Search NASA⌕ Search

SEARCH · Search NASA

Results for “performance data”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 667 records · Page 37

CMS HGCAL ECON-D ASIC : Impact of CMOS fabrication process tuning on performance and radiation tolerance

The CMS experiment’s High Granularity Calorimeter (HGCAL) upgrade will replace CMS’s existing endcap calorimeters in preparation for the High Luminosity LHC. To effectively use over 6 million channels of this “imaging”calorimeter, CMS has developed two novel Endcap Concentrator (ECON) ASICs to perform data compression/selection on detector. The ECON-D ASIC operates on the 750kHz data path, and the ECON-T ASIC on the 40MHz trigger path. These 65 nm CMOS ASICs are radiation tolerant to 200 Mrad and low power, operating at less than 2.5 mW/channel. The first full-functionality prototype ECONs were produced and characterized in 2021-23, and an initial engineering run was performed in 2024. ECON-D radiation testing for the engineering run revealed that the chip’s internal SRAMs produce intermittent read errors for a non-negligible fraction of chips. Further investigation indicated that the SRAM performance is highly sensitive to the exact parameters of the CMOS fabrication process. To both study this process sensitivity and mitigate SRAM performance issues, twenty ECON wafers were produced in 2025 with a range of doping concentrations designed to tune the underlying transistor threshold voltage by 0%, 5%, 10%, and 15% from nominal. This talk will present first measurements of ECON-D performance, power consumption, and radiation tolerance for these four variations of CMOS process.

Syal, Chinar [Fermilab]↗

Effective sampling-driven performance tools for GPU-accelerated supercomputers, In: SC '13 Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis

Performance analysis of GPU-accelerated systems requires a system-wide view that considers both CPU and GPU components. In this paper, we describe how to extend system-wide, sampling-based performance analysis methods to GPU-accelerated systems. Since current GPUs do not support sampling, our implementation required careful coordination of instrumentation-based performance data collection on GPUs with sampling-based methods employed on CPUs. In addition, we also introduce a novel technique for analyzing systemic idleness in CPU/GPU systems. We demonstrate the effectiveness of our techniques with application case studies on Titan and Keeneland. Some of the highlights of our case studies are: 1) we improved performance for LULESH 1.0 by 30%, 2) we identified a hardware performance problem on Keeneland, 3) we identified a scaling problem in LAMMPS derived from CUDA initialization, and 4) we identified a performance problem that is caused by GPU synchronization operations that suffer delays due to blocking system calls.

Computer Science↗

Design and implementation of HuREX analysis supporting interface - 171

Human error is recognized as one of the main causes to affect to the socio-technical system including nuclear power plant. Therefore, reducing human error through human reliability analysis (HRA) is a critical issue. Moreover, sufficient and reliable human performance data collection is a prerequisite for ensuring the safety of nuclear power plants. Korea Atomic Energy Research Institute developed the HuREX (Human Reliability data Extraction) framework to collect and analyze the responses of human operators from simulator training records of nuclear power plant. Thus, the development and computer software interfaces of the database are essential to efficiently support human reliability analysis. In this study, OPERA database was developed in accordance with the following process. First, the essential data required for the HRA analysis (e.g., plant data, procedures, system and equipment info, operating crews, results of simulator experiments, and so forth) was stored in the database. In addition, standardized procedures and related items were stored in the database to support HRA quantification. A data structure was developed to store crew responses conducted by operators to cope with an emergency situation during simulator experiments. In addition, various interfaces that escalate efficiency of the HuREX applications were developed. The HuREX analysis support interface was developed with HuREX Video Analyzer for observation of simulator collection data (e.g., audio-visual records, plant parameters, operator's action logs) and HuREX Data Analyzer for analysis of operator's conducted task. By using the OPERA database and HuREX analysis support interface developed through this research, maximizing the efficiency of human reliability analysis based on simulator training data. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Sandia Cloud Forensics and Incident Response Platform v 1.0

