Automatically Expanding User-Management System for Massive Users in the Cloud Platform

Round 1

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

The paper is interesting. There are some grammatical errors(e.g.-, when a user want to share data to other users). The authors are required to correct all the errors. What is the physical reason behind optimising the Weighted Load Balancing algorithm for the improvement of scheduling efficiency? Explain in details how three categories of user information methods could be applied to the large scale cloud platform to implement the synchronization of massive users.

Comments on the Quality of English Language

Grammatical errors to be corrected.

Author Response

We checked the paper again. All grammatical errors we found had been corrected.

In section 3.1

Load balancing algorithms in Nginx Among these load balancing algorithms, the weighted load balancing algorithm is more flexible and scalable for the cloud platform. All the Nginx worker nodes could be given different weights depending on their server specifications to realize load balancing. However, the unchanged weights for worker nodes were proved limited for load balancing. So we chose the weighted load balancing algorithm to optimize by adjusting the weights of worker nodes.

In section 2.2.

The module of users automated expanding was composed of three parts, that is, user information synchronization module, user identification and authentication module, and user registration and request module. For a large scale cloud platform, it is indispensable to synchronize user information and authentication between two types of user management system. So we would rewrite some classes and interfaces of user information synchronization. In the meantime, the original function of user identification and authentication within the cloud platform is insufficient to manage massive users and huge amounts of data. Some of important methods for user identification and authentication were deficient, including user role mapping, user group dividing, and user privilege creating. Therefore, these three categories of user information methods are requisite for expanding users account and synchronizing users authority automated.

Author Response File: Author Response.pdf

Reviewer 2 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

The authors have made significant improvements by the comments given in the previous round of the review. The paper can now be published.

Author Response

Thanks a lot for your review.

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

1. The abstract lacks clarity, and the proposed scheme is not clearly articulated. 
2. To address this, the author should incorporate a table within the existing work to highlight its weaknesses, aiding in the identification of research gaps. Compliance with this requirement is crucial. Furthermore, the authors must provide a lucid and comprehensive explanation of the paper's objectives.
3. A noticeable omission in the paper is a comparison with existing schemes. The authors introduce their attack generation scheme without benchmarking it against established methods. It is imperative to conduct a thorough comparison between the proposed scheme and other state-of-the-art approaches to validate its efficacy.
4. The novelty and contribution of the paper are deemed low. The author should explicitly underscore the significant contributions made in the paper to enhance its overall impact.
5. Several figures in the paper are challenging to comprehend, hindering a clear understanding of their purpose. A concerted effort should be made to improve the clarity of these figures for the benefit of the readers.
6. The conclusion should focus on scientific findings and should not be a mere modified version of the abstract. It should provide a concise summary of the key outcomes and their implications.
7. The author should define the notation upon its first occurrence and refrain from repeatedly defining the notation throughout the paper, promoting consistency and clarity in the presentation of the research.

Comments on the Quality of English Language

N/A

Author Response

We have revised the abstract as below.

Cloud computing has become one of the key technologies for big data processing and analytics. User management on cloud platforms is a growing challenge, as the number of users and the complexity of systems increase. The cloud services that provided by major cloud service providers include Alibaba Cloud and Huawei Cloud are no more than resource of computing and storage, while the relationship between the cloud platform and the huge amounts of user registration is deficient. When the users want to use the computing and storage resource provided by cloud platform, they need to register user information first and synchronize the authentication information between the cloud resource user and data user. Especially for the large-scale cloud platform, which usually have tens of thousands to a few hundred thousand of registered users. Registering users and synchronizing the user authentication would be very time consuming. This article hypothesizes that a scaling user management system for authorization synchronization can improve the efficiency and scalability of user management on cloud platforms. As the numbers of total registered customers and concurrent online customers were ever-increasing, a significant problem is how to keep the time of user registration and authentication synchronization acceptable for users. A scale-out automated expanding user management is presented in this article for authorization synchronization across cloud platforms. With the help of automated expanding user management, data accessing authority and cloud resource utilizing authority were all allocated automated as the user role changing. This user management system is proved effective through optimizing the weighted load balancing algorithms in Nginx.

Based on all above mentioned research and analysis, we can draw the following conclusion that there was no multi-user management system automatic synchronization in past research, as shown in Table 1.

Table 1. A comparison of some system architecture

No.	System architecture	Load balance of workload	Multi-user management
1	Infrastructure of multiple data centers(MDC)	An open source scheduler was introduced to make load balance	No multi-user management system automatic synchronization within all the architecture
2	Hierarchical architecture of EcoMultiCloud	An efficient management of the workload was included.
3	Major cloud service providers, such as Alibaba Cloud and Huawei Cloud	Some open source and commercial scheduling software within cloud service providers
4	OpenStack open source cloud	Some open source scheduling software were integrated in OpenStack

Therefore, this paper would reestablish a multi-user management system to break through the obstacle between cloud resource and application data (remote sensing data and other scientific data).

Besides, it was impractical to add each user manual as the number of users grows. Especially for our cloud platform, it is calculated that the number of users had been reached five thousand at present, and is expected to grow with time passed. An automated expanding multi-user management system could definitely improve the efficiency of cloud platform resource utilization, but how to construct the system that covered user authorization management and user organization relationship was extremely difficult. The user management system of OPCP was initialized based on user’s architecture of Stack 6.5.1 that was belonged to Huawei private cloud. As a matter of fact, Stack 6.5.1 of Huawei private cloud was derived from OpenStack architecture. The framework of user organization within Stack 6.5.1 was already in use, so we would make an innovation on expanding plenty of cloud users automatically based on Stack 6.5.1.

