CPU Resources Research Articles

Decomposition analysis of influencing factors of GPU-centric supercomputing demand: LMDI-based approach

With the introduction of AI technology, the supercomputing industry is transitioning from CPU-centric to GPU-centric, and many countries are making efforts to build new GPU-centric resources. The purpose of this paper is to discover new factors in demand management for efficient construction and operation of future national supercomputing GPU resources. Reflecting industry characteristics, we decompose the factors affecting existing CPU use into intensity effect, structure effect, and production effect indicators targeting CPU-only resources and GPU-only resources, and compare and analyze the influence of each factor. To estimate the influence of each factor, the Logarithmic Mean Divisia Index methodology was used, and annual CPU usage data from the Republic of Korea's national supercomputing center was used. As a result of the analysis, it was confirmed that CPU resources show a similar trend every year, and that the effects of the intensity and production indicators are continuously increasing. In the case of GPU resources, all indicators had an influence in the direction of increasing demand, and it was confirmed that the information/communication field was overwhelmingly showing the greatest effect.

Retrospecting Available CPU Resources: SMT-Aware Scheduling to Prevent SLA Violations in Data Centers

Retrospecting Available CPU Resources: SMT-Aware Scheduling to Prevent SLA Violations in Data Centers

Server Performance Analysis Against Distributed Denial of Service (DDoS) Attacks (Case Study: Mataram University Academic Information System)

Information technology advancement has significantly impacted Indonesia's education sector, including the University of Mataram. The university relies on the Sistem Informasi Akademik (SIA) for various academic activities. However, increasing users has led to challenges such as excessive traffic and server downtime. This research analyzes the impact of DDoS attacks on the SIA server's performance and utilizes Zabbix for server monitoring. The results show that DDoS attacks mainly burden the network traffic component, with a significant increase compared to normal conditions. CPU and RAM resources also increased but less significantly. Monitoring with Zabbix improves the efficiency of performance management, enabling real-time monitoring and easy-to-understand visualization. The findings help administrators identify the need to increase server capacity and implement DDoS detection and mitigation systems to maintain the quality and availability of academic services.

Comparison and Analysis of Scheduling Algorithms: Exploring Performance, Time, Fairness, and Applicable Scenarios

Scheduling algorithm is an important part of CPU work and plays a key role in the program system. Successful scheduling algorithms can effectively improve system efficiency and stability. This paper aims to analyze the characteristics of several scheduling algorithms comparatively through some general criteria to help determine the applicable algorithms in different situations. This paper first introduces the important role of scheduling algorithms. It then discusses and analyzes in detail the complexity of scheduling algorithms, allocation methods, average waiting time, and starvation. Finally, it draws the advantages and disadvantages of different scheduling algorithms through comparison. The types of scheduling algorithms applicable to batch systems and real-time/interactive systems respectively are concluded. The study results provide a theoretical basis for system designers to achieve the optimal utilization of CPU resources in the system. By analyzing and comparing the advantages and disadvantages of different scheduling algorithms, this paper lays a solid foundation for the future research with more scheduling methods.

Enhancing task execution: a dual-layer approach with multi-queue adaptive priority scheduling.

Efficient task execution is critical to optimize the usage of computing resources in process scheduling. Various task scheduling algorithms ensure optimized and efficient use of computing resources. This article introduces an innovative dual-layer scheduling algorithm, Multi-Queue Adaptive Priority Scheduling (MQAPS), for task execution. MQAPS features a dual-layer hierarchy with a ready queue (RQ) and a secondary queue (SQ). New tasks enter the RQ, where they are prioritized, while the SQ contains tasks that have already used computing resources at least once, with priorities below a predefined threshold. The algorithm dynamically calculates the time slice based on process priorities to ensure efficient CPU utilization. In the RQ, the task's priority level defines its prioritization, which ensures that important jobs are completed on time compared to other conventional methods where priority is fixed or no priority parameter is defined, resulting in starvation in low-priority jobs. The simulation results show that MQAPS better utilizes CPU resources and time than traditional round-robin (RR) and multi-level scheduling. The MQAPS showcases a promising scheduling technique ensuring a balanced framework for dynamic adjustment of time quantum and priority. The MQAPS algorithm demonstrated optimization, fairness, and efficiency in job scheduling.

