Optimizing Cloud Service Composition with Cuckoo Optimization Algorithm for Enhanced Resource Allocation and Energy Efficiency

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript adapts COA for cloud service composition aiming at multi-objective enhancements, validated via CloudSim against PSO/ACO. However, it is marred by vague adaptations, superficial surveys, unvalidated assumptions, and confined simulations, demanding rejection pending major revisions for precision, depth, and broader validations.

The abstract proposes applying the Cuckoo Optimization Algorithm (COA) to cloud service composition for optimizing resource allocation, load balancing, task scheduling, and energy efficiency, with simulations in CloudSim showing superiority over PSO and ACO in metrics like execution time and power consumption. The abstract inadequately defines service composition mappings to COA elements (e.g., high/low-demand services as eggs) and lacks baseline parameter details, undermining its methodological clarity and risking overstated comparative advantages.

The Introduction highlights cloud computing challenges in dynamic workloads and variability, justifying bio-inspired optimization for service composition. The introduction recycles generic issues without empirical evidence on COA's novelty in clouds or critiques of prior metaheuristics, revealing superficial problem framing and failure to position against recent hybrid algorithms.

Section 2 surveys optimization techniques in cloud scheduling, categorizing metaheuristics and noting gaps in energy-aware composition. This section is cursory and repetitive, omitting quantitative syntheses of cited COA variants or integrations with ML for adaptive tuning, exposing insufficient depth.

Section 3 details COA adaptation where high-demand services mimic host eggs and low-demand ones cuckoo eggs, with fitness based on QoS metrics. Methodological mappings contain ambiguous pseudocode and undefined hyperparameters, neglecting convergence analyses and handling of premature stagnation common in COA, indicating procedural laxity and scalability neglect in large-scale clouds.

Section 4  describes CloudSim 5.0 simulations with varied tasks, VMs, and datacenters, evaluating metrics like SLA violations and VM utilization. Setup relies on arbitrary configurations without sensitivity tests or real traces (e.g., Google clusters), pointing to simulated biases and limited generalizability beyond controlled scenarios.

Section 5 reports COA's outperformance in all metrics, attributing to efficient search spaces. Results lack statistical validations (e.g., ANOVA) or error bars, with discussions recycling data without addressing outliers or practical deployment hurdles, highlighting weak empirical rigor.

The Conclusion reaffirms COA's benefits and suggests hybrid extensions. The conclusion evades limitations such as computational intensity in real-time environments, repeating claims without forward-looking critiques, betraying absent self-examination essential.

Author Response

We sincerely thank the reviewers for their valuable and constructive comments. All remarks have been carefully addressed and incorporated to enhance the quality and clarity of the manuscript. Detailed responses are provided below.

- The abstract proposes applying the Cuckoo Optimization Algorithm (COA) to cloud service composition for optimizing resource allocation, load balancing, task scheduling, and energy efficiency, with simulations in CloudSim showing superiority over PSO and ACO in metrics like execution time and power consumption. The abstract inadequately defines service composition mappings to COA elements (e.g., high/low-demand services as eggs) and lacks baseline parameter details, undermining its methodological clarity and risking overstated comparative advantages.

Response :
We would like to thank the reviewer for this valuable comment. To enhance methodological clarity, we revised the abstract to better define as follows:

Revised Abstract:

The composition of cloud services plays a vital role in optimizing resource allocation, load balancing, task scheduling, and energy management. However, it remains a significant challenge due to the dynamic nature of workloads and the variability of resource demands, where addressing these challenges is essential for ensuring seamless service delivery. This research investigates the implementation of the Cuckoo Optimization Algorithm (COA) in a cloud computing environment to optimize service composition. In the proposed approach, each service is treated as an egg, where high-demand services represent the host’s original eggs, while low-demand services represent the Cuckoo bird’s eggs that compete for the same resources. This implementation enables the algorithm to balance workloads dynamically and allocate resources efficiently while optimizing load balancing, task scheduling, cost reduction, processing and response times, system stability, and energy management. The simulations were conducted using CloudSim 5.0, and the results were compared with the Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) algorithms across key performance metrics. Experimental results clearly demonstrate that COA outperforms both PSO and ACO across all evaluated metrics. COA achieved higher efficiency in task scheduling, dynamic load balancing, and energy-aware resource allocation. It consistently maintained lower operational costs, reduced SLA violations, and achieved superior task completion and VM utilization rates. These findings underscore COA’s potential as a robust and scalable approach for optimizing cloud service composition in dynamic and resource-constrained environments.