In Section 4.3. Comparison with other schemes, we added the descriptions as below.

As shown in Table 1, there was no multi-user management system that synchronization between two types of users automatically in past research works. So, we selected two previous architectures in contrast with the automated expanding user management system proposed in OPCP.

Table 4. Time of user registration and authentication synchronization within three types of system architecture

No.	System architecture	Time of user registration (s)	Time of user authentication synchronization (s)	Total time (s)
1	OpenStack Private Cloud Platform (OPCP)	6	3	9
2	Infrastructure of multiple data centers(MDC)	7	10	17
3	Huawei Cloud platform	6	12	18

The time of user registration in OPCP is broadly in line with another two system architecture. User registration by filling and submitting a user form took 6 seconds approximately. In the OPCP, user authentication synchronization can be completed by the automated expanding user management system. In the MDC and Huawei Cloud platform, the resource user authentication can be synchronized with data user manually. Because the synchronization consists of two distinct phases within MDC and Huawei Cloud platform, time of user registration and authentication synchronization is more time-consuming. The total time of user registration and authentication synchronization in the cloud platform could be marked as benchmarking. In comparison, this benchmarking of the automated expanding user management system in OPCP is generally more effective.

In the meantime, the limitations of the proposed system is that we only completed user registration and authentication synchronization between two categories of user management system. For three or more types of user’s authentication synchronization, one user authority should be synchronized to other users simultaneously.

Therefore, in order to realize massive user’s registration and authentication synchronization among multiple different user management system, we would study the database synchronization to avoid the authentication information inconsistency.

 

Author Response File: Author Response.pdf

Reviewer 4 Report (New Reviewer)

Comments and Suggestions for Authors

OVERVIEW: Managing users on cloud platforms is a growing challenge as the number of users and the complexity of systems increase. An expanding user management system is needed to automate the authorization registration and synchronization process, making it more efficient and scalable.

 

User management system for authorization synchronization across cloud platforms. The system uses weighted load-balancing algorithms to optimize performance based on Huawei's Stack 6.5.1 user architecture.

 

The system was evaluated on a private cloud platform with five thousand users. The results showed that the system can record and synchronize user authorization efficiently and efficiently.

 

User management system for authorization synchronization on cloud platforms. The system is designed to be efficient and scalable and was evaluated on a private cloud platform with five thousand users.

 

Of this paper is that a scaling user management system for authorization synchronization can improve the efficiency and scalability of user management on cloud platforms.

 

 

IMPROVEMENT POINTS:

 

 

SUMMARY: I found the summary a little confusing, but I could identify a brief explanation of what had already been studied, the object, a gap or problem, and the results. However, I could not identify anything about the project methodology. It remained to identify the type of algorithm used to optimize user management. This algorithm is cited in the conclusions (Weighted Load balancing). This would complete the understanding of the article.

 

Suggestions for improving the summary:

1. A sentence summarizing the problem in the article, such as “User management on cloud platforms is a growing challenge, as the number of users and the complexity of systems increase.”
2. A sentence that summarizes the article's objective, such as “This article presents a scale-out user management system for authorization synchronization across cloud platforms.”
3. A sentence summarizes the article's hypothesis, such as, “This article hypothesizes that a scaling user management system for authorization synchronization can improve the efficiency and scalability of user management on cloud platforms.”

 

INTRODUCTION: The current introduction to the article provides an overview of the topic but could be expanded to provide more context and theoretical foundation.

 

Some specific suggestions for improving the introduction include:

 

·         Provide a more in-depth discussion of the challenges of managing users on cloud platforms.

·         Review relevant literature on user management systems and authorization synchronization.

·         Provide a more detailed discussion of the proposed system architecture.

 

 

DISCUSSIONS: The current discussion section of the article provides a brief discussion of the results but could be expanded to discuss the following questions:

 

·         The limitations of the proposed system.

·         The benefits of the proposed system compared to other systems.

·         The implications of the proposed system for the practice of user management on cloud platforms.

·         Future research areas that could be explored based on the work presented in the article.

 

 

CONCLUSION: It is well described and links with the project's objective stipulated in the summary.

Author Response

 We have revised the abstract as below.

Cloud computing has become one of the key technologies for big data processing and analytics. User management on cloud platforms is a growing challenge, as the number of users and the complexity of systems increase. The cloud services that provided by major cloud service providers include Alibaba Cloud and Huawei Cloud are no more than resource of computing and storage, while the relationship between the cloud platform and the huge amounts of user registration is deficient. When the users want to use the computing and storage resource provided by cloud platform, they need to register user information first and synchronize the authentication information between the cloud resource user and data user. Especially for the large-scale cloud platform, which usually have tens of thousands to a few hundred thousand of registered users. Registering users and synchronizing the user authentication would be very time consuming. This article hypothesizes that a scaling user management system for authorization synchronization can improve the efficiency and scalability of user management on cloud platforms. As the numbers of total registered customers and concurrent online customers were ever-increasing, a significant problem is how to keep the time of user registration and authentication synchronization acceptable for users. A scale-out automated expanding user management is presented in this article for authorization synchronization across cloud platforms. With the help of automated expanding user management, data accessing authority and cloud resource utilizing authority were all allocated automated as the user role changing. This user management system is proved effective through optimizing the weighted load balancing algorithms in Nginx.