Efficient Management of Containers for Software Defined Vehicles

Containerization technology, such as Docker, is gaining in popularity in newly established software-defined vehicle architectures (SDVA). However, executing those containers can quickly become computationally expensive in constrained environments, given the limited CPU, memory, and energy resources in the Electric Control Units (ECU) of SDVA. Consequently, the efficient management of these containers is crucial for enabling the on-demand usage of the applications in the vehicle based on the available resources while considering several constraints and priorities, including failure tolerance, security, safety, and comfort. In this article, we propose a dynamic software container management approach for constrained environments such as embedded devices/ECUs in SDVA within smart cars. To address the conflicting objectives and constraints within the vehicle, we design a novel search-based approach based on multi-objective optimization. This approach facilitates the allocation, movement, or suspension of containers between ECUs in the cluster. Collaborating with our industry partner, Ford Motor Company, we evaluate our approach using different real-world software-defined scenarios. These scenarios involve using heterogeneous clusters of ECU devices in vehicles based on real-world software containers and use-case studies from the automotive industry. The experimental results demonstrate that our scheduler outperforms existing scheduling algorithms, including the default Docker scheduler -Spread- commonly used in automotive applications. Our proposed scheduler exhibits superior performance in terms of energy and resource cost efficiency. Specifically, it achieves a 35% reduction in energy consumption in power-saving mode compared to the scheduler employed by Ford Motor Company. Additionally, our scheduler effectively distributes workload among the ECUs in the cluster, minimizing resource usage, and dynamically adjusts to the real-time requirements and constraints of the car environment. This work will serve as a fundamental building block in the automotive industry to efficiently manage software containers in smart vehicles, considering constraints and priorities in the real world.

Estimation of Mango Fruit Production Using Image Analysis and Machine Learning Algorithms

Mango production is fundamental to the agricultural economy, generating income and employment in various communities. Accurate estimation of its production optimizes the planning and logistics of harvesting; traditionally, manual methods are inefficient and prone to errors. Currently, machine learning, by handling large volumes of data, emerges as an innovative solution to enhance the precision of mango production estimation. This study presents an analysis of mango fruit detection using machine learning algorithms, specifically YOLO version 8 and Faster R-CNN. The present study employs a dataset consisting of 212 original images, annotated with a total of 9604 labels, which has been expanded to include 2449 additional images and 116,654 annotations. This significant increase in dataset size notably enhances the robustness and generalization capacity of the model. The YOLO-trained model achieves an accuracy of 96.72%, a recall of 77.4%, and an F1 Score of 86%, compared to the results of Faster R-CNN, which are 98.57%, 63.80%, and 77.46%, respectively. YOLO demonstrates greater efficiency, being faster in training, consuming less memory, and utilizing fewer CPU resources. Furthermore, this study has developed a web application with a user interface that facilitates the uploading of images from mango trees considered samples. The YOLO-trained model detects the fruits of each tree in the representative sample and uses extrapolation techniques to estimate the total number of fruits across the entire population of mango trees.

Container-Based Electronic Control Unit Virtualisation: A Paradigm Shift Towards a Centralised Automotive E/E Architecture

The past 40 years have seen automotive Electronic Control Units (ECUs) move from being purely mechanical controlled to being primarily digital controlled. While the safety of passengers and efficiency of vehicles has seen significant improvements, rising ECU numbers have resulted in increased vehicle weight, greater demands placed on power, more complex hardware and software, ad hoc methods for updating software, and subsequent increases in costs for both vehicle manufacturers and consumers. To address these issues, the research presented in this paper proposes that virtualisation technologies be applied within automotive electrical/electronic (E/E) architecture. The proposed approach is evaluated by comprehensively studying the CPU and memory resource requirements to support container-based ECU automotive functions. This comprehensive performance evaluation reveals that lightweight container virtualisation has the potential to welcome a paradigm shift in E/E architecture, promoting consolidation and enhancing the architecture by facilitating power, weight, and cost savings. Container-based virtualisation will also enable efficient and robust online dynamic software updates throughout a vehicle’s lifetime.

