检索结果-内蒙古大学图书馆

Rethinking Cost-Efficient VM Scheduling on Public Edge Platforms: A Service Providers Perspective

IEEE TRANSACTIONS ON MOBILE COMPUTING 2025年第3期24卷 1846-1858页

作者： Li, Yanan Ma, Xiao Fu, Zhe Zhou, Ao Xu, Mengwei Wang, Shangguang Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Tsinghua Univ Dept Comp Sci & Technol Beijing 100084 Peoples R China

Prior studies on traditional centralized clouds independently optimizing static resource utilization and dynamic bandwidth cost are not applicable to edge scenarios, where edge sites are interconnected by wide area networks (WAN) rather than local area networks (LAN) as within clouds. Due to the lack of knowledge about the actual status of public edge platforms and real-world edge datasets, existing influential literature on edge scenarios demonstrates significant disparities in optimization objectives and perspectives. To bridge this gap, we collaborate with a public edge platform and perform a comprehensive measurement, which reveals limitations of the status quo VM scheduling schemes and potential opportunities for improvement. However, resolving VM scheduling considering static resource utilization, dynamic bandwidth cost, and end users' QoE in a cost-efficient manner faces several challenges, including coupled objectives, exponentially increased complexity, and spatiotemporal dynamics. To address the above challenges, in this work, we propose a holistic online framework that integrates combinatorial bandit-based VM migration and seasonality-aware VM request allocation at two distinct time granularities. Large-scale experiments based on a real-world dataset confirm that our online framework achieves near-offline bandwidth cost and resource utilization while significantly lowering time consumption.

关键词： Edge computing VM scheduling cost-efficient Edge computing VM scheduling cost-efficient

来源：评论

学校读者我要写书评

暂无评论

LMGD: Log-Metric Combined Microservice Anomaly Detection Through Graph-Based Deep Learning

引用

IEEE ACCESS 2024年 12卷 186510-186519页

作者： Liu, Xu Liu, Yuewen Wei, Miaomiao Xu, Peng China Acad Ind Internet Beijing 100083 Peoples R China Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

Microservice architecture is a high-cohesion and low-coupling software architecture. Its core idea is to split the application into a set of microservices with a single function and independent deployment. Due to their complexity and large scale, microservice systems are typically fragile and failures are inevitable. Therefore, there is an urgent need for fast and accurate anomaly detection capabilities. However, the existing microservice anomaly detection methods do not pay attention to the multi-source data of the microservice system and thus have low accuracy. To address this limitation, we propose a Log-Metric Combined Microservice Anomaly Detection approach through Graph-based Deep Learning (termed as LMGD). First, we propose a time-aware LSTM prediction neural network to improve the accuracy of service dependency mining. Secondly, based on the service dependency graph, we propose an anomaly detection method based on log-metric fusion, which can more accurately describe the running status of microservices, thereby improving the accuracy of anomaly detection. The experimental outcomes demonstrate that compared with other state-of-the-art methods, our method improves recall and F1-score by 2.63% and 1.05%.

关键词： Microservice architectures Anomaly detection Measurement Long short term memory Logic gates Monitoring Data models Real-time systems Vectors Predictive models log metrics distributed systems GCN GAT

来源：评论

学校读者我要写书评

暂无评论

DeepZoning: Re-accelerate CNN Inference with Zoning Graph for Heterogeneous Edge Cluster

引用

ACM TRANSACTIONS ON ARCHITECTURE AND CODE OPTIMIZATION 2025年第1期22卷 1-26页

作者： Wang, Jingyu Ma, Ruilong Yang, Xiang Qi, Qi Zhuang, Zirui Wang, Jing Liao, Jiangxin Guo, Song Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing Peoples R China Hong Kong Polytech Univ Hong Kong Peoples R China

Parallelizing CNN inference on heterogeneous edge clusters with data parallelism has gained popularity as a way to meet real-time requirements without sacrificing model accuracy. However, existing algorithms struggle to find optimal parallel granularity for complex CNNS, the structure of which is a directed acyclic graph (DAG) rather than a chain, and the parallel dimension is inflexible. To distribute the workload of modern CNNs on heterogeneous devices is also proven as NP-hard problem. In this article, we introduce DeepZoning, a versatile and cooperative inference framework that combines both model and data parallelism to accelerate CNN inference. DeepZoning employs two algorithms at different levels: (1) a low-level Adaptive Workload Partition algorithm that uses linear programming and takes spatial and channel dimensions into optimization during the search for feature map distribution on heterogeneous devices, and (2) a high-level Model Partition algorithm that finds the optimal model granularity and organizes complex CNNs into sequential zones to balance communication and computation during execution. Our experimental evaluations show that Deep-Zoning is effective, achieving up to a 3.02x speed improvement on our experimental prototype compared to state-of-the-art algorithms.