Some literatures were being investigated and added within the introduction.

A three-dimensional role based user management model is proposed. The three-dimensional role is defined as a vector composed of authority, scope, and per-mission time. This user management model based on the three-dimensional role can satisfy requirements of modern application systems and large scale systems. Consid-ering that users participate in an enterprise system as a particular identity, it results in that roles of different users have different authorities, and each authority is valid in a relevant scope during its permission time [4]. In a geospatial hybrid cloud platform based on multi-sourced computing and model resources, a user management module was proposed to manage the accounts for different types of end users. There were several categories of users including cloud resource contributors, cloud resource con-sumers, and cloud administrators to be managed in this user management module [5].

At the same time, we deleted some literatures uncorrelated.

Aiming at these types of data-driven problems, numerous researchers had been carry-ing out plenty of big data studies and applications. A method was proposed to use massive Big Earth Data to explore changes in snowmelt over the Antarctic ice sheet. The results shows that the abrupt changes in melt conditions linked to temperature changes over the Antarctic ice sheet were observed within the time series [4].A few-shot aircraft detection method was proposed with a feature scale selection pyramid and proposal contrastive learning for satellite videos. An evaluation of large-scale experimental data showed that the method makes full use of the advantages of the two-stage fine-tuning strategy and the characteristics of satellite video to enhance the few-shot detection performance [5]

In order to facilitate application deployment on the cloud, a conception of Cloud native technology was proposed to customize tools for cloud platform specifically. Such as microservices, containers, DevOps, and cloud native database are all belong to the cloud native technologies system. A converged cloud native infrastructure based Kubernetes was proposed in [14]. Experimental results showed that the converged cloud native infrastructure had the characteristic of scalability and reliability.

In Section 1 introduction,

Based on all above mentioned research and analysis, we can draw the following conclusion that there was no multi-user management system automatic synchronization in past research, as shown in Table 1.

Table 1. A comparison of some system architecture

No.	System architecture	Load balance of workload	Multi-user management
1	Infrastructure of multiple data centers(MDC)	An open source scheduler was introduced to make load balance	No multi-user management system automatic synchronization within all the architecture
2	Hierarchical architecture of EcoMultiCloud	An efficient management of the workload was included.
3	Major cloud service providers, such as Alibaba Cloud and Huawei Cloud	Some open source and commercial scheduling software within cloud service providers
4	OpenStack open source cloud	Some open source scheduling software were integrated in OpenStack

Therefore, this paper would reestablish a multi-user management system to break through the obstacle between cloud resource and application data (remote sensing data and other scientific data).

Besides, it was impractical to add each user manual as the number of users grows. Especially for our cloud platform, it is calculated that the number of users had been reached five thousand at present, and is expected to grow with time passed. An automated expanding multi-user management system could definitely improve the efficiency of cloud platform resource utilization, but how to construct the system that covered user authorization management and user organization relationship was extremely difficult. The user management system of OPCP was initialized based on user’s architecture of Stack 6.5.1 that was belonged to Huawei private cloud. As a matter of fact, Stack 6.5.1 of Huawei private cloud was derived from OpenStack architecture. The framework of user organization within Stack 6.5.1 was already in use, so we would make an innovation on expanding plenty of cloud users automatically based on Stack 6.5.1.

 

In Section 4.3. Comparison with other schemes, we added the descriptions as below.

As shown in Table 1, there was no multi-user management system that synchronization between two types of users automatically in past research works. So, we selected two previous architectures in contrast with the automated expanding user management system proposed in OPCP.

Table 4. Time of user registration and authentication synchronization within three types of system architecture

No.	System architecture	Time of user registration (s)	Time of user authentication synchronization (s)	Total time (s)
1	OpenStack Private Cloud Platform (OPCP)	6	3	9
2	Infrastructure of multiple data centers(MDC)	7	10	17
3	Huawei Cloud platform	6	12	18

The time of user registration in OPCP is broadly in line with another two system architecture. User registration by filling and submitting a user form took 6 seconds approximately. In the OPCP, user authentication synchronization can be completed by the automated expanding user management system. In the MDC and Huawei Cloud platform, the resource user authentication can be synchronized with data user manually. Because the synchronization consists of two distinct phases within MDC and Huawei Cloud platform, time of user registration and authentication synchronization is more time-consuming. The total time of user registration and authentication synchronization in the cloud platform could be marked as benchmarking. In comparison, this benchmarking of the automated expanding user management system in OPCP is generally more effective.

In the meantime, the limitations of the proposed system is that we only completed user registration and authentication synchronization between two categories of user management system. For three or more types of user’s authentication synchronization, one user authority should be synchronized to other users simultaneously.

We revised the conclusion as below.

Because of the complexity of two types of users existed on cloud platform, a single user management system cannot meet all the requirements of cloud platform resource users and data users. Especially for a cloud platform that possesses high volumes of user, it is impossible to configure user authentication between different user management systems manually.

In this study, a scale-out automated expanding user management system is proposed to achieve synchronization between the resource users of Virtual Data Center (VDC) in HuaWei Cloud Stack 6.5.1 and the register users from data management system. In order to achieve a better load balancing efficiency, we introduced Ngnix as the scheduler and optimized the Weighted Load Balancing algorithm to improve the scheduling efficiency. The main contributions in this paper is the three categories of user information modules could be applied to other large scale cloud platform to implement the synchronization of massive users. The optimized Weighted Load Balancing algorithm is also valuable for massive users registration concurrently based on limited cloud resources.