As systems and devices become virtualized and deployed in the cloud, the hypervisor becomes an increasingly appropriate place to collect performance data, system state, system landscape, function calls, transaction traces, and other characteristics. We developed a method by which an introspection application may be coupled with a hypervisor to “reach into” the VM with minimal intrusiveness to collect data critical to the reconstruction of events, files, and operations. Such a capability is required to take advantage of the hypervisor as an instrumentation platform and to integrate that data with more traditional collection mechanisms. The concept of a VM serviced by a lightweight hypervisor is a relatively new paradigm for forensic practitioners. Traditional forensic techniques, based on assumptions that the filesystem was directly interacting with the hardware through an abstraction, afforded the forensic practitioner the assumption that there was nothing controlling the application below the filesystem. This is not the case when using virtualized technologies. Hypervisors can covertly monitor, introspect and interact with the guest in a transparent fashion. The problems of storage and collection of actionable data are exhausting. The current challenge is most hypervisors do not expose a useful application programming interface (API) at a sufficient level to do transparent, fine-grained and customizable introspection. Scalable VM instrumentation and introspection at an in-depth level requires fast handling of events, as well as direct access to VM state. Furthermore, deep introspection benefits greatly from the ability to gather data from the hardware during the VM’s exit to the hypervisor. All of this requires identical access to the system as the hypervisor itself; improper use of this ability could easily cause system instability. It is for this reason we believe that the hypervisor developers have been hesitant to grant this much control through their APIs. However, our approach leverages other means to collect and monitor the guests in a targeted fashion. Virtual machine introspection (VMI) is a technique used to monitor the runtime state of a system-level virtual machine. The runtime state can be may include processor registers, memory, disk, network, and any other hardware-level events A review of research literature and current VMI technologies exposed several limitations and tradeoffs in VMI approaches, including: the use of in-guest agents; kernel to user space transitions (dramatically slowing down processing); VMI tool pre-configuration requirements; hypervisor version lock-in or source code patching; reliance on operating system symbols; limited processor features due to hypervisor (even if the hardware could do more). To address these constraints, a VMI tool was envisioned to provide the cloud forensic capabilities while having as few of these limitations as possible. The Kernel-based Virtual Machine Introspection (KVMi) tool was developed for hypervisors on Intel’s x86-64 architecture. To meet performance, scoping and use-case demands, the follow criteria were applied to KVMi to support live, forensic data collection: • Shall not require in-guest agents. • Shall work with any recent version of modern hypervisors. • Shall introspect upon any guest (VM) running modern versions of Windows, Linux and OS X. • Shall have the ability to find and track operating system artifacts • Shall not require operating system symbol files. • Shall be able to fully handle VM-exits, bypassing execution of the hypervisor if necessary, to facilitate new features the hypervisor may not support. • Shall be compilable/loadable on a running system with standard build tools.

Stout, WilliamM.S.↗

Development and Production of a 201 MHz, 5.0 MW Peak Power Klystron

Communications & Power Industries LLC has designed and manufactured the VKP-8201A, a high peak power, high gain, VHF band klystron. The klystron operates at 201.25 MHz, with 5.0 MW peak output power, 34 kW average output power, and a gain of 36 dB. The klystron is designed to operate between 1.0 MW and 4.5 MW in the linear range of the transfer curve. The klystron utilizes a unique magnetic field which enables the use of a proven electron gun design with a larger electron beam requirement. Experimental and predicted performance data are compared.

43 PARTICLE ACCELERATORS↗

Quantify Degradation Rates and Mechanisms of PV Modules and Systems Installed in Florida Through Comprehensive Experimental and Theoretical Analysis (Poster)

The economic viability of photovoltaic (PV) technologies is inextricably tied to both the electrical performance and degradation rate of the PV systems, which are the generators of electrical power in PV systems. Over the past 15 years, performance data have been collected on numerous PV systems installed throughout the state of Florida and will be presented.

14 SOLAR ENERGY↗

Design of the Second-Generation ILC Marx Modulator

SLAC National Accelerator Laboratory (SLAC) has initiated a program to design and build a Marx-topology modulator to produce a relatively compact, low-cost, high availability klystron modulator for the International Linear Collider (ILC). Building upon the success of the P1 Marx, the SLAC P2 Marx is a second-generation modulator whose design further emphasizes the qualities of modularity and high-availability. This paper outlines highlights of this design and presents single-cell performance data obtained during the proof-of-concept phase of the project.

