检索结果-内蒙古大学图书馆

International Conference on parallel and distributed systems (ICPADS)

作者： Qihang Wang Decheng Zuo Zhan Zhang Siyuan Chen Tianming Liu Department of Computer Science and Technology Harbin Institute of Technology Harbin China

In the field of real-time analytics, stream joins are the basis for complex queries and greatly affect system performance. In order to satisfy the real-time requirements of streaming applications, the system imposes high requirements on the latency and throughput of the stream join operator. In this paper, we model the latency and throughput of distributed stream join systems based on queuing theory. Based on the analysis of this model, we demonstrate the impact of indexing-related overhead on the latency and throughput of stream join systems and propose a new distributed stream join system, SepJoin, which is oriented to the hash join problem. SepJoin reduces the number of tuples stored in each processing unit belonging to each input stream by designing a novel partitioning scheme that uses as many processing units as possible to store tuples belonging to each input stream, thereby reducing the index-related overhead of each processing unit when performing join operations and ultimately achieving performance benefits in terms of latency and throughput. We provide both theoretical analysis and extensive experimental evaluations to evaluate the processing latency and max throughput of SepJoin.

关键词： Analytical models System performance Throughput real-time systems Indexes Low latency communication Queueing analysis

来源：评论

学校读者我要写书评

暂无评论

Islanded microgrids with output distributed delay: Sampled-data based observer design 14th

Islanded microgrids with output distributed delay: Sampled-d...

引用

14th IFAC workshop on Adaptive and Learning Control systems, ALCOS 2022

作者： Ammeh, L. Fadil, H. El Giri, F. Oulcaid, M. Kabba, A. Advanced Systems Engineering Laboratory Ibn Tofail University Kenitra Morocco Automatic Laboratory Caen University Normandie France

In this paper, two majors islanded microgrid issues are addressed: i) the voltage amplitude and frequency control and ii) the time delay introduced by the use of communication networks leading to instability. The studied system is composed of a distributed generator (DG), a voltage source inverter (VSI), an LC filter and local loads. After the system modeling, a voltage controller by means of the backstepping method is proposed. Then, the distributed time delay that comes in the VSI output current is dealt with supposing that it is the only measured signal and the measurements are available only in sampling periods. Two observers are designed to give online estimation of the state signals;the first one involves an intersample predictor and the second one uses a ZOH predictor. The predictors compensate the time delay and the sampling effects. The observers are designed on the base of the high gain principle to deal with the system nonlinearity. From simulations and formal analysis, the following results are assumed: the output voltage tracks its reference and it remains stable, the online estimated signals, given by both observers, converge exponentially to their real values with large time delay and sampling interval. It is worth noting that it is the first time that distributed time delay in islanded microgrids is addressed. © 2022 Elsevier B.V.. All rights reserved.

关键词： Backstepping

来源：评论

学校读者我要写书评

暂无评论

Improving Power System Resilience Based on Grid-Forming Converter Control and real-time Monitoring 8

Improving Power System Resilience Based on Grid-Forming Conv...

引用

8th IEEE workshop on the Electronic Grid, eGRID 2023

作者： Ungerland, Jakob Denninger, Rebekka Werner, Daniel Schroven, Kris Lickert, Benjamin Kopke, Corinna Stolz, Alexander Fraunhofer Institute for Solar Energy Systems ISE Freiburg Germany Fraunhofer Institute for High-Speed Dynamics Ernst-Mach-Institut EMI Efringen-Kirchen Germany

ISBN: (纸本)9798350327007

Power system resilience becomes more important in the context of an increasing share of renewable energy generation and heavy weather events jeopardizing the reliability of power supply. This work proposes an approach to improve power system resilience by establishing a communication between a so-called Resilience-Monitor and a grid-forming battery converter. The Resilience-Monitor is an application that calculates the power system resilience based on different key performance indicators by utilizing grid measurements and weather forecast data. As soon as critical values for the key performance indicators are reached, warnings are communicated to the grid-forming converter that adapts its battery operation according to the received signals. Results of an exemplarily conducted case study show a local improvement in power system resilience in case of an island operation of a medium voltage grid due to faults in the high voltage grid caused by heavy weather. © 2023 IEEE.