In the end, the automated expanding user management system for massive users on cloud platform in this study also leaves something to be desired for more complicated user authority. For instance, if users want to share their data to other multiple users, the database synchronization would be studied to avoid the authentication information inconsistency. This is the further research issues of the dynamical expansion of user management system.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

The paper can be accepted

Author Response

Thanks for your approval !

Reviewer 4 Report (New Reviewer)

Comments and Suggestions for Authors

Scale-out user management system for authorization synchronization on cloud platforms aims to improve the efficiency and scalability of user management, addressing the growing challenge of managing users on cloud platforms.

The reviewer is grateful for the changes already made in the first version of the article and asks the authors to address further points for improvement in the scientific article that deals with user management in a multi-cloud environment and highlights its benefits in service management. The article is innovative due to its prototyped approach. It has excellent publication potential when it addresses the limitation of cloud services companies in not providing adequate guarantees for their services (QoS).

 

The text between lines 161 to 169 clarifies the construction tool for the constructed prototype's controlled environment, and Table 1 gives a good view of the research gap.

 

The improvements are listed below:

 

Point 1. Summary: Please consider inserting and reviewing the summary, as there are missing items. The summary must be both concise and complete (approximately 300 words). The authors did not focus on reviewing this item sufficiently.

In summary, it is essential to have at least the following items: how the bibliography review was carried out (what was the review's research methodology?); what gap was identified; the objectives of the study; note that the general objective described in the study summary is different from those mentioned in paragraphs 80 and 81; which is also different from the text between lines 135 to 138 (what was the fundamental objective of the research? see item that informs... "this paper proposed an architecture"; "This paper introduced a way to overcome current Cloud service limitations.. ."); the research hypothesis in a very concise way; the preliminary results found; what was completed (?); the limitations of the present study; and, what future studies can be carried out?

 

Point 2. The reviewer is grateful for including some companies, such as HUAWEI and Alibaba, in the summary. However, since the services were only tested in the HUAWEI cloud, they should only keep the company, as if they were to mention them all, they would have to include AWS (mentioned in lines 91 and 92 with "AWS Aerospace & Satellite program..."), Oracle Cloud and others.

 

It is essential to highlight that the new prototype was only mentioned with the environment built on the HUAWEI platform for companies with cloud computing. This is described in the prototype construction methodology.

 

Point 2.1. Line 120. The same problem occurs with Microsoft Azure and Oracle Cloud cloud services. It is best to cite all researched sources. The reviewer appreciates that the authors work in a multi-cloud environment to test the prototype.

 

Point 3. Cite all references to the themes within the item "Literature review". This point requested for expansion by the reviewer was not fully addressed: the authors focus a lot on the innovative solution but little on supporting the literature review with more authors who have previously researched the topics. This article deserves and needs a robust review of the bibliographical references used.

 

Point 4. Lines 98 to 102. What is the centralized solution that is the authors' reference? To quote. "Performance analysis has proven that the hierarchical approach achieves nearly the same quantitative results as a reference centralized solution, but offers better functionalities in terms of flexibility."

 

Point 5. Lines 108 to 110. However, these two papers should have mentioned the user management system among geographically distributed data centers and cloud platform nodes.

 

Once again, the authors do not mention which two scientific articles they refer to ("...in these two papers"). They appear to refer to quote number [15] from the text, but this needs to be clarified.

 

Point 6. Lines 185 to 187. Do the authors consider the innovation developed as incremental or disruptive? Cite a bibliography to support the answer.

 

Point 7. At the end of the "Introduction," which contains the "Literature Review," the reviewer requests the insertion of a paragraph defining the complete structure of the scientific article.

 

Example: This article is structured in sections, as follows: Summary with keywords; (1) Introduction with literature review; (2) Materials and methods describing the methodological approach, dividing it into XXXXXX stages: bibliographic references and choice of research participants with the criteria adopted (reference descriptors, period of articles and chosen databases of research); method of conception and development of the work and the techniques applied for XXXXXX (development of YYYYYY, ZZZZZZ, ..., and DDDDDD); development of a usable software; (3) Discussions based on insights and review of the application with an emphasis on the experience of expert users; identification of improvements, research limitations and future studies; (4) Conclusions and suggestions; (5) Bibliographic references.

 

Point 8. In each figure (all of them), the authors must identify who created each figure in the description. "Figure 1. The architecture of the user management system."

Examples: "Title" by authors? "Title", adapted to CCCCCC? This is a standard for MDPI publications.

 

Point 9. Refer to the text before each figure or table, introducing what information the figure or table will provide. Explain below each figure or table in at least one paragraph what each image contributed to the present research work.

 

Point 10. The images in Figures 3 and 4 are illegible; please check the possibility of subdividing them into blocks to make the texts inserted within the images explicit. Authors may need to reconstruct them with graphic design tools or editor software.

 

Point 11. Lines 177 to 179. "...reached five thousand at present, and is expected to grow with time passed." and lines 423 and 424. "...number of users had been reached four thousand at present, and is expected to grow with time passed.".

 

What is the correct number of users? No matter how many estimates the authors made, both must be coherent. I believe fatigue may have occurred due to the extensive review required in round 1 of the review.