关键词： Cooperative CNN inference edge computing model deployment graph partition

来源：评论

学校读者我要写书评

暂无评论

Joint Communication and Sensing System Performance Evaluation and Testbed: A Communication-Centric Approach

引用

IEEE NETWORK 2024年第5期38卷 286-294页

作者： Zhang, Qixun Ji, Kejia Wei, Zhiqing Feng, Zhiyong Zhang, Ping Beijing Univ Posts & Telecommun Key Lab Universal Wireless Commun Minist Educ Beijing 100876 Peoples R China Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

As one of the promising technologies for 6G mobile system, the joint communication and sensing (JCS) technology can substantially empower communication network with the accurate environment perception ability, as well as enhanced communication performances enabled by the sensing ability. In this article, a communication-centric enabled JCS system has been designed, and the simulation platform has been developed by using different signaling and data payloads in each frame to evaluate the performances of communication and sensing functions. Hardware testbed has been implemented and the numerous field trail results prove the feasibility of the proposed JCS system, paving a fundamental way for the maturity of JCS enabled future 6G mobile system.

关键词： Sensors Robot sensing systems System analysis and design Symbols Hardware Payloads 6G mobile communication Joint communication and sensing 5G New Radio Cram & eacute r-Rao lower bound Hardware testbed

来源：评论

学校读者我要写书评

暂无评论

Joint Beamforming and Dynamic Beam Hopping Based on MAPPO for LEO Satellite Communication System

引用

IEEE WIRELESS COMMUNICATIONS LETTERS 2025年第5期14卷 1461-1465页

作者： Meng, Meng Hu, Bo Chen, Shanzhi Kang, Shaoli Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China China Acad Telecommun Technol State Key Lab Wireless Mobile Commun Beijing 100191 Peoples R China

To meet time-varying and nonuniform service requirements, how to flexibly allocate resource-constrained to meet the traffic demands is an important topic in LEO satellite communication system. We introduce joint beamforming and dynamic beam hopping (BH) algorithm to support hybrid wide-spot beam coverage in LEO satellite communication system, which can transmit control signalling and user data simultaneous. Considering that the joint decision of beamforming and BH will lead to explosive growth of the action space dimension, a cooperative Multi-Agent Proximal Policy Optimization (MAPPO) algorithm is presented to solve this problem in hybrid wide-spot beam coverage scenario. The beamforming problem is decomposed into two subproblems, power allocation and analog beamforming problems. The analog beamforming problem is solved by ZF beamforming algorithm, and the power allocation and BH problems are solved by MAPPO algorithm. In MAPPO algorithm, each agent only undertakes the BH or power allocation decision of one beam. The agents learn to cooperate with others via shared rewards to achieve a common goal of maximize the system throughput and minimize the delay fairness (DF) while ensuring the minimum rate requirement of the control beam. Simulation results show that the MAPPO algorithm can achieve real-time BH and power allocation to match time-varying traffic requests. In addition, the proposed algorithm can achieve about 8 Mbps throughput gain compared with GABH.

关键词： Low earth orbit satellites Satellites Array signal processing Satellite communications Resource management Throughput Heuristic algorithms Vectors Precoding Satellite broadcasting Beamforming beam hopping PPO

来源：评论

学校读者我要写书评

暂无评论

Towards a transmission control mechanism for data centres based on flow deadline and size over software defined networking

引用

INTERNATIONAL JOURNAL OF INTERNET PROTOCOL techNOLOGY 2023年第3期16卷 148-160页

作者： Yu, Tianfang Qiu, Xuesong Beijing Univ Posts & Telecommun Sch Comp Sci State Key Lab Networking & Switching Technol Beijing Peoples R China

