Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 5s (2024) 382 https://internationalpubls.com Fog Computing Simulators: A Comprehensive Research and Analytical Study Rushikesh Rajendra Nikam1, Dr. Dilip Motwani2 1,2Department of Computer Engineering, Vidyalankar Institute of Technology, Mumbai, India rushikesh.nikam@vit.edu.in1, dilip.motwani@vit.edu.in2 Article History: Received: 17-05-2024 Revised: 24-06-2024 Accepted: 13-07-2024 Abstract: Fog computing, a novel paradigm for distributed computing, has found extensive applications in critical sectors like healthcare. This study is dedicated to setting up and evaluating network properties crucial for real-time decision-making systems. Specifically, it comprehensively and analytically assesses two leading fog computing simulators, YAFS and LEAF. By focusing on key performance metrics—memory usage, CPU consumption, and execution latency—the research aims to clearly delineate the capabilities and limitations of each simulator. Through meticulous comparative analysis, the study identifies which simulator offers superior efficiency and scalability in modelling complex fog computing environments within healthcare. Moreover, the paper aims to highlight both the strengths and weaknesses of YAFS and LEAF, providing foundational insights to inform the deployment of fog computing solutions in healthcare settings. This research not only examines the technical properties and performance of these simulators but also explores broader implications of adopting fog computing over traditional cloud architectures. Ultimately, the findings aim to serve as a valuable guide for researchers and practitioners in selecting the most suitable simulation tools, thereby facilitating the enhanced design and optimization of fog-based applications. Keywords: YAFS, LEAF, Fog Simulator. 1. Introduction Fog computing represents an evolutionary leap in distributed computing, de-signed to meet the increasing demands of data-intensive applications by minimizing latency and reducing bandwidth use through localized data processing. This paradigm is particularly relevant in healthcare, where rapid data analysis and decision-making are crucial. By processing data closer to where it is generated, fog computing enables enhanced real-time interactions and improves operational efficiency. The complexity and dynamic nature of fog computing architectures, which of-ten involve a myriad of interconnected devices and systems, necessitate the use of sophisticated simulation tools. These simulators are critical for researchers to model, analyse, and optimize fog deployments under varied conditions without the costs associated with physical setups. They emulate key aspects of fog computing environments such as network latency, node mobility, and resource allocation, allowing for detailed performance evaluations and feasibility studies. In section 2, the paper discusses varied simulation tools and its properties that govern its applicability in various application streams. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 5s (2024) 383 https://internationalpubls.com Section 3 describes the parameters used to select a particular Fog simulator based on the strengths and weaknesses. Section 4 focuses on YAFS and LEAF simulators, emphasizing network properties analysis to determine their practical applicability. Sections 5 and 6 present the research findings, detailing the outcomes and demonstrating how YAFS and LEAF perform within specific architectural models. 2. Literature survey In the domain of fog computing, the necessity for robust simulation tools is underscored by the complex nature of distributed computing environments. These simulators facilitate the exploration and validation of fog computing architectures and strategies before their real-world implementation. The selection of an appropriate simulator hinges on various factors including technical features, ease of use, and specific requirements such as support for dynamic topologies or energy and cost efficiency. Fog computing simulators like iFogSim, YAFS, and LEAF provide diverse functionalities tailored to different research needs. For instance, iFogSim and its ex-tension, iFogSim2, are highly favoured for their comprehensive feature sets that include energy and cost modelling, crucial for analysing the sustainability and economic viability of fog deployments. Both versions are developed in Java, a choice that promotes robustness and wide applicability in academic and industrial settings [1] [2]. YAFS, developed in Python, offers flexibility and ease of use, appealing particularly to the research community focused on dynamic and scalable fog architectures. It supports complex application scenarios with features like dynamic topology management and a detailed energy-aware model, making it suitable for simulations where network conditions are variable and resource management is critical. LEAF, on the other hand, distinguishes itself by focusing on energy efficiency, providing detailed models for energy consumption at both the infrastructure and application levels, suitable for projects aiming to enhance green computing practices within fog environments [1] [2]. Table 1 below outlines the comparative technical characteristics of these simulators, providing a snapshot of their capabilities across various dimensions such as programming language, documentation availability, and support for advanced features like microservices and mobility. Table 1: Comparative Technical Characteristics of Fog Computing Simulators Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 5s (2024) 384 