Exploiting Anytime Algorithms for Collaborative Service Execution in Edge Computing

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper presents a well-researched and innovative approach to collaborative service execution in edge computing environments. The proposed anytime algorithms are a valuable addition to the field, demonstrating significant potential for improving service quality and responsiveness. The topic of the paper is interesting. My concerns of the paper are as follows.

1. Algorithm 1 Anytime cluster formation is quite straightforward and intuitive. It is simple as the primary contribution of the paper. 

2. Section 4.3 is also quite not necessary since those properties are obvious.

3. In Section 5，the authors might consider including a section on the ethical implications of anytime algorithms, particularly regarding decision-making under uncertainty.

4. While the simulations are robust, the lack of real-world testing is a drawback needing be addressed in Section 6.

5. The paper could benefit from a discussion on the limitations of the proposed approach, including potential scalability issues in Section 6.

Author Response

We appreciate the reviewer's positive feedback on our paper and
insightful comments and suggestions for improvement. We have addressed
each of the concerns raised to enhance the quality and relevance of
our work.

1 - We acknowledge the reviewer's perspective on the simplicity and intuitiveness of Algorithm 1, "Anytime cluster formation," and recognize its significance as a primary contribution. We will ensure that its importance is appropriately emphasized in the paper.

2 - Regarding Section 4.3, which aims to demonstrate the alignment of the cluster formation algorithm with the desirable properties of anytime algorithms, we believe it is essential to retain this section in the final version of the paper. This inclusion ensures consistency with Section 5.3, where the service proposal anytime algorithm is similarly evaluated.

3 -  We recognize the importance of addressing the ethical
  implications of anytime algorithms, particularly in decision-making
  under uncertainty. We incorporated a dedicated portion of Section 2
  to discuss these ethical considerations thoroughly.

4 - While our simulations are robust, we acknowledge the importance of real-world testing to validate the effectiveness of our approach. We addressed this limitation by discussing potential avenues for real-world testing and practical implementation in Section 6.

5 - We agree that discussing the limitations of our proposed
  approach, including scalability issues, is essential for providing a
  comprehensive analysis. We included a detailed discussion of
  these limitations in Section 6 to provide a balanced assessment of
  our work.

Reviewer 2 Report

Comments and Suggestions for Authors

This article is comprehensive enough showing the use of an interesting algorithm for a collaborative service execution under edge computing facilities. However, there is a big issue which deserves the authors to think about the possibility of structural modification as follow.

The authors have mentioned the "heterogenous computing environment" and also the "timely responses to events" as well as "resource-constrained IoT". All these favors the use of 'Events driven processing" rather than "Service oriented processing". The major differences between the two are the former one is loosely coupled while the latter is decoupled which is ideal for heterogenous situation. Besides, the former one is synchronous while the later one is asynchronous which is ideal for resource-constrained situation. In fact, event driven architecture has a big improvement in the past few years, e.g. complex event processing engine. Should the authors consider using event rather than services to handle such complicated and time-critical environment?

Comments on the Quality of English Language

Just few careless grammatical mistakes found.

Author Response

We would like to thank the reviewer for the insightful comments and
suggestions regarding our article.

We acknowledge the importance of considering alternative processing
models, especially in scenarios characterized by heterogeneity and
resource constraints. Event-driven architecture indeed offers several
advantages, including loose coupling and asynchronous processing,
which may be well-suited for the challenges posed by edge computing
environments.

We will carefully evaluate the feasibility and potential benefits of
transitioning from a service-oriented processing approach to an
event-driven one. We recognize the significance of timely responses to
events and the potential improvements that event-driven architecture,
with its recent advancements, could bring to our proposed solution.

We will explore the implications of this suggestion further and
consider whether structural modifications to our approach are
warranted to better align with the unique requirements of
heterogeneous and resource-constrained IoT environments.

These concerns will be addressed in future work, as outlined in
Section 6 of the paper.

Once again, we thank the reviewer for their valuable feedback and look
forward to incorporating their suggestions into our future work.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The paper can be accepted for publication since the authors have addressed all the concerns of the paper.