Data centres (DCN) support lots of online services, which possess strict deadline constraints. Some research shows that TCP protocol cannot provide deadline-aware transmission services for intra-DCN traffic, because TCP protocol cannot learn about upper application information and always fairly allocate bottleneck bandwidth. As a result, those throughput-sensitive flows have bigger sending windows, yet deadline-aware flows have to queue for forwarding along with longer delays. To address the issue above, we present DSATC, a transmission control mechanism based on flow deadline and size. DSATC uses double ECN thresholds to mark packets, takes the remaining completion time and size of a flow to calculate the flow urgency and redefines the adjustment method of the congestion window (CWND) to balance congestion control and flow deadline completion rate. The experimental results show that DSATC not only efficiently raises flow deadline completion rate and reduces flow completion time but also lifts network throughput.

关键词： data centre SDN deadline-aware congestion control TCP protocol

来源：评论

学校读者我要写书评

暂无评论

Distributed Multi-Agent Reinforcement Learning for Collaborative Path Planning and Scheduling in Blockchain-Based Cognitive Internet of Vehicles

引用

IEEE TRANSACTIONS ON VEHICULAR techNOLOGY 2024年第5期73卷 6301-6317页

作者： Chang, Huigang Liu, Yiming Sheng, Zhengguo Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China China Intelligent & Connected Vehicles Beijing Res Beijing 100176 Peoples R China Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Univ Sussex Dept Engn & Design Brighton BN1 9RH England

The collaborative path planning and scheduling can overcome the limitations of single vehicle intelligence to obtain a globally optimal decision strategy in cognitive Internet of Vehicles (CIoVs). The collaboration of vehicles necessitates the exchange of environmental and decision information, generating massive collaborative computing tasks with strict latency requirements. Leveraging mobile edge computing (MEC) technology, computing tasks can be processed near the vehicles to reduce latency. However, traffic congestion and computational load imbalance seriously affect traffic efficiency and computational latency. In hybrid driving scenarios, it is challenging to fulfill the diverse service requirements of vehicles with different intelligence levels. Moreover, non-collaborative tend to result in traffic congestion due to vehicle aggregation effects, while centralized solutions lack flexibility and have high computational complexity. To address these concerns, a distributed multi-agent reinforcement learning (DMARL) algorithm is proposed for collaborative path planning and scheduling in a blockchain-based collaboration framework. In this framework, we model the communication, traffic situation and task processing of the system and formulate a joint optimization problem to minimize both travel time and computation latency. Last, we convert the scheduling problem for different types of vehicles into Markov decision processes (MDPs) and propose Q-learning-based DMARL algorithm to achieve proactive load balancing of both road infrastructures and MEC nodes (MECNs). Simulation results demonstrate that the proposed approach outperforms the comparison schemes in terms of load balance indexes of roads and MECNs, travel time, and computation latency.

关键词： Collaboration Path planning Roads Task analysis Blockchains Processor scheduling Load management Cognitive Internet of Vehicles (CIoVs) mobile edge computing (MEC) path planning and scheduling multi-agent reinforcement learning load balancing

来源：评论

学校读者我要写书评

暂无评论

引用

IEEE INTERNET OF THINGS JOURNAL 2025年第9期12卷 12034-12047页

作者： Fan, Senran Bao, Zhicheng Dong, Chen Liang, Haotai Xu, Xiaodong Zhang, Ping Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Peng Cheng Lab Dept Broadband Commun Shenzhen 518066 Peoples R China

With the rapid development of Internet of Things (IoT) technology, more sensors are required to operate in complex channel scenarios and under limited communication resources. Semantic communication, as an emerging paradigm, extracts, transmits, and reconstructs information at the semantic level, offering advantages, such as high compression rates and strong noise resistance. These features are expected to find widespread application across various IoT scenarios. However, widely used image similarity evaluation metrics like peak signal-to-noise ratio and multiscale structural similarity index primarily focus on pixel or structural features, making it challenging to accurately measure the loss of semantic-level information during transmission. This limitation poses challenges for the performance evaluation of visual semantic communication systems and restricts the emergence of more novel and efficient systems. To address this issue, we propose a new semantic evaluation metric-semantic similarity score (SeSS). This metric is based on Scene Graph Generation and graph matching techniques, transforming image similarity scores into graph matching scores. By manually annotating thousands of image pairs, we fine-tuned the hyperparameters within SeSS to align it more closely with human semantic perception. The performance of SeSS has been tested across various image datasets and specific IoT visual tasks. Experimental results demonstrate the effectiveness of SeSS in measuring differences in semantic-level information between images, making it a valuable tool for evaluating visual semantic communication systems. This development is expected to encourage the emergence of more robust systems suited for diverse IoT scenarios. The code of SeSS is openly available on https://***/FSR3340/Semantic_Similarty_ScoreGitHub.

