system performance measurement

AI infrastructure confronts multidimensional physical and economic constraints—including power, thermal management, water usage, interconnect bandwidth, memory capacity, and data throughput—while existing metrics (e.g., PUE, TCO) are siloed and fail to capture the coupled trade-offs among energy efficiency, performance, and cost, hindering cross-layer co-optimization. To address this, we propose a unified measurement architecture grounded in a 6×3 cross-layer taxonomy—spanning facility, network, compute, storage, software, and application layers, each annotated with physical, computational, and economic semantics—and introduce the Measurement Propagation Graph (MPG) to enable, for the first time, system-level, three-dimensional relational modeling. Leveraging systematic literature review, meta-analysis, and graph-based modeling, our framework integrates heterogeneous, multi-source metrics. It supports benchmarking, capacity planning, and total cost of ownership analysis, substantially enhancing interpretability of AI cluster efficiency frontiers and enabling rigorous multi-objective optimization.

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.

Cloud-native applications operate in multi-tenant, shared cloud environments, where conventional energy-efficiency evaluation methods—relying solely on isolated local metrics (e.g., CPU utilization)—fail to capture holistic system-level energy behavior. To address this, we propose the first automated, scalable energy-efficiency experimentation framework tailored for Kubernetes-native applications, enabling joint quantification of energy consumption and QoS metrics (latency, throughput, error rate) across container, platform, and infrastructure layers. The framework tightly integrates eBPF for fine-grained observability, Prometheus for metric collection, and hardware power interfaces (RAPL/ACPI) for accurate energy measurement, and establishes an end-to-end sustainability assessment pipeline. Evaluation across multiple open-source cloud-native applications demonstrates up to 42% energy-efficiency variation among architectural variants, while precisely characterizing their trade-offs with P99 latency and service availability.

This study addresses the energy–latency–reliability trade-off in data compression and transmission for resource-constrained wireless devices in edge computing. We propose an application-driven end-to-end latency budgeting mechanism, departing from conventional hard real-time constraints. A joint optimization model is formulated, with compression ratio and device processing speed as key decision variables, to characterize their nonlinear interdependencies and compute the Pareto-optimal frontier. Theoretical analysis and experiments demonstrate that modest relaxation of end-to-end latency yields exponential reductions in energy consumption—minor latency increases enable substantial energy savings. The proposed framework provides a quantifiable, configurable design paradigm for low-power, adaptive edge communication, while rigorously guaranteeing reliability requirements.

Traditional performance metrics for AI-driven Distributed Computing Continua (DCC) lag behind the demands of generative AI and large language models. This paper identifies structural gaps in existing cloud-edge-device co-design metric frameworks—particularly concerning sustainability, energy efficiency, and system observability. Through a systematic literature review and cross-layer workload analysis, we propose a multidimensional, AI-centric metric framework that introduces four novel dimensions: energy-efficiency ratio, carbon-aware latency, resource coordination rate, and observability entropy, along with principled criteria for metric selection. Moving beyond conventional throughput- and latency-centric paradigms, our framework provides both theoretical foundations and practical guidance for designing and evaluating DCC systems. It advances distributed AI toward greater efficiency, environmental sustainability, and interpretability.

This study addresses the multi-objective trade-offs among generation quality, energy consumption, latency, and memory when deploying large language models on edge devices. The authors construct a reproducible empirical evaluation framework to systematically analyze the energy efficiency, performance, and privacy characteristics of models ranging from 0.5B to 9B parameters on a real-world Android device (Samsung Galaxy S25 Ultra). Leveraging non-intrusive, fine-grained power monitoring and mixed-precision inference, they uncover a “quantization-energy paradox”: model architecture—not quantization strategy—dominates energy consumption. Notably, Mixture-of-Experts architectures disrupt conventional scaling–energy relationships, and medium-scale models such as Qwen2.5-3B emerge as the optimal choice, balancing high output quality with energy efficiency, thereby offering practical deployment guidelines for on-device intelligence.

Existing system runtimes struggle to efficiently obtain application-level QoS metrics—such as tail latency and throughput—because these cannot be directly observed via hardware counters or kernel interfaces, leading to complex and costly feedback mechanisms. This work proposes the first non-intrusive eBPF-based framework that accurately infers application-layer QoS metrics using only kernel-observable events like system calls, combined with lightweight modeling techniques. The approach requires no application instrumentation or runtime feedback, is fully decoupled from the system runtime, and supports plug-and-play deployment. Evaluated across diverse latency-sensitive workloads, the framework achieves high correlation with ground-truth QoS metrics while significantly reducing application intrusiveness. The implementation has been open-sourced to facilitate adoption and further research.

This work addresses the challenge in self-powered streaming networks where dynamic power management, while energy-efficient, incurs switching delays that degrade throughput and hinder real-time signal processing. The paper presents the first cycle-based scheduling framework tailored to such networks, formulating a linear program to compute the maximum achievable throughput and a mixed-integer linear program to minimize energy consumption under throughput constraints. To efficiently explore the trade-off between energy and throughput, the authors introduce a novel “Hop and Skip” multi-objective search strategy that rapidly generates a high-quality Pareto frontier. Experimental results demonstrate that the proposed approach significantly accelerates design space exploration on both benchmark and random graphs, and in practical case studies, it achieves superior energy-throughput trade-offs compared to always-on or purely self-powered baselines.

This work addresses the challenge of ensuring service continuity for multi-stage industrial workflows in B5G/6G networks, where conventional per-request QoS mechanisms fall short. To overcome this limitation, the authors propose a capability-aware collaborative planning framework that proactively exposes sustainable QoS capabilities within a finite network planning window. Industrial applications leverage this foresight to map workflow phases and submit demand trajectories, enabling workflow-level, forward-looking joint evaluation and dynamic coordination updates. By integrating network capability modeling, demand mapping, and adaptive coordination, the approach transcends traditional request-granularity constraints. Experimental validation on a real B5G system and large-scale simulations demonstrates that the proposed method significantly enhances service continuity, reduces request rejection rates, and substantially improves workflow completion rates under high network loads.

Distributed scientific workflows often exhibit highly unpredictable behavior, making it challenging to satisfy Quality-of-Service (QoS) constraints such as execution time or resource limits. To address this issue, this work proposes QoSFlow, an interpretable modeling approach based on statistical sensitivity analysis that partitions the configuration space into regions of similar behavioral characteristics, enabling efficient and accurate QoS-aware scheduling. By integrating performance modeling with analytical reasoning, QoSFlow avoids the need for exhaustive empirical testing. Experimental evaluation on three representative workflows demonstrates that configurations recommended by QoSFlow outperform those from the best heuristic methods by an average of 27.38%, with consistently stable real-world performance. This study presents the first interpretable, high-precision optimization framework for QoS in distributed scientific workflows.

Traditional efficiency metrics struggle to accurately assess resource utilization in heterogeneous high-performance computing systems that combine CPUs and accelerators. This work extends the POP efficiency model by introducing a hardware-agnostic, host-device dual-branch hierarchical efficiency framework. It uniquely defines a multiplicative efficiency decomposition on the device side, symmetric to that on the host, separately capturing mixed execution/offload efficiency and device parallel efficiency. Implemented via the lightweight TALP monitoring library, the approach supports both runtime and post-mortem analysis and outputs results in human-readable and machine-readable formats. Experiments on synthetic benchmarks and three real-world HPC applications demonstrate that the proposed methodology effectively uncovers performance bottlenecks related to offloading, load balancing, and task scheduling, offering developers actionable insights for optimization.