关键词： distributed power generation

来源：评论

学校读者我要写书评

暂无评论

Underwater Waste Detection with YOLO: A High-Performance Computing Technique

Underwater Waste Detection with YOLO: A High-Performance Com...

引用

International Conference on parallel distributed and Grid Computing (PDGC)

作者： Vivek Kumar Umangi Nigam Manan Goel Neeraj Rathore Ajitesh Kumar Mayank Srivastava School of Comp. Sc. Engg. & Tech. Bennett University Greater Noida India Operations & Technology Management UPES Dehradun India Computer Engineering and Application GLA University Mathura India

ISBN: (数字)9798331521349

ISBN: (纸本)9798331521356

Waste accumulation in the oceans poses severe environmental challenges that require efficient mechanisms for detection and removal. The methods usually employ slow and resource-intensive techniques like sonar-based detection and manual inspection, hence the need for automated mechanisms of detection. Waste detection under water is a regular monitoring task. Thus, the work proposes rapid model training which allows real-time deployment and enhancement for prevention of degradation of water bodies for various type of wastes. The work uses high-performance computing techniques such as parallel processing, NVIDIA GPUs, CUDA, and YOLO. A deep learning model featuring Convolutional Neural Networks (CNN) is integrated with YOLO (You Only Look Once) to augment detection speed along with accuracy. The model training time is particularly important for real-time applications which allows rapid iteration for the researchers and practitioners. This work points to the fact that the GPU architecture has accelerated 41X in training deep models and improved over precision, and inference time. In all aspects the proposed model claims to perform real-time waste detection within permissible computational latency. This work is in demand for autonomous systems like underwater drones which demand the fast and timely decision-making capability to operate and maneuver properly in dynamic operational environments.

关键词： YOLO Training Deep learning Computational modeling High performance computing Graphics processing units Computer architecture parallel processing real-time systems Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

VILLASnode-based Co-simulation of Local Energy Communities 1

VILLASnode-based Co-simulation of Local Energy Communities

引用

1st International workshop on Open Source Modelling and Simulation of Energy System (OSMSES)

作者： Carta, Daniele Weber, Moritz Gluecker, Philipp Pesch, Thiemo Hagenmeyer, Veit Benigni, Andrea Kuhnapfel, Uwe Forschungszentrum Julich IEK 10 Julich Germany Karlsruher Inst Technol IAI Karlsruhe Germany

ISBN: (纸本)9781665410083

In this paper, a flexible co-simulation environment that allows to test the coupling among different energy communities is presented. The presented solution allows each community to simulate its system and the coupling with the others, with the desired level of detail, while preserving the share of sensitive information. The tests have been conducted by connecting the laboratories of Forschungszentrum Julich and Karlsruhe Institute of Technology by means of VILLASnode. The campuses of the two research centers are considered as local energy communities (LECs) under tests with an interconnecting transmission grid, and different test scenarios have been considered. The results show a successful co-simulation of the LECs through heterogeneous digital real-time simulators, as well as limitations regarding the modeling of dynamic loads by controlled current sources.

关键词： Communities Co-simulation Power system simulation real-time systems Geographically distributed systems

来源：评论

学校读者我要写书评

暂无评论

Hierarchical Heterogeneous Cluster systems for Scalable distributed Deep Learning

Hierarchical Heterogeneous Cluster Systems for Scalable Dist...