关键词： Measurement Visualization Data mining PSNR Noise Image coding Focusing Feature extraction Image similarity Internet of Things (IoT) semantic communication semantic similarity score (SeSS) semantic communication semantic similarity score (SeSS)

来源：评论

学校读者我要写书评

暂无评论

On Analytical Solution of Channel Spatial Correlation for Efficient MIMO OTA Testing

引用

IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS 2025年第4期24卷 933-937页

作者： Zhao, Changsheng Wang, Heng Tian, Lei Zhang, Yuxiang Hu, Zhenyun Yang, Xiaohang Zhang, Jianhua Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China China Acad Informat & Commun Technol Beijing 100191 Peoples R China

This letter presents an analytical (closed-form) expression for computing the channel spatial correlation in multiple-input-multiple-output over-the-air (OTA) testing. Specifically, by leveraging the Jacobi-Anger expansion, the target channel spatial correlation can be accurately characterized using finite-order Bessel functions. Compared to conventional approaches that require complex integration to determine the channel spatial correlation, the proposed method significantly reduces computation time with minimal impact on accuracy. This improvement is crucial for enhancing the efficiency of OTA testing, especially in dynamic channel scenarios where spatial correlation must be calculated for each snapshot. Simulation results show that our method can reduce computation time to 1.4% of that required by the conventional integration method while achieving more than 99.99% computational accuracy.

关键词： Correlation Testing Probes Mathematical models Azimuth Antennas Jacobian matrices Vectors Closed-form solutions Standards Channel reconstruction dynamic over-the-air (OTA) testing Jacobi-Anger expansion spatial correlation

来源：评论

学校读者我要写书评

暂无评论

Group-Based Distinctive Image Captioning with Memory Difference Encoding and Attention

引用

INTERNATIONAL JOURNAL OF COMPUTER VISION 2025年第4期133卷 1435-1455页

作者： Wang, Jiuniu Xu, Wenjia Wang, Qingzhong Chan, Antoni B. City Univ Hong Kong Dept Comp Sci Kowloon Hong Kong Peoples R China Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing Peoples R China

Recent advances in image captioning have focused on enhancing accuracy by substantially increasing the dataset and model size. While conventional captioning models exhibit high performance on established metrics such as BLEU, CIDEr, and SPICE, the capability of captions to distinguish the target image from other similar images is under-explored. To generate distinctive captions, a few pioneers employed contrastive learning or re-weighted the ground-truth captions. However, these approaches often overlook the relationships among objects in a similar image group (e.g., items or properties within the same album or fine-grained events). In this paper, we introduce a novel approach to enhance the distinctiveness of image captions, namely Group-based Differential Distinctive Captioning Method, which visually compares each image with other images in one similar group and highlights the uniqueness of each image. In particular, we introduce a Group-based Differential Memory Attention (GDMA) module, designed to identify and emphasize object features in an image that are uniquely distinguishable within its image group, i.e., those exhibiting low similarity with objects in other images. This mechanism ensures that such unique object features are prioritized during caption generation for the image, thereby enhancing the distinctiveness of the resulting captions. To further refine this process, we select distinctive words from the ground-truth captions to guide both the language decoder and the GDMA module. Additionally, we propose a new evaluation metric, the Distinctive Word Rate (DisWordRate), to quantitatively assess caption distinctiveness. Quantitative results indicate that the proposed method significantly improves the distinctiveness of several baseline models, and achieves state-of-the-art performance on distinctiveness while not excessively sacrificing accuracy. Moreover, the results of our user study are consistent with the quantitative evaluation and demonstrate the r

关键词： Image caption Vision and language Distinctiveness Memory attention

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：