The Proposal and Validation of a Distributed Real-Time Data Management Framework Based on Edge Computing with OPC Unified Architecture and Kafka
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis manuscript describes the overall architecture of managing real-time data in the intelligent manufacturing environment using the combination of edge computing, OPC UA (Open Platform Communications Unified Architecture), and Apache Kafka. The authors develop a multi-layered architecture that supports standard data capture from multi-vendor industrial devices in a heterogeneous manner, real-time processing at the edge layer, and scalable, high-throughput data transmission with Kafka for persistent storage and computation. The proposed system includes:
- Modeling and acquisition of data from an OPC UA-based device,
- Python-based distributed edge modules (PDCs),
- Performance in production monitoring and inventory management in real time,
- A data streaming and storage infrastructure based on Kafka,
- PostgreSQL is an integrated backend with a visualized web interface.
The framework is tested via a realistic experimental testbed, a simulated six-station assembly platform with detailed performance evaluations (latency, throughput, and multi-threaded message handling). The following recommendations must be addressed.
- Some sections are awkward (e.g., “this paper... this paper...”) or repetitive, and even have inconsistent tenses. It is advisable to have a complete language edit pass.
- Despite citing a wide variety of references, there is room to discuss recent high-impact work (2021 – 2024) on edge-AI in manufacturing, TSN in OPC UA, and cloud-edge.
- Some latest references from 2022 onwards must be cited, like On AI-based predictive maintenance with edge analytics, and The Architecture of a smart factory with OPC UA on TSN.
- Although the Kafka throughput test evidence is quite satisfactory, a comparison with baseline methods ( e.g., MQTT, REST APIs ) would provide more context for the improvements.
- The testbed is based on VM-based Kafka nodes that may not adequately represent actual industrial deployments. It would be a good move to recognize this limitation in the discussion.
- Many figures' data (for example, Figures 12-17) are important but have no meaning or are not used the Python code as a figure. It is recommended to move the code to an appendix section.
- Tables (eg, Table 1 on edge vs cloud) are useful but could use more quantifiable metrics or the actual latency/throughput values from the experiments.
- Although future integration with ML is touched upon, a separate section on future directions may be useful for the manuscript, which should include, among other things, the following: Real-time anomaly detection, Integration with edge AI models, OPC UA over TSN for critical timeliness environments, and Principles of security/privacy for OPC UA-Kafka setups deployed in a distributed fashion.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThis paper proposes a multi-layer industrial data-management framework that couples OPC UA-based data collection at edge nodes with lightweight analytics and an Apache Kafka cluster plus PostgreSQL for high-throughput streaming and storage. Experiments on a six-station assembly platform demonstrate up to 51.96 MB/s average throughput and end-to-end latencies below 7 ms, indicating the approach can support real-time production monitoring in smart-factory scenarios. The manuscript is generally well written, logically structured and reports promising performance results that will interest researchers and practitioners working on Industry 4.0 data pipelines. I have however the following comments:
- The paper could include a brief comparison with MQTT options to clarify why Kafka suits the workload.
- Figure 3 outlines the PDC workflow yet the OPC UA object modeling steps remain abstract. A small example or pseudocode could improve clarity.
- Section 3.3.4 mentions a thread pool strategy but the pool size and tuning parameters are not specified. Sharing those details would help reproducibility.
- Since the message payload uses JSON the work might benefit from reporting the average message size and discussing possible compression choices such as lz4 versus gzip.
- Making the edge client source code or configuration scripts publicly available would support independent validation and wider adoption.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe article presents a well-structured examination of a timely and sensitive topic, specifically focusing on real-time data management frameworks and edge computing. It holds significant promise for contributing valuable insights to the readership of Energies. However, certain areas require careful revision before a final approval for publication can be granted. The following points warrant attention:
Firstly, the title may be somewhat perplexing for readers lacking specialized knowledge in this domain. It is advisable to minimize the use of acronyms in both the title and keywords to enhance clarity.
The abstract would benefit from a more explicit articulation of the research's motivation. It is essential to clarify the reasoning behind undertaking this study, given the importance of the research problem at hand. In addition, a comprehensive overview of the research methodology is necessary. It should address the implications of the findings: how will they bring about changes in the real world for both industry practitioners and academic scholars? What merits are associated with improving data throughput and real-time responsiveness?
