network performance evaluation

Addressing the challenges of network performance evaluation under traffic fluctuations and the difficulty of quantifying individual flow impact, this paper proposes a flow-level influence assessment framework integrating percentile statistics, sample distribution analysis, and an improved Shapley value. We innovatively design a utilization scoring metric and leverage cooperative game theory to quantify the marginal contribution of each flow to network resource consumption. From an initial set of 11 candidate metrics, we identify three core indicators balancing interpretability, maintainability, and scalability. Extensive experiments across multiple traffic scenarios demonstrate that our method significantly improves detection sensitivity to abrupt network state transitions induced by anomalous flows. Results confirm that the framework enables efficient, robust performance attribution and optimization decision-making in dynamic networks, exhibiting strong potential for real-world deployment.

Non-technical users often misinterpret network measurement data (e.g., latency, packet loss, throughput), leading to erroneous conclusions. To address this, we propose the first systematic framework leveraging large language models (LLMs) for semantic interpretation of network measurements. Our method integrates historical measurement data with context-aware prompt engineering to automatically translate raw metrics into natural-language performance explanations, enabling scenario-adaptive and personalized feedback. Key contributions include: (1) introducing the first context-aware LLM reasoning paradigm specifically designed for network measurements, overcoming the limitations of conventional threshold-based alerting; and (2) significantly improving non-experts’ comprehension accuracy of critical metrics—experiments show a 37.2% average improvement—while delivering real-time, interpretable, and low-barrier diagnostic support.

This study addresses the challenge of isolating performance anomalies in the middle-mile segment of Internet paths—such as topology errors, suboptimal routing policies, and interconnection congestion—from end-host effects. Leveraging Measurement Lab (M-Lab) data, the authors employ a natural experiment design: users from the same access ISP connect to multiple geographically proximate M-Lab servers, enabling an A/B comparison that effectively controls for client-side, access-network, and temporal variability. This approach, applied at scale for the first time, uncovers previously masked middle-mile anomalies and enables joint detection of topological, routing, and congestion issues. Using a sparse multidimensional histogram method on BigQuery, the system computes Kolmogorov–Smirnov distances and geometric mean throughput ratios in a single pass over millions of samples, efficiently identifying bandwidth bottlenecks, traffic shaping, and suboptimal routes. Results are made publicly accessible through a metropolitan-level real-time dashboard supporting fine-grained analysis.

Existing speed measurement tools focus on peak throughput and poorly reflect users’ perceived responsiveness; emerging metrics such as “latency under load” show promise but their sensitivity to Active Queue Management (AQM) configurations remains unclear. Method: We empirically evaluate three mainstream AQM schemes—CoDel, FQ-CoDel, and SFQ—in a controlled network environment, systematically analyzing their impact on throughput and latency distributions, particularly latency under load. Results: AQM significantly alters speed test outcomes, with distinct latency-throughput trade-offs observed across algorithms under high load. Current measurement platforms, if uncalibrated for AQM, yield misleading latency estimates, undermining the reliability of policy and regulatory decisions. This study is the first to quantitatively characterize the structural impact of AQM on emerging speed metrics, providing critical empirical evidence to inform standardization of measurement tools and evidence-based network governance.

This paper addresses two critical issues in low-earth-orbit satellite networks (LSNs): the conflation of capacity and throughput, and the systematic overestimation of throughput by conventional flow models. It rigorously distinguishes infrastructure-inherent capacity—determined solely by inter-satellite link (ISL) topology and reliability—from scenario-dependent throughput, which is governed by routing policies and traffic load. To this end, it proposes cap-uISL, a capacity model independent of routing or congestion control, enabling quantitative calibration via ISL-level parameters. It further introduces THP-CPE, a throughput computation method based on constrained path expansion and empirically grounded traffic path sets, achieving path-aware, load-adaptive quantification. Evaluated across four emerging inter-satellite transport network (ISTN) architectures, THP-CPE yields significantly more accurate throughput estimates than traditional flow models, with markedly reduced estimation bias; path utilization remains consistently ≤1, ensuring physical feasibility and result credibility; and cap-uISL supports continuous, parameter-driven capacity adjustment—e.g., in response to uISL reliability variations.

This study addresses the accuracy limitations of network simulation in research on the Differentiated Services (DiffServ) architecture by proposing a hybrid validation methodology that integrates real-world experimentation with simulation. The authors deploy representative scenarios in a physical testbed and replicate identical configurations in a re-engineered simulation environment, enabling systematic comparison of key performance metrics. To enhance fidelity, they substantially refactor the DiffServ modules of widely used network simulators, significantly improving their modeling capabilities. Their findings demonstrate that uncalibrated simulations can yield misleading conclusions in advanced networking studies. The proposed approach not only exposes critical shortcomings in current simulation practices but also offers a practical, actionable pathway toward higher simulation accuracy for DiffServ-based research.

This work addresses the lack of a standardized observability framework in quantum networks, which hinders effective fault diagnosis and adaptive control. It proposes the first multidimensional performance metric system tailored for quantum networks, encompassing key parameters such as entanglement fidelity, quantum bit error rate, dark count rate, and timing jitter, while integrating environmental sensor data. Building on this foundation, the authors design and implement a non-intrusive, integrable real-time monitoring prototype, which has been deployed and validated at Oak Ridge National Laboratory. The system enables real-time data acquisition, performance alerting, and dynamic feedback, thereby establishing a critical observability infrastructure for quantum software-defined networking and autonomous control.

This work addresses the limitations of existing performance evaluation approaches for distributed computing continua, which often focus on a single dimension and fail to holistically characterize the behavior of cross-layer heterogeneous systems. The paper presents the first systematic framework that establishes a comprehensive taxonomy of performance metrics spanning three layers—computation, networking, and application/user—as well as emerging non-functional attributes such as sustainability and observability. By integrating mathematical modeling with cross-layer analysis, the study rigorously defines the applicability, measurement phases, and specifications for each metric category. The resulting framework is both clearly structured and extensible, offering a solid theoretical foundation and practical guidance for unified performance assessment in dynamic, heterogeneous environments.

This paper identifies a critical gap in high-performance data transfer research: an overemphasis on network bandwidth while neglecting end-to-end bottlenecks—including latency, TCP congestion control, host CPU limitations, and virtualization—leading to severe discrepancies between benchmark results and real-world production performance. To address this, the authors propose a hardware–software co-design paradigm and develop a latency-programmable testbed. Leveraging high-fidelity wide-area network (WAN) modeling and cross-continental 100 Gbps measurements (Switzerland–California), they systematically isolate key constraints at the network edge and host side. Results demonstrate that primary bottlenecks reside predominantly at the network edge—not the core—and that stable, predictable data movement is achieved across 1–100+ Gbps. This significantly enhances performance fidelity in complex, heterogeneous environments.

This study addresses the unpredictable end-to-end latency in cloud virtualized environments, which stems from virtualization overheads in CPU, I/O, and network resources. Through systematic network measurement experiments across diverse virtualization platforms—including KVM, LXC, and Docker—under multidimensional workload conditions, the authors collect packet round-trip time data to construct a high-quality dataset suitable for machine learning–based network performance modeling. By integrating data preprocessing, correlation analysis, dimensionality reduction, and clustering techniques, this work presents the first quantitative evaluation of latency impacts across multiple virtualization technologies. The resulting dataset effectively supports network performance prediction and intelligent resource scheduling, providing an empirical foundation for performance optimization in cloud environments.