 

- The Introduction highlights cloud computing challenges in dynamic workloads and variability, justifying bio-inspired optimization for service composition. The introduction recycles generic issues without empirical evidence on COA's novelty in clouds or critiques of prior metaheuristics, revealing superficial problem framing and failure to position against recent hybrid algorithms.

Response:
Thank you for your valuable comment. We believe that this concern has been addressed in the Literature Review section, as follows:
"According to previous studies, cloud computing still faces numerous challenges, such challenges include task scheduling, resource allocation, load balancing, consolidation, VMs placement, service composition, energy consumption, and replication, where most of the challenges are classified as NP-hard problems. In this study, researchers integrate multiple performance metrics related to the cloud computing environment, including Execution Time, Power Consumption, Cost, The SLA violations, Latency, Response Time, load balancing efficiency, VM Utilization, and task completion rate, providing a more comprehensive evaluation. Moreover, as shown in Table 1, no prior study has addressed all these performance metrics within a single framework. This highlights a significant gap in the literature and justifies the need for a more integrative and holistic optimization approach, as proposed in this study. "

To further clarify our contribution, we added the following statement:

Unlike prior metaheuristic and hybrid approaches that focused on isolated objectives, this study’s model unifies multiple metrics within one optimization framework, positioning it as a more comprehensive solution for cloud service composition.

- Section 2 surveys optimization techniques in cloud scheduling, categorizing metaheuristics and noting gaps in energy-aware composition. This section is cursory and repetitive, omitting quantitative syntheses of cited COA variants or integrations with ML for adaptive tuning, exposing insufficient depth.

Response:

The following paragraphs were added in the Literature Review section:

However, no prior COA-based approach has applied similar adaptive mechanisms for real-time adjustment in cloud environments, highlighting a research gap in COA–ML integration for self-optimized service composition.

- Section 3 details COA adaptation where high-demand services mimic host eggs and low-demand ones cuckoo eggs, with fitness based on QoS metrics. Methodological mappings contain ambiguous pseudocode and undefined hyperparameters, neglecting convergence analyses and handling of premature stagnation common in COA, indicating procedural laxity and scalability neglect in large-scale clouds.

Response: thank you for your comment, we added the following paragraph to the future work section in the research conclusion.

Future research may focus on tuning and optimization of COA hyperparameters, convergence analysis, and strategies to further study the proposed ACO implementation in large-scale or real-time cloud environments. This would enhance its robustness, scalability, and practical applicability, in addition to prevent premature stagnation.

- Section 4  describes CloudSim 5.0 simulations with varied tasks, VMs, and datacenters, evaluating metrics like SLA violations and VM utilization. Setup relies on arbitrary configurations without sensitivity tests or real traces (e.g., Google clusters), pointing to simulated biases and limited generalizability beyond controlled scenarios.

Response: Thank you for your valuable observation. We agree that the study’s simulation setup using CloudSim 5.0 represents a controlled environment, which may not fully reflect real-world workloads. This limitation has already been acknowledged in the conclusion, where we noted that the simulations may not capture the full complexity and unpredictability of actual cloud infrastructures, thereby addressing the generalizability concern raised.

"Future studies may also benefit from incorporating accurate real-time workload prediction to improve resource allocation and scalability, particularly during fluctuating demand scenarios. Finally, deploying and evaluating the proposed approach on commercial cloud platforms such as Amazon Web Services (AWS), Microsoft Azure, will provide deeper insights into its practical effectiveness under operational constraints. However, it is important to understand that the COA might involve significant computational intensity in real-time implementation or large-scale services' environment, which could impact its scalability and responsiveness. such limitations provide a clearer understanding of the COA algorithm applicability and can be considered as a potential direction for future research."

 

-  Section 5 reports COA's outperformance in all metrics, attributing to efficient search spaces. Results lack statistical validations (e.g., ANOVA) or error bars, with discussions recycling data without addressing outliers or practical deployment hurdles, highlighting weak empirical rigor.

Response: Thank you for your valuable comment, the experiments in this study were conducted using CloudSim simulator, which provides a deterministic and repeatable simulation environment rather than a stochastic empirical one. Therefore, traditional statistical significance tests such as ANOVA were not applied. So, we added this paragraph to the conclusion section according to the reviewer's comment:

It is important to understand that the COA might involve significant computational intensity in real-time implementation or large-scale services' environment, which could impact its scalability and responsiveness. such limitation provides a clearer understanding of the COA algorithm applicability and can be considered as a potential direction for future research.