43 PARTICLE ACCELERATORS↗

Testing Results from the Low Activity Waste Pretreatment System (LAWPS) Full-Scale Ion Exchange Column Test and Engineering Scale Integrated System Test - 18634

The U.S. Department of Energy (DOE), Office of River Protection's (ORP) primary mission is to retrieve and treat Hanford's tank waste and close the Tank Farms to protect the Columbia River. Mixed radioactive waste is stored in 177 underground tanks at the Hanford Site as reported in DOE/ORP 2003- 02, Environmental Impact Statement for Retrieval, Treatment, and Disposal of Tank Waste and Closure of the Single Shell Tanks at the Hanford Site, Richland WA, Inventory and Source Term Data Package. The 177 underground tanks are estimated to contain about 56 million gallons of waste. The Low Activity Waste Pretreatment System (LAWPS) Project provides for the early production of immobilized low activity waste (ILAW) by feeding LAW directly from Tank Farms to the Hanford Waste Treatment Plant's (WTP's) LAW Facility, bypassing the Pretreatment Facility. Prior to the transfer of feed to the WTP LAW Vitrification Facility, tank supernatant waste will be pretreated in the LAWPS to meet the WTP LAW waste acceptance criteria. The LAWPS will also facilitate the return of secondary liquid wastes from the WTP LAW Vitrification Facility to the Tank Farms. In November of 2015 DOE's Tank Operations Contractor, Washington River Protection Solutions (WRPS), awarded a contract to the AECOM team to design, fabricate and operate two test platforms to generate performance data at two different scales and demonstrate technology maturation. The first platform models the ion exchange (IX) columns at full scale, and the second platform models the integrated flowsheet (cross flow filter (CFF) and IX) at 1/9. scale. Overarching technical objectives of performing the full-scale ion exchange test and the engineering-scale integrated test are to: 1. Test interactions between unit process operations; 2. Test removal of non-radioactive cesium from process stream; 3. Test removal of undissolved solids from process stream; 4. Confirm operational flow rate range and stability; 5. Confirm volumetric throughput of waste stream over flow operating range; 6. Confirm range of process temperatures over planned operating range; 7. Establish system hydraulics and process criteria; 8. Coordinate key control and monitoring components; 9. Establish treated product chemistry meets requirements (cesium, solids, and pH); and 10. Test the hazard control strategy for a no flow condition through the IX column. Data collected during testing will be used by WRPS to validate the preliminary design, inform the final design of the permanent plant equipment, and demonstrate that the critical technology elements have advanced to a technology readiness level of 6 (prototypical system validation in a relevant environment). Testing on both platforms started in fall of 2016 generating NQA-1 data. This paper covers test results from the first five phases of testing on the 1/9. scale integrated test platform. A further three phases of testing are currently being executed, these including a 5.6 M sodium simulant containing foulant, a high viscosity simulant and finally a run performed with a lower Cs concentration (all earlier tests were spiked to 4 x Cs to reduce volume and cost). Results from these tests were not available at the time of writing available. Results are also presented from the full scale IX testing. Eight test were performed covering column loading, elution and regeneration; column loading, elution and regeneration with upflow displacement; IX column loading and gas migration; resin removal; solids accumulation; and IX column emergency recirculation and vent. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Fundamentals of Liquid Tin Anode Solid Oxide Fuel Cell (LTA-SOFC) Operation

An alternative high temperature fuel cell system, called Liquid Tin Anode Solid Oxide Fuel Cell (LTA-SOFC) technology, is presently under consideration by NETL for the ability to directly convert coal. Before such a fuel cell concept can be considered in system studies, a detailed assessment for the electrochemical activity and SnO and O-atom diffusion within the liquid tin needs to be performed. In addition, the fundamental thermodynamic operation of such a concept needs to be properly analyzed. Initial research efforts to characterize the tin electrochemistry on a button cell at 900°C and with a tin thickness of 6mm showed a peak power density of 40 mW/cm2, and an activation energy for total resistivity of 185600 J/gm-mol. Higher power densities are expected with a more optimized electrolyte interface, and additional tests are being planned. A review of the liquid tin anode cell theoretical operation and performance data will be presented.