 

Point 12. The authors' conclusions need to respond to the research objectives, and the authors need to do this sufficiently. It is better to ask the Editors for a deadline extension if the authors cannot answer all the reviewers' questions within the deadline stipulated by the magazine.

 

 

Final considerations: There was a significant improvement in the scientific article, but due to the extensive corrections made by the reviewers, there are still points to be adjusted, and therefore, the article needs at least one more round of adjustments. A refined English review also becomes appropriate and desirable at the end of the entire article.

Comments on the Quality of English Language

A suggestion for the new version with English adjustments only in the actual summary:

"Cloud computing has become one of the critical technologies for big data processing and analytics. User management on cloud platforms is a growing challenge as the number of users and the complexity of systems increase. The cloud services provided by major cloud service providers, including Alibaba Cloud and Huawei Cloud, are no more than a computing and storage resource. In contrast, the relationship between the cloud platform and the massive user registration must be improved. When the users want to use the computing and storage resource provided by the cloud platform, they need to register user information first and synchronize the authentication information between the cloud resource user and the data user, especially for the large-scale cloud platform, which usually has tens of thousands to a few hundred thousand of registered users. Registering users and synchronizing the user authentication would be very time-consuming. This article hypothesizes that a scaling user management system for authorization synchronization can improve the efficiency and scalability of user management on cloud platforms. As the number of total registered and concurrent online customers was ever-increasing, a significant problem was how to keep the time of user registration and authentication synchronization time acceptable for users. A scale-out automated expanding user management is presented in this article for authorization synchronization across cloud platforms. With the help of automated expanding user management, data accessing authority and cloud resource utilizing authority were all allocated automated as the user role changed. This user management system has proved effective by optimizing Nginx's weighted load balancing algorithms."

Author Response

1.

We rewrote the summary to cover all the contents above.

Cloud computing has become one of the key technologies for big data processing and analytics. User management on cloud platforms is a growing challenge, as the number of users and the complexity of systems increase. In light of the user management system that provided by major cloud service providers could not manage multiple types of user systems, this article proposed a scale-out automated expanding user management for authorization synchronization can im-prove the efficiency and scalability of user management on cloud platforms. Three modules for user automated expanding was designed and implemented to synchronize the authentication information from the cloud platform resource user to data processing user. Additionally, an op-timized dynamic weighted load balancing algorithm in Nginx was presented in this article that would adjust the weight according to the load information such as CPU and memory usage, and a better load balance could be achieved. The effectiveness of the proposed user management system is substantiated through comparison with two existing infrastructures, including multi-ple data centers and Huawei cloud platform. The experimental results validate the finding that the scale-out automated expanding user management across Huawei cloud platform could effec-tively synchronize data accessing authority with cloud resource utilizing authority. And the op-timized Weighted Load Balancing algorithm is also valuable for massive users registration concurrently based on limited cloud resources. In the future, this scale-out user management system would be applied to other cloud platform and extended by database synchronization to satisfy the needs of data sharing among multiple types of users belongs to different cloud plat-form.

2.

As a matter of fact, the prototype proposed in this paper was derived from Huawei cloud platform. In the future, we would apply this architecture and the methods of user authority synchronization to Alibaba cloud platform.

So we revised the introduction and conclusion.

3.

We removed some literatures that are less relevant to this paper.

4.

The centralized solution is a simple and efficient approach. User registration and user authority synchronization could be executed within one master worker node.

5.

We revised the two cited literatures about user management system and the quote number.

Performance analysis has proven that the hierarchical approach achieves nearly the same quantitative results as a centralized architecture. User management system was con-structed for job submitting and monitoring across distributed data centers. Authority for data access was not able to be synchronized with resource access [9] ……. It follows that user management system among different geographically distributed data centers and cloud platform nodes was only for job submission in [9] and [10].

6.

The innovation of this paper was designed based on methods provided by Stack cloud platform,

The framework of user management within Stack 6.5.1 was already in use, so we would make an expansion to accommodate plenty of cloud users automatically based on Stack 6.5.1 [19].

7.

This article is structured in sections, as follows: Summary with keywords; (1) Introduction with literature review; (2) An architecture of the user management was proposed for implementation of user authority synchronization, dividing it into three parts: user information synchronization module; user identification and authentication module; user registration and request module; (3) A dynamic weighted load balancing algorithm was presented to adjust the weight of worker node for improving scheduling efficiency; (4) Experimental environment and comparison with other methodologies were illustrated; (5) Conclusions; (6) Bibliographic references.

8.

We supplemented the description in the Figure 1, Figure 2, and Figure 6. I am not sure whether the description meet requirement.

9.

We explained the effect of figures and tables in this paper.

10.

We redrew Figure 3 and Figure 4 with Microsoft vision to make sure these two images are legible

11

I am sorry that this is a clerical error. It is calculated that the number of users had been reached four thousands at present

12.

We revised the conclusions as follow.

This research aimed to address the challenge encountered while managing multiple types of user management systems across cloud platform. In light of the complexity of two types of users existed on cloud platform, it is difficult to implement synchronization of user authority information. This challenge would significantly hinder the application of cloud platform.

Author Response File: Author Response.pdf

Round 3

Reviewer 4 Report (New Reviewer)

Comments and Suggestions for Authors

Congratulations to the authors on the significant improvement of the article. The reviewer is happy to see their commitment to refining the work, which is almost ready for final publication.