 

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript investigates an important issue related to the composition of clous services with Cuckoo Optimization Algorithm (COA). The introduction states the purpose of the study, namely, an implementation of COA in cloud computing environment to optimize service composition. The references are relevant and complete with a few exceptions. The manuscript is written clearly and it is well organized. In the Introduction, the authors stress on some of the Key Performance Indicators of cloud computing and on the works of other authors, and clearly declare the study objectives and key contributions. Section 2. Literature review analysis the related works in the area of using the Machine learning method in cloud computing with focus on service composition. In the next section, the methodology used is explained. The CloudSim (5.0) simulation environment is used to implement the COA, where the following stages are explained: the parameter initialization, the initial population generation, the generation of a new solution, the fitness function, the discovery and replacement strategy, the nest selection and elitism, the task scheduling strategy, and the energy management model. Section 4, presents and discusses the simulation results. The conclusion summarizes the contribution and outlines the directions of future work. The authors explain the significance of the research. The main contribution is the development of cloud service composition model based on COA and demonstration of its capabilities in comparison to the Particle Swarm Optimization (PSO) and the Ant Colony Optimization (ACO) algorithms.

Some remarks that could improve the research presentation are as follows:

1. Lines 7, 8 must be removed.
2. It is not appropriate to use abbreviations as keywords. It would be better to replace them with machine learning. The keywords need to be arranged semantically, namely Cloud computing, service, composition, Optimization machine learning.
3. The abbreviation SLA must be given in full in the abstract.
4. The reference list is not formatted as to journal style - the references are formatted in IEEE style. The MDPI requirements are e.g., first is the family and that the first and second author name, names of the journal must be abbreviated, etc.
5. Not all of the references are relevant, e.g. “Electronic Governmental Cloud (e-GCloud)” does not have room in the context of manuscript topic [Ref. [6], [7].
6. The first time the abbreviations VM and SLA (lines 93, 94) are not defined, but later VM and SLA are defined multiple times (lines 106, 118, 188). PSO and ACO are defined twice.
7. Compare->Compares (line 161)
8. Load Balancing Efficiency, resource allocation, load balancing, Energy Management, Service Level, Execution Time (all in lowercase)
9. Table 1 must not be split
10. The figure captures are not formatted correctly.
11. In Figure 1 must not contain references, they have to be explained in the text.
12. The process of the COA algorithm must be explained.
13. The table of abbreviations is missing.
14. SLA violations->The SLA violations (line11), Response Time-> The response time (line120), etc.
15. The Section 2. Literature Review must begin in the previous page.
16. Service composition->service composition (line 196)
17. The abbreviation PSO is defined several times (lines 85, 86, 177, 244
18. Figure (2) Figure 2 ->Figure 2 (line 282)
19. Few words about the steps of the COA algorithm would make the manuscript even better
20. The font size in text and in some equations is different (line 382, 385, 421, etc.)
21. Figure (6) represents the Fitness Function pseudocode-> Figure 6. The Fitness Function pseudocode
22. The Figure 13 contains numerical result but the runs seem to be not enough in order to conclude for any kind of trend as far as 10 runs only are statistically negligible. If the trend is declining and we have 100 runs how low would the execution time go in milliseconds? The same is probably applicable to Figures 14, 15, 16, 17, 18, 19.

Comments on the Quality of English Language

There are some typos.

Author Response

We sincerely thank the reviewers for their valuable and constructive comments. All remarks have been carefully addressed and incorporated to enhance the quality and clarity of the manuscript. Detailed responses are provided below.

1. Lines 7, 8 must be removed.

Response: Removed

1. It is not appropriate to use abbreviations as keywords. It would be better to replace them with machine learning. The keywords need to be arranged semantically, namely Cloud computing, service, composition, Optimization machine learning.

Response: Modified

Keywords: Cloud Computing, Cuckoo Optimization Algorithm, Particle Swarm Optimization, Ant Colony Optimization,  CloudSim,  Services Composition

 

1. The abbreviation SLA must be given in full in the abstract.

Response: Modified

The results were compared with the Particle Swarm Optimization (PSO) and the Ant Colony Optimization (ACO) algorithms across key performance metrics including: Execution Time, Power Consumption, Load Balancing Efficiency, Service Level Agreement (SLA) Violations, Cost, Latency, Response Time, Virtual Machine (VM) Utilization, and Task Completion Rate.

1. The reference list is not formatted as to journal style - the references are formatted in IEEE style. The MDPI requirements are e.g., first is the family and that the first and second author name, names of the journal must be abbreviated, etc.

Response: we updated the reference list using the MDPI reference format. 

1. Not all of the references are relevant, e.g. “Electronic Governmental Cloud (e-GCloud)” does not have room in the context of manuscript topic [Ref. [6], [7].