Moreover, it is recommended to refrain from using lumped references like [1-2], as they obstruct the reader’s ability to cross-reference your claims with other sources.
The paragraph beginning with "The efficiency of traditional workshop production lines…" (line 64) lacks citations, despite containing multiple verifiable statements. While it is your discretion, I suggest incorporating a citation, such as https://doi.org/10.3390/asi7020024, to substantiate the claims made in this section. Additionally, consider avoiding overly dramatic adjectives, such as “urgently,” which can detract from the professionalism of your writing.
When first introducing the OPC UA protocol, it is important to present it formally. The same applies to KAFKA and other specific terminology that may not be readily understood by a global audience.
Furthermore, it is critical to theoretically derive a clear research gap that your article aims to address. Even though you have outlined expected contributions, clarity on how they fill existing gaps in knowledge is essential. Provide evidence demonstrating that these issues have not been thoroughly explored in recent, high-impact literature and emphasize their importance in advancing the state of the art in the field.
Before concluding the introduction, formally state the article's purpose, the research methods employed, and give a succinct overview of the article's structure.
Chapter 2 should be titled "Review" and should incorporate foundational concepts necessary for readers less familiar with the research problem. In Chapter 3, titled "Materials and Methods," provide a clear and straightforward description of the methodology utilized, followed by a flowchart that outlines the steps taken (complementing your existing list with a visual representation of checkpoints and procedures to ensure the reliability of your findings). If any specific equipment was used, please include a table detailing the model, year of release (especially for software), and the manufacturer or dealer.
Avoid using third-level indentation (3.1.1) as it can disrupt the reading flow unnecessarily.
The experiment appears to be well-designed and executed—congratulations on that achievement. However, it is critical to clarify how reliability was ensured in your findings. Why should the audience have confidence that the results are consistent over time? Additionally, what measures were taken to address repeatability? After line 511, please provide the probability of a message being correctly received both with and without the enhancements introduced by your contribution.
When you state, “In order to test the feasibility…” (line 493), it is vital to provide a precise definition of "feasibility" and clearly differentiate it from reliability—the latter referring to the ability to achieve the same result in repeated experiments.
Lastly, please remove the internal captions from Figures 18 and 19, as they are redundant in light of the external captions.
The concluding chapter requires substantial enhancement. I recommend expanding it by adding a subsection titled "Implications," in which you outline the future directions of the research. This subsection should detail the prospective applications of your results from the viewpoints of both practitioners and scholars.
Best regards.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsDear Authors,
I have been invited to review your paper "A Distributed Real-Time Data Management Framework Based on Edge Computing with OPC UA and Kafka"
I have the following comments:
- Title: You might wish to clearly state that you are making a proposal for a framework or that you are testing a new framework. This shall facilitate the understanding of potential readers.
- Introduction: Your paper lacks a self-standing literature review that should lead you to identify gaps in the existing body of knowledge, which you would, in turn, seek to address fully or partly with your research. The justification and rationale of your paper are lacking; the statements contain in the Introduction do not suffice to introduce the issues at stake and the rationale of your work. The paper does not properly underline the novelty and contributions of your work -as compared to previous research undertaken in the field(-s) of interest.
- Moreover, it is difficult to know if you are quoting the work of previous authors and whether you are citing them as no references are provided in multiple parts of the text, e.g., lines 39-414143, 100-101, etc. Please make sure that whenever you introduce the results of previous research, you provide the relevant reference(-s). In case, the statements derive from your research, please make sure this is obvious and ensure it is clear to the reader how.
- While the framework you introduce might be of interest, the paper should be substantially rewritten in the form of a research article, including a fully comprehensive literature review, the identification of gaps in the existing body of knowledge, the explanation of the rationale and focus of your work, clearly indicating its novelty and contributions.
- Conclusions: please make sure you spell out the limitations of your work and based on them you put forward ideas for future research.
Thank you and best regards,
Anonymous reviewer
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors incorporated my concerns, now it is suitable for publication.
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors addressd most (but not all) my remarks and provided good rebuttals. The only thing I´d like to sign is the presence of an acronym in the title. If possible, it should be replaced by the full expression. Nonetheless, I let it at the discretion of the authors an the editor.
Reviewer 4 Report
Comments and Suggestions for AuthorsThank you for the revised version of your paper and the responses to my comments.
Best regards,
Anonymous reviewer