 

Point 1. Please adjust the figures "Figure 2. Basic synchronization flows from HuaWei Cloud Stack users to user authority management is created and adapted by Wanfeng Zhang." and "Figure 5. The interface class of user registration and request" so that they have the same visual quality as figures 3 and 4.

 

Point 2. Inform the methodology item of the selection criteria for the target audience, the selection criteria for the theoretical framework, and the scientific databases researched. Please check whether the keywords in the abstract match the research descriptors and inform which period the articles were researched in and in which language. Example: "Chinese" and "English".

 

Point 3. Regarding the bibliographic reference, it was expanded, but the authors barely cited MDPI authors and journals; I ask that you redo the search in the MDPI search tools to identify articles that can corroborate the authors' references, include references of the magazines where you want to publish.

 

Point 4. "Data Availability Statement: Not applicable.". Is research data public or private data? If they are private, they can be inserted as non-disclosable attached data, available only to MDPI. The editor can indicate the magazine, an observation that this data is restricted to access by MDPI editors; after all, a database was created for the item to carry out the tests.

 

Point 5. "Informed Consent Statement: Not applicable.". Is there any explanation why the data does not have a declaration of consent or why the authors used fictitious data? The question was already asked in a previous Cover Letter, and the authors did not respond. Suppose they were fictitious OK, as for the explanation. Otherwise, tell them how you got so many hits without telling them they would be used for research.

 

One way to meet point 4 is to include in the methodology item the criteria for selecting the target audience and that they gave their consent when carrying out the research. Please evaluate the best explanation, but it needs to be in the text.

 

In addition to the items above, I have no other new observations.

Author Response

1. We redrew the Figure 2 and Figure 5 so that they could be identified clearly.

2. We revised the keywords as follows.

Keywords: Automated expanding, Users management system, Cloud computing platform, User authority synchronization.

This paper would be valuable for researchers who may concern about user management system across cloud platform. All the works in this article were carried out in year 2022 and in Chinese.

3. We searched the MDPI magazines again and found that studies on user management system based on cloud platform were not many. We chose a bibliography highly related in user management system named “A User-Centered Mobile Cloud Computing Platform for Improving Knowledge Management in Small-to-Medium Enter-prises in the Chilean Construction Industry”.

4. Research data in this paper was derived from our project. Because of the requirements of our project on the safety, data in this paper would not be generally public to all the readers. At the same time, the target audience of this paper could offer consent about their research work to obtain data in this paper.

5. Sorry about the mistake of “Informed Consent Statement”. All the data in this article is real and authorized a declaration of consent.

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Please explain why is the introduction based around the aerospace engineering. Nor in abstract, nor in the rest of the paper the connection with aerospace engineering is not explained. 

 

The need for development of dynamical expansion of user management system for massive users in the cloud platform must be explained. It would be good to explain it based on the results of a literature review. No preview of related works is given in the paper.

 

The experiment and results should be provided in more detail while in the end add the section about the implications of this work for scientific and industrial community. It remains unclear how this solution is better than the others from this field, if there are any, and this must be described.

Author Response

1. Please explain why is the introduction based around the aerospace engineering. Nor in abstract, nor in the rest of the paper the connection with aerospace engineering is not explained. 

The cloud platform that we represented in this paper was implemented upon a Chinese aerospace data processing system. The name of the aerospace data processing system was private to the public so far. So nor in abstract, nor in the rest of the paper the connection with aerospace engineering is not explained. Aiming at this question, a supplementary description was added in abstract.

2. The need for development of dynamical expansion of user management system for massive users in the cloud platform must be explained. It would be good to explain it based on the results of a literature review. No preview of related works is given in the paper.

We modified the paragraph 5 and 6 as follows.

The huge amounts of aerospace data were collected and stored on our system that constructed on the Huawei Cloud. For users access to this system, it is a urgent require-ment to invoke cloud resource and scientific data resource simultaneously. But there were no relationship for user management between computing resource and scientific data.

Besides, it was impractical to increase each user manual as the number of users grows. Especially for our cloud platform, it is estimated that the number of users had been reached several thousand at present, and is expected to grow with time passed. A dynamically expanding user management system could definitely improve the efficiency of cloud platform resource utilization.

 

3. The experiment and results should be provided in more detail while in the end add the section about the implications of this work for scientific and industrial community. It remains unclear how this solution is better than the others from this field, if there are any, and this must be described.

 

In section 4.2.

In fact, there was no other ready-made method to contrast. As a contrast, we added cloud platform users and synchronized user authentication through the ManageOne software provided by Huawei Cloud Stack 6.5.1.Because the user authentication synchronization was realized by deploying the ManageOne software manual, it should be create two types of users at first. It would spend twice time on creating users than the dynamical expansion of user management system proposed within this study.

The result reflects that the time of user registration and user authority synchroniza-tion once through the dynamical expansion of user management system was less than 10 seconds. While it would take 13 seconds to achieve the same process by the ManageOne software. More importantly, there existed five thousand users in our cloud platform at present. In the future, this platform would keep on running at least ten years and deal with many new users registration every day.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper proposes a flexible extensible user management system that can connect cloud resource users with data users.When one user type is added, another user type is automatically increased. Data access rights and cloud resource usage rights are both assigned when customer roles change. This user management system has been proven to be effective through its cloud-based application.The authors claim that it can be used to solve the problem of disconnection between cloud resource users and data users, and the capacity to carry the number of clients can exceed tens of thousands.They claim that when tens of thousands of clients are online simultaneously, user registration and authentication synchronization times will remain stable.

