检索结果-内蒙古大学图书馆

IDAD: An improved tensor train based distributed DDoS attack detection framework and its application in complex networks

引用

FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE 2025年 162卷

作者： Fan, Qiyuan Li, Xue Wang, Puming Jin, Xin Yao, Shaowen Miao, Shengfa An, Min Zhao, Yuqing Yunnan Univ Sch Software Kunming 650091 Peoples R China Henan Univ Sci & Technol Sch Elect Informat Engn Xinxiang 453003 Peoples R China Hubei Univ Automot Technol Shiyan 442000 Peoples R China

With the vigorous development of Internet technology, the scale of systems in the network has increased sharply, which provides a great opportunity for potential attacks, especially the distributed Denial of Service (DDoS) attack. In this case, detecting DDoS attacks is critical to system security. However, current detection methods exhibit limitations, leading to compromises in accuracy and efficiency. To cope with it, three key strategies are implemented in this paper: (i) Using tensors to model large-scale and heterogeneous data in complex networks;(ii) Proposing a denoising algorithm based on the improved and distributed tensor train (IDTT) decomposition, which optimizes the tensor train(TT) decomposition in terms of parallel computation and low-rank estimation;(iii) Combining (i), (ii) and Light Gradient Boosting Machine (LightGBM) classification model, an efficient DDoS attack detection framework is proposed. Datasets CIC-DDoS2019 and NSL-KDD are used to evaluate the framework, and results demonstrate that accuracy can reach 99.19% while having the characteristics of low storage consumption and well speedup ratio.

关键词： Tensor train decomposition DDoS attack detection distributed computing System storage efficiency Machine learning

来源：评论

学校读者我要写书评

暂无评论

Multi-attribute-based self-stabilizing algorithm for leader election in distributed systems

引用

JOURNAL OF SUPERcomputing 2025年第4期81卷 1-35页

作者： Biswas, Amit Singh, Manisha Baranwal, Gaurav Tripathi, Anil Kumar Aknine, Samir Motilal Nehru Natl Inst Technol Allahabad Dept Comp Sci & Engn Prayagraj India Maulana Azad Natl Inst Technol Dept Comp Sci & Engn Bhopal India Banaras Hindu Univ Dept Comp Sci Varanasi India Indian Inst Technol BHU Dept Comp Sci & Engn Varanasi India Univ Claude Bernard Lyon 1 LIRIS lab Lyon France

Cloud computing, edge computing, replicated databases, and various modern technologies adopt distributed computing concepts for a reliable, high-performance, large-scale computational platform. In distributed computing, leader election is a fundamental problem because the elected leader helps coordinate and utilize the resources efficiently. Several state-of-the-art works mentioned that a good quality leader is essential as it improves system performance, simplifies management procedures, reduces coordinational complexity, and makes the system more fault-tolerant. However, designing a self-stabilizing leader election algorithm with weak assumptions in a failure-prone distributed environment is challenging. This paper proposes a multi-attribute-based, self-stabilizing, network partition-tolerant leader election method for failure-prone distributed systems. Here, based on the system requirements, the pertinent attributes of a good quality leader are identified and assigned weights according to their importance. Then, the identified attributes and their weight are used to elect a suitable node as the leader. We show that the algorithm is self-stabilizing and can tolerate multiple nodes and link failures. Further, we analyze the proposed algorithm's time, communication, and bit complexities. We consider a distributed database system scenario to simulate the proposed method and compare it with existing approaches to evaluate and validate the proposed method's performance and the elected leader's quality.

关键词： Leader election distributed computing distributed systems Multi-attribute decision-making Group decision-making

来源：评论

学校读者我要写书评

暂无评论

Scaling Up Optuna: P2P distributed Hyperparameters Optimization

引用

CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE 2025年第4-5期37卷

作者： Cudennec, Loic Minist Armed Forces Ministerial Agcy Def Artificial Intelligence AMIAD Bruz France