引用

International Symposium on Object-Oriented real-time distributed Computing

作者： Yibo Wang Tongsheng Geng Ericson Silva Jean-Luc Gaudiot Electrical Engineering and Computer Science University of California Irvine Irvine USA Electrical Engineering and Computer Science Pontifical Catholic University of Minas Gerais Minas Gerais Brazil

ISBN: (数字)9798350371284

ISBN: (纸本)9798350371291

distributed deep learning framework tools should aim at high efficiency of training and inference of distributed exascale deep learning algorithms. There are three major challenges in this endeavor: scalability, adaptivity and efficiency. Any future framework will need to be adaptively utilized for a variety of heterogeneous hardware and network environments and will thus be required to be capable of scaling from single compute node up to large clusters. Further, it should be efficiently integrated into popular frameworks such as TensorFlow, PyTorch, etc. This paper proposes a dynamically hybrid (hierarchy) distribution structure for distributed deep learning, taking advantage of flexible synchronization on both centralized and decentralized architectures, implementing multi-level fine-grain parallelism on distributed platforms. It is scalable as the number of compute nodes increases, and can also adapt to various compute abilities, memory structures and communication costs.

关键词： Deep learning Training Costs Scalability parallel processing real-time systems Inference algorithms

来源：评论

学校读者我要写书评

暂无评论

An intelligent surveillance video analytics framework using NACT-Hadoop/MapReduce on cloud services

引用

distributed AND parallel DATABASES 2021年第4期39卷 873-889页

作者： Nirmalan, R. Gokulakrishnan, K. Bannari Amman Inst Technol Dept Comp Sci & Engn Sathyamangalam Tamil Nadu India Anna Univ Dept Elect & Commun Engn Reg Campus Tirunelveli Tamil Nadu India

Video analytics has gradually increased in recent years. The intelligent CCTV cameras in public places, you-tube videos, etc. generate an enormous amount of video data. Generally, video analytics required more time as it contains several processes like encoding, decoding, etc. There are several existing approaches are evolved in improving the efficiency of video analytics but performance delay and loss of data still existing challenges. With our analysis, we strongly state VM migration will be an effective solution to overcome this delay and performance issues. In this paper, we propose NACT based map reducing mechanism (NACT-Map) for processing the real-time streaming videos. The NACT (Novel Awaiting Computation time) enables the prediction of VM allocation and automatic migration. The scheduling and allocating of the optimal resource are done by task monitor who utilizes the Task manager (TM) system. The NACT based VM migration and MapReduce technique with Hadoop simplifies the process and minimizes the execution time. The splitting of video into chunks of frames speedup the process. Further efficiency is improved by the Map Reduce technique which uses video and its related content for clusters. The performance of our proposed system is executed in the cloudsim with a large dataset contains two real-time videos. Further, the result is compared with the existing methodologies such as distributed video decoding mechanism with extended FFmpeg and VideoRecordReader (VDMFF) (Yoon et al. in distributed video decoding on Hadoop. IEICE Trans Inf Syst E101-D(1):2933-2941, 2018) and distributed Video Analytics Framework for Intelligent Video Surveillance (SIAT) (Uddin et al. in SIAT: a distributed video analytics framework for intelligent video surveillance. Symmetry 11:911, 2019). The obtained result shows our proposed NACT_Map consumes minimum Task processing time (p(tix)) and about 90% of efficiency in overall system performance is increased.

关键词： real time systems Performance Video analytics distributed video processing Hadoop MapReduce

来源：评论

学校读者我要写书评

暂无评论

Edge Robotics: Edge-Computing-Accelerated Multirobot Simultaneous Localization and Mapping

引用

IEEE INTERNET OF THINGS JOURNAL 2022年第15期9卷 14087-14102页

作者： Huang, Peng Zeng, Liekang Chen, Xu Huang, Luo Zhou, Zhi Yu, Shuai Sun Yat Sen Univ Sch Comp Sci & Engn Guangzhou 510006 Guangdong Peoples R China