 While dealing with user information synchronization module, user identification and authentication module, and user registration and request module new methods were rewritten. What is the physical basis of that?

  While introducing  Ngnix as the load-balancing algorithm how the authors ensured that the computing server was running at an equalized level?

 Authors are required to clearly state the problem statement and modify the abstract and the introduction part. Also, some more current references are required; particularly of 2023.

 State clearly the approximations considered.

What are the limitations?

 

 

Author Response

1. While dealing with user information synchronization module, user identification and authentication module, and user registration and request module new methods were rewritten. What is the physical basis of that?

 

As showed in Figure 1, the class UserInfoController, CommonController, and UserRegisterController were all inherited from the class BaseController and belonged to the user information process tier. There were several private and public methods within the class BaseController. However, these methods were insufficient for realizing the three types of functions. One option is to rewrite the method within the class BaseController. Another option is to rewrite some functions inherited from the class BaseController. In comparison, the latter option is succinct and transplantable.

 

2. While introducing Ngnix as the load-balancing algorithm how the authors ensured that the computing server was running at an equalized level?

 

The Weighted Load Balancing algorithm was chosen to assign user registration requests to servers. Meanwhile, we provided three virtual machines with different specifications in the cloud platform. The virtual machine with high specification would be endowed with a higher weight so as to handling more user registration requests. For example, the virtual machine with 64 threads (virtual cores) could be allocated 32 requests, while the virtual machine with 32 threads would be allocated 16 requests. This weighted load balancing algorithm would make sure that the two virtual machines running at an equalized level with 50% CPU utilization.

 

3. Authors are required to clearly state the problem statement and modify the abstract and the introduction part. Also, some more current references are required; particularly of 2023.

 State clearly the approximations considered.

What are the limitations?

We modified the paragraph 5 and 6 as follows.

The huge amounts of aerospace data were collected and stored on our system that constructed on the Huawei Cloud. For users access to this system, it is a urgent requirement to invoke cloud resource and scientific data resource simultaneously. But there were no relationship for user management between computing resource and scientific data.

Besides, it was impractical to increase each user manual as the number of users grows. Especially for our cloud platform, it estimates that the number of users had been reached several thousand at present, and is expected to grow with time passed. A dynamically expanding user management system could definitely improve the efficiency of cloud platform resource utilization.

The fourteenth reference was added.

In section 5,we gived a summery for the limitations.

In the end, the dynamical expansion of user management system for massive users in the cloud platform in this study also leaves something to be desired for more complicated user authority. For instance, when a user want to share data to other users, the data access authority after sharing would not be synchronized to the next user. This is the further research issues of the dynamical expansion of user management system.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

A Dynamically Expanding User Management System for Massive Users in the Cloud Platform

Comments:

‘”This paper proposed a dynamically expanding users management system, which could associate the cloud resource users with the data users.” How to test the performance is acceptable?

“the time of user registration and authentication synchronization would still remain stable” how to justify this statement?

“numerous researchers had been carrying out plenty of big data applications [3-10]” The authors should clearly identify the work done from each reference. At present, the presentation is poor.

“Under normal conditions, the user’s management system of cloud platform and user’s utility application were independent respectively [17-22].” Improve English.

“In this research, the expandability and reliability of user’s management system are two most important points to consider.” What are expandability and reliability?

“Therefore, this paper would reestablish a user management system to break through the obstacle between cloud resource and scientific data.” Why scientific data? What happen to other types of data?

What are the original contributions? The structure of the paper is not clear.

The caption for Figure 3 is very long.

 “Among all the base classes provided by OPCP, the class BaseController was the most significant for user’s management system.” Why?

“As showed in Figure 1, the class UserInfoController” But the reviewer cannot find UserInfoController in the figure.

“Experiment results showed that the achievement time of massive user’s registration and authentication was successful.” It is not clear under what conditions did the experiment carry out. Also, what are the assessment criteria for results? The paper looks like written as a report. The substance is very little.

More results are needed and scentific explanations should be included. 

Comments on the Quality of English Language

A Dynamically Expanding User Management System for Massive Users in the Cloud Platform

Comments:

‘”This paper proposed a dynamically expanding users management system, which could associate the cloud resource users with the data users.” How to test the performance is acceptable?

“the time of user registration and authentication synchronization would still remain stable” how to justify this statement?

“numerous researchers had been carrying out plenty of big data applications [3-10]” The authors should clearly identify the work done from each reference. At present, the presentation is poor.

“Under normal conditions, the user’s management system of cloud platform and user’s utility application were independent respectively [17-22].” Improve English.

“In this research, the expandability and reliability of user’s management system are two most important points to consider.” What are expandability and reliability?

“Therefore, this paper would reestablish a user management system to break through the obstacle between cloud resource and scientific data.” Why scientific data? What happen to other types of data?

What are the original contributions? The structure of the paper is not clear.

The caption for Figure 3 is very long.

 “Among all the base classes provided by OPCP, the class BaseController was the most significant for user’s management system.” Why?

“As showed in Figure 1, the class UserInfoController” But the reviewer cannot find UserInfoController in the figure.

“Experiment results showed that the achievement time of massive user’s registration and authentication was successful.” It is not clear under what conditions did the experiment carry out. Also, what are the assessment criteria for results? The paper looks like written as a report. The substance is very little.