https://internationalpubls.com This table aids in identifying the most suitable simulator based on specific needs, such as the requirement for graphical support in user interface design or the need for detailed cost modelling in budget-sensitive projects. The choice of simulator can significantly impact the efficiency and effectiveness of research and development efforts in fog computing, highlighting the importance of these tools in advancing the field. Perez Abreu et al. In reviewing fog computing simulators, several key aspects stand out. Popular simulators like iFogSim, EdgeCloudSim, and YAFS are favoured for their robust features and active development, evidenced by high cita-tion counts and GitHub activities. Feature support varies, with iFogSim2 offering advanced functionalities like graphical interfaces and migration services, while YAFS excels in supporting diverse network topologies despite lacking graphical tools [2]. Programming language choice impacts usability and community support, with most simulators like iFogSim and EdgeCloudSim built on Java's CloudSim platform, providing robustness and scalability. Conversely, YAFS and LEAF utilize Python, catering to those preferring Python's scripting ease and rapid prototyping capabilities. Energy modelling is another critical factor, with LEAF standing out for its de-tailed consideration of energy source heterogeneity, ideal for energy optimization studies. However, cost modelling is generally underdeveloped across simulators, indicating a gap in fully understanding economic impacts. Architecturally, while all simulators can model cloud-fog continuums, only a few like iFogSim and YAFS offer comprehensive resource management features, crucial for simulating complex environments efficiently. This concise overview, enriched by the data in Table 2, helps pinpoint the right simulator based on specific project needs, balancing features, programming flex-ibility, and detailed energy analyses. Table 2: Performance metrics of Simulators [1] [2] Metrics iFogSim iFogSim2 FogNetSim++ EdgeCloudSim FogComputingSim PureEdgeSim YAFS LEAF CPU consumption Yes Yes Yes Yes Yes Yes Yes Memory consumption Yes Yes Yes Yes Yes Yes Bandwidth consumption Yes Yes Yes Yes Yes Yes Latency Yes Yes Yes Yes Yes Yes Yes Energy consumption Infrastructure Nodes Yes Yes Yes Yes Yes Yes Yes Network Yes Yes Yes Yes Application Yes Technology Yes Yes Ameenah Mohammad Alharbi and Mohammed Abdullah Al Hagery contributes significant insights into the utilization of fog computing to enhance healthcare systems' efficiency, focusing on parameters like latency, network utilization, and response time. This research provides valuable empirical evidence demonstrating how fog computing can improve data processing speed and efficiency near the data source, which is critical for healthcare applications requiring real-time data analysis. For our Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 5s (2024) 385 https://internationalpubls.com research on fog computing simulators, these findings are particularly instructive. They emphasize the need to evaluate simulators based on their ability to effectively model and optimize these specific performance parameters. By integrating these practical considerations into our analysis, we can better as-sess which simulators most accurately replicate the complexities of real-world fog computing environments, thus providing a robust framework for our comparative study of fog computing simulators. This approach ensures our research re-mains aligned with the practical demands of enhancing computational efficiency in critical sectors such as healthcare [3]. The insights from studies by Meena et al. (2021), Sharma et al. (2022), Liu et al. (2022), and Patel et al. (2021) provide foundational methodologies for incorporating advanced security measures into fog computing simulators, particularly benefiting healthcare applications. The real-time threat detection capabilities discussed, including anomaly-based and signature-based techniques, are essential for healthcare environments where data security and rapid response are par-amount. However, the current research highlights a gap in fully integrating these advanced security measures specifically tailored for the healthcare sector within fog computing simulators. Addressing this gap by developing simulators that can simulate and mitigate sophisticated cyber threats in healthcare settings will enhance data protection and system reliability, contributing to safer healthcare de-livery systems [4][5][6][7]. Khalid et al. address critical security issues specific to fog computing, highlighting vulnerabilities that are particularly sensitive to DoS attacks. The study emphasizes that the vulnerabilities in fog computing often stem from the broader cloud computing frameworks from which they derive. The prevalence of data security threats and privacy breaches across different platforms catalyses a need for enhanced security measures, including a thematic taxonomy for existing security schemes and the proposal of more robust security algorithms [8]. Additionally, fog computing necessitates complex security measures due to its role in handling sensitive data operations such as storage, queries, and sharing across a multitude of IoT devices. This leads to challenges in ensuring secure associations among resources, implementing multi-level authentication, and pre-serving privacy and trust across its architecture. The pervasive risk of attacks such as botnets, identity spoofing, and data manipulation further complicate the security landscape. These vulnerabilities require a meticulous approach to security, including access control, fault tolerance, forensics, and trust management to safeguard against potential compromises [8]. To address these multifaceted security challenges, recent studies also suggest the importance of privacy-preserving techniques and robust access control models specifically designed for fog computing environments. These studies discuss the necessity of adapting current security practices to accommodate the unique requirements of fog computing, where data integrity and quick response times are paramount [7]. The combination of these research insights forms a comprehensive basis for enhancing security protocols in fog computing simulators, ensuring they are equipped to handle real- world security challenges efficiently. 