Intelligence data acquisition based on embedded system in Chinese cuisine cooker (CCICR V1.0)

The Chinese cooking process, intricate and diverse, encompasses a wide range of dishes, generating substantial data that often poses significant challenges for automation. To address these challenges, this paper introduces a versatile, highly reliable, user-friendly, and intelligent data recorder specifically designed for Chinese cuisine-the Chinese Cuisine Intelligent Cooker Recorder-V1.0 (CCICR-V1.0). This device is intended to record data generated during the cooking process and provide essential support for intelligent cooking robot systems. The core of CCICR-V1.0 is a STM32 microcontroller, equipped with sensors that monitor temperature, pressure, and attitude. These sensors facilitate the intelligent collection of data related to dishes, ingredients, and the movements of cookware throughout the cooking process. To ensure efficient data acquisition and storage, CCICR-V1.0 employs a multi-task data buffering storage method that effectively allocates CPU resources. NAND Flash is used as the storage medium, guaranteeing secure storage, management, and preservation of high-frequency, multi-channel data. Experimental results demonstrate the effectiveness of CCICR-V1.0 in achieving multi-channel, high-frequency data acquisition and storage in complex environments. This leads to an increase in automation efficiency of 89.9%. The weighing module demonstrates a maximum relative error of only 0.288%, while the attitude sensor experiment shows an attitude information error of just 0.22°C. Furthermore, the non-contact infrared temperature measurement module exhibits impressive performance with a maximum absolute error of 0.258∘C\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^\\circ C$$\\end{document} and a maximum relative error of 0.88%. Notably, this measurement module has achieved a cost reduction of approximately 91.7% and an accuracy improvement of around 40%. Additionally, the response time of CCICR-V1.0 is less than 3 ms. With its ability to effectively capture data related to Chinese cooking, CCICR-V1.0 holds commercial value for widespread adoption and application in the field of intelligent cooking.

A Comparative Study of CPU and GPU Power Consumption while using Open-Source and Proprietary Media Players

This study presents a comparative analysis of power consumption between open-source and proprietary media players when playing open-media format videos (.webm). As media consumption grows, energy-efficient software is critical for both environmental sustainability and device performance. Using tools like HWiNFO, key metrics such as GPU and CPU power consumption, memory usage, and efficiency were evaluated for popular open-source (e.g., VLC, Kodi) and proprietary (e.g., GOM Player, KMPlayer) players. The results reveal that open-source players generally consume less GPU power but more CPU resources, while proprietary players balance CPU and GPU usage with higher memory demands. The findings suggest that careful selection of media players can lead to significant energy savings over time, offering insights for developers and users focused on energy-efficient computing.

Efficiently Enlarging RDMA-Attached Memory with SSD

RDMA-based in-memory storage systems offer high performance but are restricted by the capacity of physical memory. In this paper, we propose TeRM to extend RDMA-attached memory with SSD. TeRM achieves fast remote access on the SSD-extended memory by eliminating page faults of RDMA NIC and CPU from the critical path. We also introduce a set of techniques to reduce the consumption of CPU and network resources. Evaluation shows that TeRM performs close to the performance of the ideal upper bound where all pages are pinned in the physical memory. Compared with existing approaches TeRM significantly improves the performance of unmodified RDMA-based storage systems, including a file system and a key-value system.