 

Author Response

Review 3:

1. “This paper proposed a dynamically expanding users management system, which could associate the cloud resource users with the data users.” How to test the performance is acceptable?

As shown in Figure 8, we had been tested the time of user registration and authentication synchronization. It is acceptable to register one user within 8 seconds.

2. “The time of user registration and authentication synchronization would still remain stable” how to justify this statement?

As shown in Figure 8, the experimental result shows that increasing the number of users to ten thousand only need extra two seconds for user registration. For this project, several thousands of users were more common. So registering a new user would spend about six seconds in most cases.

3. “numerous researchers had been carrying out plenty of big data applications [3-10]” The authors should clearly identify the work done from each reference. At present, the presentation is poor.

In paragraph 1, we have modified the literature research as follows.

A method was proposed to use massive Big Earth Data to explore changes in snowmelt over the Antarctic ice sheet. The results shows that the abrupt changes in melt conditions linked to temperature changes over the Antarctic ice sheet were observed within the time series[3].

A few-shot aircraft detection method was proposed with a feature scale selection pyramid and proposal contrastive learning for satellite videos. An evaluation of large-scale experimental data showed that the method makes full use of the advantages of the two-stage fine-tuning strategy and the characteristics of satellite video to enhance the few-shot detection performance [4].

Some researchers designed an infrastructure of multiple data centers (MDC) for managing and processing massive remote sensing images by introducing access security and information service. Using this MDC, they succeeded in working out the concrete problems regarding procedures in processing applications collaboratively and transfer the massive remote sensing dataset fast and with stable cross-MDC [5].

In order to dealing with time-series adaptive change detection based on massive remote sensing data, some researchers proposed the similarity-measurement-based deep transfer learning model (SDTL-TSACD), time-Series Classification approach based on Change Detection (TSCCD), and high-accuracy land-cover classification method with the Informer network. Change detection results using the large scale datasets showed that these algorithms performed well in both accuracy and efficiency[6-8].

 

4. “Under normal conditions, the user’s management system of cloud platform and user’s utility application were independent respectively [17-22].” Improve English.

Under normal circumstances, the user’s management system and application software were two different systems which ran on cloud platform [17-22].

5. “In this research, the expandability and reliability of user’s management system are two most important points to consider.” What are expandability and reliability?

For some workload reasons, the expandability and reliability of user’s management system were not considered and verified in this study, so this sentence should be deleted.

6. “Therefore, this paper would reestablish a user management system to break through the obstacle between cloud resource and scientific data.” Why scientific data? What happen to other types of data?

Therefore, this paper would reestablish a user management system to break through the obstacle between cloud resource and application data (aerospace data, remote sensing data, and other scientific data).

7. What are the original contributions? The structure of the paper is not clear.

This paper proposed a dynamically expanding user management system which was used to improve efficiency of user registration and user authority synchronization for massive users. This user management system was designed and realized in previous Commercial cloud platform.

8. The caption for Figure 3 is very long.

The caption for Figure 3 had been shorted.

9. “Among all the base classes provided by OPCP, the class BaseController was the most significant for user’s management system.” Why?

As showed in Figure 1, the class UserInfoController, CommonController, and UserRegisterController were all inherited from the class BaseController. There were several private and public methods within the class BaseController. So the class BaseController was the most significant for user’s management system. However, these methods were insufficient for realizing the three types of functions. One option is to rewrite the method within the class BaseController. Another option is to rewrite some functions inherited from the class BaseController. In comparison, the latter option is succinct and transplantable.

10. “As showed in Figure 1, the class UserInfoController” But the reviewer cannot find UserInfoController in the figure.

In figure 1, the class UserInfoController is located in User Information Processing Tier.

11. “Experiment results showed that the achievement time of massive user’s registration and authentication was successful.” It is not clear under what conditions did the experiment carry out. Also, what are the assessment criteria for results? The paper looks like written as a report. The substance is very little.

In section 4.1, the dynamically expanding users management system was designed and implemented based on Stack 6.5.1 of Huawei private cloud platform. This cloud platform was both the experiment environment and the practical running environment. It is estimated that the number of users had been reached four thousands at present, and is expected to grow with time passed.

As shown in Table 1, the maximum number of threads of virtual machine was 64. The physical host computing server was assembled with Intel 6100 series CPUs that sup-port the actual maximum frequency of 2666 MHz.

In section 4.2, we modified the description as follows.

In order to meet the requirement of time effectiveness for users accessing our cloud platform, we invited two hundred participants to answer questions randomly from thousands of users. According to the user feedback, when they completed the process of user registration and user authority synchronization within 10 seconds, they hand a responsive user experience.

12. More results are needed and scentific explanations should be included.

In section 4.2, we added the description as follows.

As a contrast, we added cloud platform users and synchronized user authentication through the ManageOne software provided by Huawei Cloud Stack 6.5.1.Because the user authentication synchronization was realized by deploying the ManageOne software manual, it should be create two types of users at first. It would spend twice time on creating users than the dynamical expansion of user management system proposed within this study.

The result reflects that the time of user registration and user authority synchroniza-tion once through the dynamical expansion of user management system was less than 10 seconds. While it would take 13 seconds to achieve the same process by the ManageOne software. More importantly, there existed five thousand users in our cloud platform at present. In the future, this platform would keep on running at least ten years and deal with many new users registration every day.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have provided all the needed corrections, therefore the article can now be published.

Reviewer 2 Report

Comments and Suggestions for Authors

may be accepted