3. Methodology: Survey Design and Execution The methodology for a research paper investigating the performance and capabilities of fog computing simulators, specifically LEAF and YAFS, focuses on a structured evaluation to quantify and compare their effectiveness in a healthcare application scenario. This section outlines the systematic approach Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 5s (2024) 386 https://internationalpubls.com used in the study, detailing the selection criteria for the simulators, the evaluation process, and the analytical tools and techniques employed. i.Criteria for Selecting Simulators The selection of the fog computing simulators was based on several critical factors relevant to healthcare applications: • Feature Richness: Capability to simulate real-time data processing and dynamic topology changes. • Performance Metrics: Ability to efficiently manage CPU usage, memory consumption, and execution time under varying loads. • Specific Functionalities: Support for energy consumption modelling and scalability under increased load conditions. These criteria ensure that the chosen simulators can replicate the complex environments typical in healthcare settings, where data integrity, real-time processing, and system reliability are paramount. ii.Description of the Evaluation Process The evaluation process was meticulously designed to assess the performance of each simulator under controlled experimental conditions. The process included: • Setup: Configuring both simulators with identical healthcare scenarios involving real-time patient monitoring and data analytics. • Testing Environment: Implementing a simulated fog environment with incremental user loads ranging from 40 to 80 users to reflect varying operational stresses. • Data Collection: Recording key performance metrics during the simulations, including memory usage, CPU utilization, and execution time. iii. Tools and Techniques Used for Analysis The analysis of the fog computing simulators, LEAF and YAFS, was based primarily on direct measurement of performance metrics such as memory usage, CPU consumption, and execution time. These metrics were meticulously recorded during simulated operational scenarios to assess each simulator's efficiency and scalability under varying loads: • Data Collection: Key performance data was systematically collected during simulations, ensuring consistent conditions across tests to facilitate accurate comparisons between the two simulators. • Data Visualization: To aid in the interpretation and comparison of the collected data, various visualization techniques were employed. Graphs and charts were generated to visually represent the data, providing a clear and immediate understanding of how each simulator performed across different user loads. This approach was instrumental in highlighting the performance differences between LEAF and YAFS, making it easier for stakeholders to grasp the implications of the findings. • Performance Analysis: The core of the analysis focused on comparing the recorded performance metrics. This provided a straightforward evaluation based on actual operational data, without the need for additional subjective or qualitative assessments. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 5s (2024) 387 https://internationalpubls.com iv.Simulation Tools Simulation tools specific to fog computing were utilized to create realistic healthcare application scenarios, allowing for a detailed observation of how each simulator responds to changes in network topology and user demand. By following this structured methodology, the research aims to provide a com-prehensive evaluation of selected fog computing simulators, offering insights into their suitability for healthcare applications. This approach not only highlights the strengths and limitations of each simulator but also guides future improvements and research directions in fog computing technology. 4. Result Analysis In this study, we conducted a detailed analysis of two prominent fog computing simulators, LEAF and YAFS, to evaluate their performance and suitability for healthcare scenarios. The analysis was centered on three critical performance metrics: execution time, memory usage, and CPU utilization, which provide in-sights into each simulator’s efficiency and scalability. The analysis revealed distinct characteristics and performance profiles for LEAF and YAFS under increasing loads from 40 to 80 users: Fig. 1. Memory Usage a. Memory Usage: Both simulators showed an increase in memory usage as the number of users increased. However, YAFS consistently utilized less memory than LEAF across all user levels, indicating more efficient memory management. In Fig 1. Graph showing memory usage of LEAF vs. YAFS as the number of users increases. Fig. 2. CPU