Response: reference (Electronic Governmental Cloud (e-GCloud)) was removed, and the References List was updated.

1. The first time the abbreviations VM and SLA (lines 93, 94) are not defined, but later VM and SLA are defined multiple times (lines 106, 118, 188). PSO and ACO are defined twice.

Response: VM and SLA are defined in the abstract, so we remove all the abbreviations, also the duplication of the PSO and ACO definitions is removed

1. Compare->Compares (line 161)

Response: The sentence starts with “Compare” because it is in the imperative form. We would use “compares” only if the sentence had a subject performing the action, which does not apply in this case. However, we can modify the sentence to make it clearer, as follows:

• Compare the performance of COA with that of PSO and ACO, highlighting the effectiveness of swarm intelligence-based methods in cloud optimization

1. Load Balancing Efficiency, resource allocation, load balancing, Energy Management, Service Level, Execution Time (all in lowercase)

Response: modified to lowercase

1. Table 1 must not be split

Response: modified (in one page)

1. The figure captures are not formatted correctly.

Response: all figures' captures were modified according to the journal Format

1. In Figure 1 must not contain references, they have to be explained in the text.

12. Response: modified (references move to the previous paragraph)
13. COA is an Algorithm inspired by the brood parasitism behavior of cuckoo birds [17].  To implement the COA, we used the CloudSim (5.0) simulation environment.  Figure (1) illustrates the COA Mechanisms, including the Cuckoo Egg Placement (Lévy Flight), Fitness Function, Nest Selection and Replacement, and Elitism [17], [18], [19]. Meanwhile, Figure (2) Figure 2 shows the COA Algorithm Pseudocode.

 

 

1. The process of the COA algorithm must be explained.

Response: sections (3 Methodology explains the process in details (6 pages))

 

1. The table of abbreviations is missing.

Response:  we added the abbreviations Table to the research paper as required :

SLA	Service Level Agreement
VM	Virtual Machine
COA	Cuckoo Optimization Algorithm
PSO	Particle Swarm Optimization
ACO	Ant Colony Optimization
SaaS	Software As a Service
PaaS	Platform As a Service
IaaS	Infrastructure Asa Service
SOA	Oriented Architecture
CPU	Central Processing Unit
MCE	Multi-Cloud Environment
GA	Genetic Algorithms
OEJSR	Optimized Efficient Job Scheduling Resource
LJFP	Longest Job to Fastest Processor
MCT	Minimum Completion Time
MSF	Multistage Forward Search
SMO	Spider Monkey Optimization
AWCO	Advanced Willow Catkin Optimization
GWO	Grey Wolf Optimization
AWS	Amazon Web Services

1. SLA violations->The SLA violations (line11), Response Time-> The response time (line120), etc.

Response: Modified

1. The Section 2. Literature Review must begin in the previous page.

Response: Modified

1. Service composition->service composition (line 196)

Response: Modified

1. The abbreviation PSO is defined several times (lines 85, 86, 177, 244

Response: VM and SLA are defined in the abstract, so we remove all the abbreviations, also the duplication of the PSO and ACO definitions is removed

1. Figure (2) Figure 2 ->Figure 2 (line 282)

Response: All Figures modified according to the Journal Template

1. Few words about the steps of the COA algorithm would make the manuscript even better

Response: sections (3 Methodology explains the process in details (6 pages))

1. The font size in text and in some equations is different (line 382, 385, 421, etc.)

Response: Some equations exceeded the page margins. Accordingly, we adjusted the font size to ensure proper formatting (Eq(4) ), other equations were modified.

1. Figure (6) represents the Fitness Function pseudocode-> Figure 6. The Fitness Function pseudocode

24. Response: Modified

1. The Figure 13 contains numerical result but the runs seem to be not enough in order to conclude for any kind of trend as far as 10 runs only are statistically negligible. If the trend is declining and we have 100 runs how low would the execution time go in milliseconds? The same is probably applicable to Figures 14, 15, 16, 17, 18, 19.

Response: we appreciate your reviewer's valuable comment, in this study, the results shown in the Figure (13-19)  were based on the average of 10 independent simulation runs. This practice is the one commonly used in similar CloudSim-based studies to ensure reasonable computational COSTS while keeping the results stable. However, we acknowledge that increasing the number of runs could provide a more statistically robust representation of the performance trends. To address this concern, we have conducted additional simulations with 100 independent runs, and the updated results confirm the same declining trend in execution time and similar for the other performance metrics.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The revised manuscript is recommended for publication.