01 COAL, LIGNITE, AND PEAT↗

TU-F-CAMPUS-I-05: Investigation of An EMCCD Detector with Variable Gain in a Micro-CT System

Purpose: To investigate the performance of a newly built Electron Multiplying Charged Coupled Device (EMCCD) based Micro-CT system, with variable detector gain, using a phantom containing contrast agent of different concentrations. Methods: We built a micro- CT system with an EMCCD having 8 microns pixels and on-chip variable gain. We tested the system using a phantom containing five tubes filled with different iodine contrast solutions (30% to 70%). First, we scanned the phantom using various x-ray exposures values at 40 kVp and constant detector gain. Next, for the same tube currents, the detector gain was increased to maintain the air value of the projection image constant. A standard FDK algorithm was used to reconstruct the data. Performance was analyzed by comparing the signal-to-noise ratio (SNR) measurements for increased gain with those for the low constant gain at each exposure. Results: The high detector gain reconstructed data SNR was always greater than the low gain data SNR for all x-ray settings and for all iodine features. The largest increases were observed for low contrast features, 30% iodine concentration, where the SNR improvement approached 2. Conclusion: One of the first implementations of an EMCCD based micro- CT system was presented and used to image a phantom with various iodine solution concentrations. The analysis of the reconstructed volumes showed a significant improvement of the SNR especially for low contrast features. The unique on-chip gain feature is a substantial benefit allowing the use of the system at very low x-ray exposures per frame.Partial support: NIH grant R01EB002873 and Toshiba Medical Systems Corp. Partial support: NIH grant R01EB002873 and Toshiba Medical Systems Corp.

60 APPLIED LIFE SCIENCES↗

Savannah River Site H-Canyon Advancing Technologies for Remote Inspections - 20345

In 2017, the DOE Environmental Management Office of Technology Development (DOE-EM TD) sponsored the H-Canyon Advanced Technology Demonstration (ATD) to demonstrate to DOE facilities the value of using new commercial-off-the-shelf (COTS) and near-ready technologies to solve difficult problems and enhance worker safety. The DOE Savannah River Site (SRS) H-Canyon Air Exhaust Tunnel (HCAEX) inspection task was identified as representative of the hazardous, human denied environments which could benefit from advanced technologies. The HCAEX underground concrete tunnel is visually inspected biannually using a camera mounted on a remotely operated vehicle (ROV) designed and built by SRNL. While tunnel images have provided valuable visual information, it is desirable to have a higher order of understanding of the environment to support a more thorough structural integrity (SI) analysis and for long term planning purposes. As part of the ATD, the Concrete Integrated Product Team (CIPT) was formed to identify and evaluate available sensors and methods mature enough to remotely obtain tunnel concrete characterization data of high value and with a high probability of success. The team included SMEs and H-Canyon stakeholders in the field of concrete, nondestructive examination (NDE), structural integrity, sensors and remote systems from SRNL, SRNS, LANL, DOE-SR and the Army Corps of Engineering. The CIPT completed an in-depth identification of customer concrete inspection needs and potential technology solutions. Sensors and methods were evaluated on performance, data usefulness, cost and the feasibility of a successful deployment given the unique tunnel access challenges and environment. Two technologies were identified as promising by the CIPT for near term demonstration and evaluation: Lidar (Light Detection and Ranging) 3-dimensional (3D) mapping and remote robotic deployment of NDE instrumentation. Laser spectroscopy to characterize tunnel surface chemical changes was also of interest, but presently cost prohibitive. This paper will include a discussion of the two efforts underway to evaluate and implement the CIPT recommendations. First, the status of the November 2019 deployment of Lidar at a single location into the tunnel is presented. This initial deployment provided the team a learning curve and lessons learned on the challenges of tunnel deployment to include remote operation and data collection, stabilization of the sensor in high air flow (∼30 mph), ability to achieve a tolerance accuracy of 0.25-inches, and the probability to identify change in tunnel dimensions over time. Secondly, a discussion on the development of the Robotic Arm Concrete Inspection Test Bed capable of deploying NDE instruments to examine custom concrete forms will be presented. Concrete forms simulating the rough concrete surfaces, strength, composition and potential structural defects that can be found at our DOE EM facilities have been designed and built for the test bed. Two state-of-the art concrete NDE instruments have been identified as having potential to work on rough concrete walls, they are being tested and characterized as to their ability to provide desired structural integrity data to include wall thickness and defect identification on the developed test beams. Lastly, lessons learned, and the path forward will be presented. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Low Activity Waste Pretreatment System Bench-Scale Testing: Supporting Integrated Testing and Facility Safety Analyses - 17171

