I would recommend publication. 

Author Response

Dear Honorable Reviewer,

We are truly thankful for reviewing our manuscript and providing us the significant comments and suggestions for augmenting manuscript quality. Please kindly check the attachment for response details.

Best Regards,

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors did all the suggestions. Inserting the information among the sections increase the comprehension for the readers. I think that the paper is ready for publication.