In machine learning (ML), hyperparameter optimization (HPO) is the process of choosing a tuple of values that ensures an efficient deployment and training of an AI model. In practice, HPO not only applies to ML tuning but can also be used to tune complex numerical simulations. In this context, a numerical model of a given object is created to be used in realistic simulations. This model is defined by a set of values describing properties such as the geometry of the object or other unknown parameters related to physical quantities. While HPO for ML usually requires finding a few parameters, a numerical model can involve the tuning of more than a hundred parameters. As a consequence, a large number of tuples have to be explored and evaluated before finding a relevant solution, offering new challenges in high-performance computing for efficiently driving the optimization. In this work we rely on the Optuna HPO framework, primarily designed for ML tasks and including state-of-the-art sampling and pruning algorithms. We report on its use to optimize a complex numerical model onto a 1024-core machine. We suggest 1.5M tuples and evaluate 5M simulations using different Optuna-distributed layouts to build several tradeoffs between performance and energy consumption metrics. In order to further scale up the optimization process onto resources, we introduce OptunaP2P, an extension of Optuna based on the peer-to-peer paradigm. This allows to remove any bottleneck in the management of the shared knowledge between optimization processes. With OptunaP2P, we were able to compute up to 3 times faster compared to the regular Optuna-distributed implementation and to obtain close-to-similar results in terms of quality in this reduced time-frame.

关键词： distributed computing energy consumption high-performance computing hyperparameters optimization machine learning operational research optuna peer-to-peer

来源：评论

学校读者我要写书评

暂无评论

A modularized parallel distributed High-Performance computing framework for simulating seasonal frost dynamics in Canadian croplands

引用

COMPUTERS AND ELECTRONICS IN AGRICULTURE 2023年 212卷

作者： Li, Ziwei Qi, Zhiming Liu, Yuchen Zheng, Yunjia Yang, Yi McGill Univ Dept Bioresource Engn Ste Anne De Bellevue PQ H9X 3V9 Canada Indiana Univ Dept Comp Sci Bloomington IN 47408 USA McGill Univ Sch Comp Sci Montreal PQ H3A 2A7 Canada DevOps Fortinet Technol Burnaby BC V5C 6C6 Canada Mcgill Univ 21111 Lakeshore Rd Ste Anne de Bellevue PQ H9X 3V9 Canada

The winter/spring season in cold climate regions has been recognized as a critical period for cropland nutrient loss and greenhouse gas emissions and has been predicted to be vulnerable to climate change. Agricultural system models serve as essential tools in anticipating potential outcomes. The RZ-SHAW model, combining the Simultaneous Heat and Water Model (SHAW) with the Root Zone Water Quality Model 2 (RZWQM2), comprehensively describes snow and soil freezing dynamics. Showing promising results in simulating winter frost dynamics it, however, remains computationally intensive. High-performance computing has been recognized as a promising approach to reducing simulation runtimes. A modularized parallel distributed highperformance computing (HPC) framework (RS-DPCF) with both multi-threading (MT) and multi-processing (MP) capability was developed for RZ-SHAW calibration and simulation. Two case studies demonstrated that the developed RS-DPCF program could significantly reduce RZ-SHAW calibration time, allow improved computation resource scalability, and perform field-scale simulations of overwintering conditions for multiple croplands across Canada. For the RZ-SHAW model, MT provided a parallel-computing efficiency superior to that of MP.

关键词： Parallel computing RZ-SHAW Winter hydrology distributed computing RZWQM2

来源：评论

学校读者我要写书评

暂无评论

Parallel Committees: a scalable, secure, and fault-tolerant distributed NoSQL database architecture

引用

CLUSTER computing-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS 2025年第4期28卷 1-48页

作者： Solat, Siamak Univ Paris Cite 45 Rue St Peres F-75006 Paris France

Most existing Byzantine fault-tolerant algorithms are slow and not designed for large participant sets trying to reach consensus. Consequently, distributed databases that use consensus mechanisms to process transactions face significant limitations in scalability and throughput. These limitations can be substantially improved using sharding, a technique that partitions a state into multiple shards, each handled in parallel by a subset of the network. Sharding has already been implemented in several data replication systems. While it has demonstrated notable potential for enhancing performance and scalability, current sharding techniques still face critical scalability and security issues. In this article, we propose a novel, fault-tolerant, self-configurable, scalable, secure, decentralized, high-performance distributed NoSQL database architecture. Our approach employs an innovative sharding technique to enable Byzantine fault-tolerant consensus mechanisms in very large-scale networks. We introduce a new sharding method for data replication that leverages a classic consensus mechanism, such as PBFT, to process transactions. Our approach modifies node allocation among shards through the public key generation process, effectively reducing the frequency of cross-shard transactions, which are generally more complex and costly than intra-shard transactions. Additionally, our method eliminates the need for a shared ledger between shards, which typically imposes further scalability and security challenges on the network. We also explain how to automatically form new committees based on the availability of candidate processor nodes. This technique optimizes network capacity by employing inactive surplus processors from one committee's queue in forming new committees, thereby increasing system throughput and efficiency. Our approach maximizes processor node utilization as well as computational and storage capacity across the network, enhancing both processing and storage sha