The Low Activity Waste Pretreatment System (LAWPS) is being designed to enable the direct feed of waste to the Hanford Tank Waste Treatment and Immobilization Plant (WTP) Low Activity Waste (LAW) facility to be immobilized. Prior to construction of the LAWPS, pilot-scale integrated testing of the key unit operations (crossflow filtration, ion exchange using spherical resorcinol-formaldehyde (sRF) resin) will be conducted by a team led by Washington River Protection Solutions (WRPS) to increase the technology maturation level of the facility’s critical technology elements. As a part of this effort, Pacific Northwest National Laboratory (PNNL) has conducted a series of bench-scale (or engineering-scale) tests to perform two major objectives: (1) support pilot-scale integrated testing of the LAWPS by supplying information or performance data in advance of operating the pilot-scale facility; and (2) collect data needed to establish or confirm assumptions/approaches planned for implementation in the LAWPS safety basis. The first objective was focused in two technical areas: developing simulants that are representative of expected waste feed and can be produced at larger scales, and using these simulants in a bench-scale crossflow filter to establish expected solid-liquid separation performance. The crossflow filter was also used to observe the efficacy (with respect to filter production rate) of selected operational strategies. The second objective also included two technical areas: measuring the effect of sRF resin on hydrogen generation rate under irradiation, and demonstrating that the planned hydrogen management approach is effective and robust. The hydrogen management strategy involves fluidization of the sRF resin bed in the ion exchange columns and recirculating the liquid, a scenario that is planned for testing at full column height. The full height tests at PNNL also supported full-scale IX column testing conducted as part of the technology maturation plan. The experimental approaches used at PNNL in these four technical areas are summarized and selected key preliminary results are provided.

cross flow filtration↗

Low Activity Waste Pretreatment System Bench-Scale Testing: Supporting Integrated Testing and Facility Safety Analyses - 17171

The Low Activity Waste Pretreatment System (LAWPS) is being designed to enable the direct feed of waste to the Hanford Tank Waste Treatment and Immobilization Plant (WTP) Low Activity Waste (LAW) facility to be immobilized. Prior to construction of the LAWPS, pilot-scale integrated testing of the key unit operations (crossflow filtration, ion exchange using spherical resorcinol-formaldehyde (sRF) resin) will be conducted by a team led by Washington River Protection Solutions (WRPS) to increase the technology maturation level of the facility's critical technology elements. As a part of this effort, Pacific Northwest National Laboratory (PNNL) has conducted a series of tests to perform two major objectives: (1) support pilot-scale integrated testing of the LAWPS by supplying information or performance data in advance of operating the pilot-scale facility (1/9. scale); and (2) collect data needed to establish or confirm assumptions/approaches planned for implementation in the LAWPS safety basis. The first objective was focused in two technical areas: developing simulants that are representative of expected waste feed and can be produced at larger scales, and using these simulants in a bench-scale crossflow filter to establish expected solid-liquid separation performance. The crossflow filter was also used to observe the efficacy (with respect to filter production rate) of selected operational strategies. The second objective also included two technical areas: measuring the effect of sRF resin on hydrogen generation rate under irradiation, and demonstrating that the planned hydrogen management approach is effective and robust. The hydrogen management strategy involves fluidization of the sRF resin bed in the ion exchange columns and recirculating the liquid, a scenario that is planned for testing at full column height. The full height tests at PNNL also supported full-scale IX column testing conducted as part of the technology maturation plan. The experimental approaches used at PNNL in these four technical areas are summarized and selected key preliminary results are provided. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Perturbation-Based Diagnosis of False Data Injection Attack Using Distributed Energy Resources

Modern smart grid relies on various sensor measurements for its operational control. In a successful false data injection attack, the attacker manipulates the measurements from the grid sensors such that undetected errors are introduced into the estimates of the system parameters leading to catastrophic situations. This paper proposes a novel perturbation based false data injection attack detection mechanism that utilizes inverter based distributed energy resources (DERs) to create low magnitude perturbation signal in the distribution system voltage that is inconsequential to the normal grid operation. Two voltage sensitivity analysis based algorithms are designed to identify the optimal set of DERs that can create the voltage perturbation signal of desired magnitude. An analytical method of voltage sensitivity analysis is used to compute the magnitude of voltage perturbation signal at each node in a computationally efficient manner. Then, a detection mechanism is developed that checks for the presence of the perturbation sequence in each sensor measurement. A sensor measurement is deemed authentic if the voltage perturbation signal is present in the data. In case of sensor malfunction or cyber-attack, the perturbation signal will not be present in the measurement data. Performance of the proposed attack detection mechanism is validated via simulation of the IEEE 69 bus test system.