FedNIC: enhancing privacy-preserving federated learning via homomorphic encryption offload on SmartNIC

Federated learning (FL) has emerged as a promising paradigm for secure distributed machine learning model training across multiple clients or devices, enabling model training without having to share data across the clients. However, recent studies revealed that FL could be vulnerable to data leakage and reconstruction attacks even if the data itself are never shared with another client. Thus, to resolve such vulnerability and improve the privacy of all clients, a class of techniques, called privacy-preserving FL, incorporates encryption techniques, such as homomorphic encryption (HE), to encrypt and fully protect model information from being exposed to other parties. A downside to this approach is that encryption schemes like HE are very compute-intensive, often causing inefficient and excessive use of client CPU resources that can be used for other uses. To alleviate this issue, this study introduces a novel approach by leveraging smart network interface cards (SmartNICs) to offload compute-intensive HE operations of privacy-preserving FL. By employing SmartNICs as hardware accelerators, we enable efficient computation of HE while saving CPU cycles and other server resources for more critical tasks. In addition, by offloading encryption from the host to another device, the details of encryption remain secure even if the host is compromised, ultimately improving the security of the entire FL system. Given such benefits, this paper presents an FL system named FedNIC that implements the above approach, with an in-depth description of the architecture, implementation, and performance evaluations. Our experimental results demonstrate a more secure FL system with no loss in model accuracy and up to 25% in reduced host CPU cycle, but with a roughly 46% increase in total training time, showing the feasibility and tradeoffs of utilizing SmartNICs as an encryption offload device in federated learning scenarios. Finally, we illustrate promising future study and potential optimizations for a more secure and privacy-preserving federated learning system.

Impact assessment of blockchain-based improved PBFT consensus mechanism for improving supply chain efficiency and security

In modern supply chain management, it is crucial to be able to track the flow of goods in real-time and maintain data integrity. Research focuses on the transparency, security, and efficiency improvements brought about by blockchain technology, and how these improvements can help address information asymmetry issues in supply chain management and improve overall operational efficiency. At the same time, the study explored the working principle of the practical Byzantine fault-tolerant consensus mechanism, proposed relevant improvement measures, and analyzed how these improvements can be effectively applied in the scenario of supply chain information sharing to improve the stability and efficiency of the overall system. The research results show that the overall latency of improved PBFT is maintained at the millisecond level, while the latency of traditional PBFT can reach several seconds or exceed 10 seconds. In terms of CPU resource consumption, the demand for improving PBFT only shows linear growth, while the demand for traditional PBFT increases sharply. In addition, improving PBFT also has significant advantages in communication costs. In an 8-node network, the communication volume of improved PBFT is only 11.27 KB, while traditional PBFT is 43.67 KB. In terms of throughput and latency performance, the improved PBFT can handle over 16000 requests per second in a 16 node network, significantly better than the traditional PBFT’s approximately 8000 requests. Meanwhile, at 64 nodes, the improved version has a latency of 301.7 milliseconds, much lower than the traditional version’s 510.2 milliseconds. These improvements demonstrate the advantages of improving PBFT in enhancing supply chain transparency, security, and efficiency, providing effective technical support for the digitization and intelligence of supply chain management. We hope that research can promote the digitization and intelligence of supply chain management, and facilitate the efficient operation and optimization of the supply chain.

Compressing and Recovering Short-Range MEMS-Based LiDAR Point Clouds Based on Adaptive Clustered Compressive Sensing and Application to 3D Rock Fragment Surface Point Clouds.

Short-range MEMS-based (Micro Electronical Mechanical System) LiDAR provides precise point cloud datasets for rock fragment surfaces. However, there is more vibrational noise in MEMS-based LiDAR signals, which cannot guarantee that the reconstructed point cloud data are not distorted with a high compression ratio. Many studies have illustrated that wavelet-based clustered compressive sensing can improve reconstruction precision. The k-means clustering algorithm can be conveniently employed to obtain clusters; however, estimating a meaningful k value (i.e., the number of clusters) is challenging. An excessive quantity of clusters is not necessary for dense point clouds, as this leads to elevated consumption of memory and CPU resources. For sparser point clouds, fewer clusters lead to more distortions, while excessive clusters lead to more voids in reconstructed point clouds. This study proposes a local clustering method to determine a number of clusters closer to the actual number based on GMM (Gaussian Mixture Model) observation distances and density peaks. Experimental results illustrate that the estimated number of clusters is closer to the actual number in four datasets from the KEEL public repository. In point cloud compression and recovery experiments, our proposed approach compresses and recovers the Bunny and Armadillo datasets in the Stanford 3D repository; the experimental results illustrate that our proposed approach improves reconstructed point clouds' geometry and curvature similarity. Furthermore, the geometric similarity increases to 0.9 above in our complete rock fragment surface datasets after selecting a better wavelet basis for each dimension of MEMS-based LiDAR signals. In both experiments, the sparsity of signals was 0.8 and the sampling ratio was 0.4. Finally, a rock outcrop point cloud data experiment is utilized to verify that the proposed approach is applicable for large-scale research objects. All of our experiments illustrate that the proposed adaptive clustered compressive sensing approach can better reconstruct MEMS-based LiDAR point clouds with a lower sampling ratio.