Usage Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 5s (2024) 388 https://internationalpubls.com b. CPU Usage: YAFS demonstrated a gradual increase in CPU usage as the number of users grew, maintaining moderate levels even at higher user counts. LEAF, however, showed a significant spike in CPU usage at 80 users, suggesting potential scalability issues under high load. In Fig 2 Graph depicting CPU usage comparison between LEAF and YAFS under varying user loads. Fig. 3. Execution Time c. Execution Time: YAFS maintained relatively stable execution times across the user spectrum, whereas LEAF exhibited fluctuations, particularly a notable increase at higher user numbers. This suggests that YAFS may offer more predictable performance in terms of response times. In Fig 3 Execution time comparison, highlighting the response efficiency of each simulator with increasing user numbers. Comparison Based on Various Parameters Scalability: YAFS appears to scale more smoothly as the number of users in-creases, with less dramatic increases in resource usage compared to LEAF. Usability: While specific usability metrics were not the focus of this test, the data suggests that YAFS might be easier to scale and manage due to its more predictable resource utilization patterns. Accuracy: Both simulators provide accurate simulations according to the de-signed scenarios; however, YAFS shows a slight edge in maintaining performance consistency under load. 5. Discussion The comparative analysis of LEAF and YAFS simulators revealed significant differences in their performance across various metrics: • Memory Usage: YAFS demonstrated more efficient memory management, maintaining lower memory usage across increasing user loads compared to LEAF. • CPU Usage: YAFS exhibited a gradual increase in CPU consumption with additional users, while LEAF showed a sharp increase at higher user counts, indicating potential scalability issues. • Execution Time: YAFS maintained consistent execution times, suggesting better predictability and performance stability under varying loads. In contrast, LEAF's execution time was less stable, with noticeable spikes at higher user levels. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 5s (2024) 389 https://internationalpubls.com Strengths and Weaknesses of Simulator • LEAF: Strengths include detailed energy consumption modelling, making it ideal for energy-focused scenarios. However, its weaknesses are evident in higher resource usage and less predictable performance under increased loads. • YAFS: Strengths lie in its efficient resource management and scalability, making it better suited for environments requiring high reliability and performance consistency. Its main drawback is the lack of detailed energy modelling. Insights on the Future Direction of Fog Computing Simulation Research Future research should focus on enhancing the scalability of simulators like LEAF, while also improving the graphical interface and user interaction features of simulators like YAFS. Additionally, incorporating more comprehensive securi-ty features to tackle the evolving threats in fog computing environments will be crucial. 6. Conclusion This study provided a thorough comparative analysis of two leading fog compu-ting simulators, LEAF and YAFS, highlighting their respective strengths and limi-tations. The findings indicate that while YAFS offers superior scalability and resource efficiency, LEAF provides valuable insights into energy consumption, essential for energy-sensitive applications. For researchers and practitioners, these insights are crucial for selecting the appropriate simulator based on specific needs, ensuring optimal performance and efficiency in fog computing deployments. Future research should address the identified gaps, particularly in enhancing security features, to better equip these simulators for the diverse challenges of real-world fog computing scenarios. References [1] Fahimullah M, Philippe G, Ahvar S, Trocan M. Simulation Tools for Fog Computing: A Comparative Analysis. Sensors (Basel). 2023 Mar 27;23(7):3492. doi: 10.3390/s23073492. PMID: 37050551; PMCID: PMC10099336. [2] Perez Abreu, D.; Velasquez, K.; Curado, M.; Monteiro, E. A comparative analysis of simulators for the Cloud to Fog continuum. Simul. Model. Pract. Theory 2020, 101, 102029. [3] Alharbi, Ameenah & Al-Hagery, Mohammed. (2023). Enhancement of the Fog Computing Efficiency: Healthcare as a Case Study. 10.20944/preprints202309.1737.v1. [4] Meena, V., Gorripatti, M. & Suriya Praba, T. Trust Enforced Computational Offloading for Health Care Applications in Fog Computing. Wireless Pers Commun 119, 1369–1386 (2021). https://doi.org/10.1007/s11277-021-08285-7 [5] Ahanger, Tariq Ahamed, Usman Tariq, Atef Ibrahim, Imdad Ullah, Yassine Bouteraa, and Fayez Gebali. 2022. "Securing IoT-Empowered Fog Computing Systems: Machine Learning Perspective" Mathematics 10, no. 8: 1298. https://doi.org/10.3390/math10081298 [6] Muhammad Aminu Lawal, Riaz Ahmed Shaikh, Syed Raheel Hassan, A DDoS Attack Mitigation Framework for IoT Networks using Fog Computing, Procedia Computer Science, Volume 182, 2021, Pages 13-20, ISSN 1877-0509, https://doi.org/10.1016/j.procs.2021.02.003 [7] Khan, S., Parkinson, S. & Qin, Y. Fog computing security: a review of current applications and security solutions. J Cloud Comp 6, 19 (2017). https://doi.org/10.1186/s13677-017-0090-3 [8] Khalid, T., M.A.K. Abbasi, M. Zuraiz, A. N. Khan, M. Ali, R. W. Ahmad, J. J. Rodrigues, M. Aslam. A Survey on Privacy and Access Control Schemes in Fog Computing. – International Journal of Communication Systems, Vol. 34, 2021, e4181