24 POWER TRANSMISSION AND DISTRIBUTION↗

FY07 LDRD Final Report Heavy Quark Jet Tomography

We propose and develop a new signature, the measurement of hadron-electron correlations to measure energy loss of heavy quarks in the quark-gluon plasma. This measurements will be used in future analyses to quantify the energy densities created in collisions of heavy ions at the Relativistic Heavy Ion Collider (RHIC) at Brookhaven National Lab and the Large Hadron Collider (LHC) at CERN. In addition we develop and implement a computing model that will leverage LLNL expertise in cost-effective high performance computing to perform data analyses and simulations for the ALICE experiment at CERN.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

5G Energy FRAME: The Design and Implementation of Data, Model, and Use Case (Year 2 Report)

This report summarizes the Year 2 work of Pacific Northwest National Laboratory’s (PNNL’s) 5G Fabricated Resource and Asset Management Encompassment for energy infrastructure (Energy FRAME) project funded by the Department of Energy Office of Science’s Advanced Scientific Computing Research (ASCR) Program. In this report, newest 5G equipment testing results are presented, along with two 5G-enabled AI/ML examples for grid applications; in addition, the work flow of grid edge, cloud, and High Performance Computing (HPC) platform is introduced, to support and interface the cross-domain simulation for power system transmission, distribution, and communication networks. Last but not least, the outlook for Year 3 work and the overarching impact of 5G Energy FRAME work to a multitude of stakeholders are provided. Additional 5G performance data now is shared through the publicly available weblink, https://www.pnnl.gov/projects/5g-energy-frame/publications

24 POWER TRANSMISSION AND DISTRIBUTION↗

Experimental investigation on hydrogen cryogenic distillation equipped with package made by ICIT

ICIT (Institute for Cryogenics and Isotopic Technologies) has used its experience in cryogenic water distillation process to propose a similar process for hydrogen distillation that can be used in detritiation technologies. This process relies on the same packages but a stainless filling is tested instead of the phosphorous bronze filling used for water distillation. This paper presents two types of packages developed for hydrogen distillation, both have a stainless filling but it differs in terms of density, exchange surface and specific volume. Performance data have been obtained on laboratory scale. In order to determine the characteristics of the package, the installation was operated in the total reflux mode, for different flow rate for the liquid. There were made several experiments considering different operating conditions. Samples extracted at the top and bottom of cryogenic distillation column allowed mathematical processing to determine the separation performance. The experiments show a better efficiency for the package whose exchange surface was higher and there were no relevant differences between both packages as the operating pressure of the cryogenic column was increasing. For a complete characterization of the packages, future experiments will be considered to determine performance at various velocities in the column and their correlation with the pressure in the column. We plan further experiments to separate tritium from the mixture of isotopes DT, having in view that our goal is to apply this results to a detritiation plant.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Leveraging History to Predict Infrequent Abnormal Transfers in Distributed Workflows

Scientific computing heavily relies on data shared by the community, especially in distributed data-intensive applications. This research focuses on predicting slow connections that create bottlenecks in distributed workflows. In this study, we analyze network traffic logs collected between January 2021 and August 2022 at the National Energy Research Scientific Computing Center (NERSC). Based on the observed patterns, we define a set of features primarily based on history for identifying low-performing data transfers. Typically, there are far fewer slow connections on well-maintained networks, which creates difficulty in learning to identify these abnormally slow connections from the normal ones. We devise several stratified sampling techniques to address the class-imbalance challenge and study how they affect the machine learning approaches. Our tests show that a relatively simple technique that undersamples the normal cases to balance the number of samples in two classes (normal and slow) is very effective for model training. This model predicts slow connections with an F1 score of 0.926.

97 MATHEMATICS AND COMPUTING↗