关键词： Sharding Data replication distributed consensus mechanisms distributed databases NoSQL distributed computing Byzantine-fault-tolerance Cryptography

来源：评论

学校读者我要写书评

暂无评论

distributed ADMM With Linear Updates Over Directed Networks

引用

IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 2025年第2期12卷 1396-1407页

作者： Rokade, Kiran Kalaimani, Rachel Kalpana Cornell Univ Sch Elect & Comp Engn Ithaca NY 14850 USA Indian Inst Technol Madras Dept Elect Engn Chennai 600036 Tamil Nadu India

distributed optimization over a network of agents is ubiquitous in applications such as power system, robotics and statistical learning. In many settings, the communication network is directed, i.e., the communication links between agents are unidirectional. While several variations of gradient-descent-based primal methods have been proposed for distributed optimization over directed networks, an extension of dual-ascent methods to directed networks remains a less-explored area. In this paper, we propose a distributed version of the Alternating Direction Method of Multipliers (ADMM) with linear updates for directed networks using balancing weights, called BW-DADMM (Balancing Weights Directed ADMM). We show that if the objective function of the minimization problem is smooth and strongly convex, then BW-DADMM achieves a geometric rate of convergence to the optimal point. Our algorithm exploits the robustness inherent to ADMM by not enforcing accurate consensus, thereby significantly improving the convergence rate. We illustrate this by numerical examples, where we compare the performance of BW-DADMM with that of state-of-the-art ADMM methods over directed graphs.

关键词： Convex functions Optimization Directed graphs Convergence Standards Vectors Linear programming Approximation algorithms Statistical learning Robustness Alternating Direction Method of Multipliers (ADMM) dual-ascent directed graphs distributed computing parallel computing

来源：评论

学校读者我要写书评

暂无评论

Energy-time modelling of distributed multi-population genetic algorithms with dynamic workload in HPC clusters

引用

FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE 2025年 167卷

作者： Escobar, Juan Jose Sanchez-Cuevas, Pablo Prieto, Beatriz Kiziltepe, Rukiye Savran Diaz-del-Rio, Fernando Kimovski, Dragi Univ Granada Dept Software Engn CITIC Granada Spain Univ Seville Dept Comp Architecture & Technol Seville Spain Univ Granada Dept Comp Engn Automat & Robot CITIC Granada Spain Karadeniz Tech Univ Dept Software Engn Trabzon Turkiye Ankara Univ Dept Software Engn Ankara Turkiye Univ Klagenfurt Inst Informat Technol Klagenfurt Austria

Time and energy efficiency is a highly relevant objective in high-performance computing systems, with high costs for executing the tasks. Among these tasks, evolutionary algorithms are of consideration due to their inherent parallel scalability and usually costly fitness evaluation functions. In this respect, several scheduling strategies for workload balancing in heterogeneous systems have been proposed in the literature, with runtime and energy consumption reduction as their goals. Our hypothesis is that a dynamic workload distribution can befitted with greater precision using metaheuristics, such as genetic algorithms, instead of linear regression. Therefore, this paper proposes anew mathematical model to predict the energy-time behaviour of applications based on multi-population genetic algorithms, which dynamically distributes the evaluation of individuals among the CPU-GPU devices of heterogeneous clusters. An accurate predictor would save time and energy by selecting the best resource set before running such applications. The estimation of the workload distributed to each device has been carried out by simulation, while the model parameters have been fitted in a two-phase run using another genetic algorithm and the experimental energy-time values of the target application as input. When the new model is analysed and compared with another based on linear regression, the one proposed in this work significantly improves the baseline approach, showing normalised prediction errors of 0.081 for runtime and 0.091 for energy consumption, compared to 0.213 and 0.256 shown in the baseline approach.

关键词： Energy-time modelling Heterogeneous clusters distributed computing Parameter optimisation Task scheduling Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

A distributed Information Granulation Method for Time Series Clustering

引用

CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE 2025年第4-5期37卷

作者： Mu, Yashuang Liu, Tian Zhang, Wenqiang Guo, Hongyue Wang, Lidong Liu, Xiaodong Henan Univ Technol Sch Artificial Intelligence & Big Data Zhengzhou Peoples R China Henan Univ Technol Coll Informat Sci & Engn Zhengzhou Peoples R China Dalian Maritime Univ Sch Maritime Econ & Management Dalian Peoples R China Dalian Maritime Univ Collaborat Innovat Ctr Transport Studies Dalian Peoples R China Dalian Maritime Univ Sch Sci Dalian Peoples R China Dalian Univ Technol Fac Elect Informat & Elect Engn Sch Control Sci & Engn Dalian Peoples R China