Securing Network Log Data Using Advance Encryption Standard Algorithm And Twofish With Common Event Format

The rapid advancement of information technology demands enhanced security for data exchange in the digital world. Network security threats can arise from various sources, necessitating techniques to protect information transmitted between interconnected networks. Securing network logs is a critical step in strengthening overall network security. Network logs are records of activities within a computer network, including unauthorized access attempts, user activities, and other key events. This research focuses on developing a network log security system by comparing the performance of the Advanced Encryption Standard (AES) and Twofish algorithms, integrated with the Common Event Format (CEF) for encrypting network logs. Tests were conducted on network log datasets to evaluate system functionality and performance. Results indicate that the AES algorithm performs encryption and decryption faster than Twofish. Across five tests with different file sizes, AES took an average of 2.1386 seconds for encryption, while Twofish required 22.8372 seconds. For decryption, AES averaged 2.451 seconds compared to Twofish’s 26.140 seconds. The file sizes after encryption were similar for both algorithms. Regarding CPU usage, AES demonstrated higher efficiency. The average CPU usage during AES encryption was 0.5558%, whereas Twofish used 23.2904%. For decryption, AES consumed 0.4682% of CPU resources, while Twofish required 13.7598%. These findings confirm that AES is not only faster in both encryption and decryption but also more efficient in terms of CPU usage. This research provides valuable insights for optimizing network log security by integrating standardized log formats, like CEF, with appropriate encryption techniques, helping to safeguard against cyber threats.

Word frequency statistics based on MapReduce on serverless platforms

This paper investigates the application of serverless computing in conjunction with the MapReduce framework, particularly in machine learning (ML) tasks. The MapReduce programming model has been widely used to process large-scale datasets by simplifying parallel and distributed data processing. This study explores how the combination of these two technologies can provide more efficient and cost-effective ML solutions. Through a detailed analysis of serverless environments and the MapReduce framework, this paper shows how the combination can advance the fields of cloud computing and machine learning. The experimental part includes the implementation of Map-Reduce model on a serverless platform, exploring the impact of different parameter settings on performance and improving efficiency by optimizing the data processing flow. In addition, the paper analyzes the use of memory and CPU resources and derives the relationship between dataset size, memory consumption and processor configuration and execution time. Through these experiments and analyses, this paper provides an empirical basis and theoretical support for the optimization of cloud computing frameworks.

GRAAFE: GRaph Anomaly Anticipation Framework for Exascale HPC systems

The main limitation of applying predictive tools to large-scale supercomputers is the complexity of deploying Artificial Intelligence (AI) services in production and modeling heterogeneous data sources while preserving topological information in compact models. This paper proposes GRAAFE, a framework for continuously predicting compute node failures in the Marconi100 supercomputer. The framework consists of (i) an anomaly prediction model based on graph neural networks (GNNs) that leverage nodes’ physical layout in the compute room and (ii) the computationally efficient integration into the Marconi100’s ExaMon holistic monitoring system with Kubeflow, an MLOps Kubernetes framework which enables continuous deployment of AI pipelines. The GRAAFE GNN model achieves an area under the curve (AUC) from 0.91 to 0.78, surpassing state-of-the-art (SoA), achieving AUC between 0.64 and 0.5. GRAAFE sustains the anomaly prediction for all the Marconi100 nodes every 120s, requiring an additional 30% CPU resources and less than 5% more RAM w.r.t. monitoring only.