With the wide penetration of smart robots in multifarious fields, the simultaneous localization and mapping (SLAM) technique in robotics has attracted growing attention in the community. Yet collaborating SLAM over multiple robots still remains challenging due to performance contradiction between the intensive graphics computation of SLAM and the limited computing capability of robots. While traditional solutions resort to the powerful cloud servers acting as an external computation provider, we show by real-world measurements that the significant communication overhead in data offloading prevents its practicability to real deployment. To tackle these challenges, this article promotes the emerging edge-computing paradigm into multirobot SLAM and proposes RecSLAM, a multirobot laser SLAM system that focuses on accelerating the map construction process under the robot-edge-cloud architecture. In contrast to the conventional multirobot SLAM that generates graphic maps on robots and completely merges them on the cloud, RecSLAM develops a hierarchical map fusion technique that directs robots' raw data to edge servers for real-time fusion and then sends to the cloud for global merging. To optimize the overall pipeline, an efficient multirobot SLAM collaborative processing framework is introduced to adaptively optimize robot-to-edge offloading tailored to heterogeneous edge resource conditions, meanwhile ensuring the workload balancing among the edge servers. Extensive evaluations show RecSLAM can achieve up to 39.31% processing latency reduction over the state of the art. Besides, a proof-of-concept prototype is developed and deployed in real scenes to demonstrate its effectiveness.

关键词： Robots Simultaneous localization and mapping Servers Robot sensing systems Cloud computing Robot kinematics Merging distributed and parallel processing edge intelligence edge offloading multirobot laser simultaneous localization and mapping (SLAM)

来源：评论

学校读者我要写书评

暂无评论

Stream Processing with Dependency-Guided Synchronization 22

Stream Processing with Dependency-Guided Synchronization

引用

27th ACM SIGPLAN Symposium on Principles and Practice of parallel Programming (PPoPP)

作者： Kallas, Konstantinos Niksic, Filip Stanford, Caleb Alur, Rajeev Univ Pennsylvania Philadelphia PA USA

ISBN: (纸本)9781450392044

real-time data processing applications with low latency requirements have led to the increasing popularity of stream processing systems. While such systems offer convenient APIs that can be used to achieve data parallelism automatically, they offer limited support for computations that require synchronization between parallel nodes. In this paper, we propose dependency-guided synchronization (DGS), an alternative programming model for stateful streaming computations with complex synchronization requirements. In the proposed model, the input is viewed as partially ordered, and the program consists of a set of parallelization constructs which are applied to decompose the partial order and process events independently. Our programming model maps to an execution model called synchronization plans which supports synchronization between parallel nodes. Our evaluation shows that APIs offered by two widely used systems-Flink and timely Dataflow-cannot suitably expose parallelism in some representative applications. In contrast, DGS enables implementations with scalable performance, the resulting synchronization plans offer throughput improvements when implemented manually in existing systems, and the programming overhead is small compared to writing sequential code.

关键词： data parallelism sharding synchronization distributed stream processing

来源：评论

学校读者我要写书评

暂无评论

Enhancing blockchain security through natural language processing and real-time monitoring

引用

INTERNATIONAL JOURNAL OF parallel EMERGENT AND distributed systems 2023年

作者： Salzano, Francesco Pareschi, Remo Univ Molise Stake Lab Campobasso Italy Univ Cagliari Dept Math & Comp Sci Cagliari Italy BB Smile Srl Rome Italy

We describe implementing and testing a security monitoring system for Blockchain-based applications by performing Log Analysis through natural language processing (NLP) techniques and enabling queries via REST API. The study focuses on the Hyperledger Fabric framework, a highly reliable and scalable open-source software platform for private blockchains. The real-time monitoring system aims to detect and prevent attack scenarios on the network structure and smart contracts, such as DDOS, Sybil, and Eclipse. The vulnerability analysis is extended to inspecting Docker containers to detect network-based attacks. Although the case study focuses on Hyperledger Fabric, the techniques used are general and applicable to all Blockchain-based systems and fulfill the need to provide adequate monitoring capabilities for blockchain implementations.

关键词： Blockchain distributed ledger technology distributed systems natural language processing security log analysis

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：