Time series clustering is an important research problem in machine learning and data mining. With the rapid increase in the amount of time series data, many traditional clustering algorithms cannot directly deal with large-scale time series due to some limitations in the memory capacity and the execution time. In this study, we suggest a distributed information granulation method for large-scale time clustering problem. First, a distributed time series partitioning method is designed to randomly divide the original time series dataset into some data blocks. Then, the distributed time series granulation method is developed in the map-reduce framework by the principle of reasonable granularity, where each time series can be described by some representative data points to show the trend state information. Finally, we introduce the large-scale time series clustering method in terms of the fuzzy C-means clustering algorithm. The experimental studies demonstrate the feasibility and the effectiveness on several UCR publicly benchmark time series datasets. Compared with the classical clustering methods, the proposed method can achieve a 4.86-9.65% improvement in average clustering accuracy. Meanwhile, the proposed method exhibits more advantages in both unequal length time series clustering and execution time.

关键词： clustering distributed computing information granulation time series

来源：评论

学校读者我要写书评

暂无评论

DRLFluent: A distributed co-simulation framework coupling deep reinforcement learning with Ansys-Fluent on high-performance computing systems

引用

JOURNAL OF COMPUTATIONAL SCIENCE 2023年 74卷

作者： Mao, Yiqian Zhong, Shan Yin, Hujun Univ Manchester Dept Mech Aerosp & Civil Engn Manchester M13 9PL England Univ Manchester Dept Elect & Elect Engn Manchester M13 9PL England

For active flow control (AFC), several frameworks have been developed to enable dynamic interactions between deep reinforcement learning (DRL) agents and computational fluids dynamics (CFD) environments. However, the highly coupled modules in the existing frameworks require massive development efforts to be applied to different flow scenarios. Moreover, these frameworks do not support applications to run flexibly on multiple nodes in high-performance computing (HPC) systems. Herein, we propose a new framework named as DRLFluent coupling an open-source DRL package with a well-developed general CFD solver, Ansys-Fluent. The distributed HPC deployment of DRLFluent is achieved by a broker architecture which enables the DRL client and CFD server written in different languages to communicate across different nodes via a pair of standardized interfaces. We have evaluated DRLFluent along with two other representative frameworks using a benchmark case in which a laminar flow around a circular cylinder (Re=100) is controlled by two synthetic jets. The strategy discovered in DRLFluent is capable of neutralizing the small asymmetry of the shedding vortices, yielding a higher drag reduction than the other two frameworks. When the number of parallel environments exceeds 20, DRLFluent exhibits the largest speedup among the frameworks evaluated. Capable of distributing workload across multiple nodes, DRLFluent is suitable for implementing DRL for AFC of complex flow scenarios which requires hundreds of cores or more. Benefitting from a loose coupling between DRL and CFD, the standardized interfaces and the user-friendly CFD solver, DRLFluent is readily applicable to a wide range of flow scenarios without laborious low-level programming efforts.

关键词： Computational fluid dynamics Reinforcement learning High -performance computing distributed computing Flow control

来源：评论

学校读者我要写书评

暂无评论

Hybrid in-network computing and distributed learning for large-scale data processing

引用

COMPUTER NETWORKS 2023年 226卷

作者： Jeon, So-Eun Lee, Sun-Jin Lee, Il-Gu Sungshin Womens Univ Dept Future Convergence Technol Engn 34 Da GilBomun Ro Seoul 02844 South Korea

Recently, the value of big data generated, processed, and shared by mobile devices and intelligent Internet of Things devices has significantly increased. The speed of data generation is accelerating, and the amount of data that needs to be collected and learned in a distributed database is increasing. However, the conventional data processing methods, centralized and distributed computing, are reaching their limits owing to the trade-off between data processing speed and learning accuracy. To solve this problem, in this study, a hybrid in -network computing (HINC) method that can process data quickly and accurately is proposed. The efficiency of the proposed HINC method was verified by comparing and analyzing its data loss, throughput, and accuracy to those of the conventional method. The experimental results show that in an environment with many nodes in the network, HINC reduces the data loss, which means data that could not be processed because it exceeded the available throughput of the node by 20.94% compared to the conventional method. Furthermore, in an envi-ronment with full workload, HINC reduced data loss by an average of 55.6% compared with the conventional centralized computation method. It also improved throughput by an average of 18.91% and model evaluation accuracy by 15.7% compared with the conventional fully distributed computing model.

关键词： Centralized computing distributed computing In-network computing Data processing Large-scale data

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：