Quick UDP Internet Connections and Transmission Control Protocol in unsafe networks: A comparative analysis

AbstractSecure data transmission and efficient network performance are both key aspects of the modern Internet. Traditionally, Transport Layer Security (TLS)/Transmission Control Protocol (TCP) has been used for reliable and secure networking communications. In the past decade, Quick User Datagram Protocol (UDP) Internet Connections QUIC has been designed and implemented on UDP, attempting to improve security and efficiency of Internet traffic. Real‐world platform investigations are carried out in this paper to evaluate TLS/TCP and QUIC/UDP in maintaining communication, security and efficiency under three different types of popular cyber‐attacks. A set of interesting findings, including delay, loss, server CPU utilisation and server memory usage are presented to provide a comprehensive understanding of the two protocol stacks in performing malicious traffic. More specifically, in terms of the efficiency in achieving short delays and low packet loss rates with limited CPU and memory resources, QUIC/UDP performs better under Denial of Service attacks but TLS/TCP overtakes QUIC/UDP when handling MitM attacks. In terms of security, the implementation of TCP tends to be more secure than QUIC, but QUIC traffic patterns are harder to learn using machine learning methods. We hope that these insights will be informative in protocol selection for future networks and applications, as well as shedding light on the further development of the two protocol stacks.

EPO‐R: An efficient garbage collection scheme for long‐term transactions

SummaryThis article proposes EPO‐R, which is an efficient garbage collection scheme designed for multi‐version concurrency control (MVCC) protocols. MVCC generates a version for each update operation, and it can exhaust physical memory for dynamically changing environments such as IoT payments. Eager pruning of obsolete versions (EPO) is a novel garbage collection technique for long‐term transactions. We found room for improvement in EPO, which triggers reclamation invocation. EPO is triggered for each write operation, and it wastes CPU resources. The proposed method EPO‐R is triggered on a read operation to address this issue. The result of experiments with 64 CPU cores and workloads with the combination of short‐term and long‐term transactions demonstrated that EPO‐R exhibited 6.3 times higher throughput than that of EPO. The reason for this can be explained by analyzing the number of unreclaimable versions.

HSM4SSL: Leveraging HSMs for Enhanced Intra-Domain Security

In a world where digitization is rapidly advancing, the security and privacy of intra-domain communication within organizations are of critical concern. The imperative to secure communication channels among physical systems has led to the deployment of various security approaches aimed at fortifying networking protocols. However, these approaches have typically been designed to secure protocols individually, lacking a holistic perspective on the broader challenge of intra-domain communication security. This omission raises fundamental concerns about the safety and integrity of intra-domain environments, where all communication occurs within a single domain. As a result, this paper introduces HSM4SSL, a comprehensive solution designed to address the evolving challenges of secure data transmission in intra-domain environments. By leveraging hardware security modules (HSMs), HSM4SSL aims to utilize the Secure Socket Layer (SSL) protocol within intra-domain environments to ensure data confidentiality, authentication, and integrity. In addition, solutions proposed by academic researchers and in the industry have not addressed the issue in a holistic and integrative manner, as they only apply to specific types of environments or servers and do not utilize all cryptographic operations for robust security. Thus, HSM4SSL bridges this gap by offering a unified and comprehensive solution that includes certificate management, key management practices, and various security services. HSM4SSL comprises three layers to provide a standardized interaction between software applications and HSMs. A performance evaluation was conducted comparing HSM4SSL with a benchmark tool for cryptographic operations. The results indicate that HSM4SSL achieved 33% higher requests per second (RPS) compared to OpenSSL, along with a 13% lower latency rate. Additionally, HSM4SSL efficiently utilizes CPU and network resources, outperforming OpenSSL in various aspects. These findings highlight the effectiveness and reliability of HSM4SSL in providing secure communication within intra-domain environments, thus addressing the pressing need for enhanced security mechanisms.