Search NASA⌕ Search

SEARCH · Search NASA

Results for “Network Performance”

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 1,531 records · Page 85

Identification of aerospace acoustic sources using sparse distributed associative memory

A pattern recognition system has been developed to classify five different aerospace acoustic sources. In this paper the performance of two new classifiers, an associative memory classifier and a neural network classifier, is compared to the performance of a previously designed system. Sources are classified using features calculated from the time and frequency domain. Each classifier undergoes a training period where it learns to classify sources correctly based on a set of known sources. After training the classifier is tested with unknown sources. Results show that over 96 percent of sources were identified correctly with the new associative memory classifier. The neural network classifier identified over 81 percent of the sources correctly.

Scott, E. A.↗

ON POTASSIUM AND OTHER ABUNDANCE ANOMALIES OF RED GIANTS IN NGC 2419

Globular clusters are of paramount importance for testing theories of stellar evolution and early galaxy formation. Strong evidence for multiple populations of stars in globular clusters derives from observed abundance anomalies. A puzzling example is the recently detected Mg–K anticorrelation in NGC 2419. We perform Monte Carlo nuclear reaction network calculations to constrain the temperature–density conditions that gave rise to the elemental abundances observed in this elusive cluster. We find a correlation between stellar temperature and density values that provide a satisfactory match between simulated and observed abundances in NGC 2419 for all relevant elements (Mg, Si, K, Ca, Sc, Ti, and V). Except at the highest densities (ρ ≳ 10{sup 8} g cm{sup −3}), the acceptable conditions range from ≈100 MK at ≈10{sup 8} g cm{sup −3} to ≈200 MK at ≈10{sup −4} g cm{sup −3}. This result accounts for uncertainties in nuclear reaction rates and variations in the assumed initial composition. We review hydrogen-burning sites and find that low-mass stars, asymptotic giant branch (AGB) stars, massive stars, or supermassive stars cannot account for the observed abundance anomalies in NGC 2419. Super-AGB stars could be viable candidates for the polluter stars if stellar model parameters can be fine-tuned to produce higher temperatures. Novae, involving either CO or ONe white dwarfs, could be interesting polluter candidates, but a current lack of low-metallicity nova models precludes firmer conclusions. We also discuss whether additional constraints for the first-generation polluters can be obtained by future measurements of oxygen, or by evolving models of second-generation low-mass stars with a non-canonical initial composition.

79 ASTRONOMY AND ASTROPHYSICS↗

Genetic Algorithm for Optimization of Neural Networks for Bayesian Inference of Model Uncertainty

The objective of this work was to develop a genetic optimization algorithm that can design a neural network capable of producing uncertainty estimates along with predictions. This algorithm is necessary because the inclusion of uncertainty modeling in a neural network greatly complicates the network’s design space, making the development of a converging model extremely difficult and time consuming. The genetic algorithm presented in this work uses a number of value ranges for various configurable neural network parameters to create a randomly generated population of network architectures. The initially generated population is then evolved over the course of several generations, with the best performing models breeding to produce novel network configurations. Mutations are randomly applied to the network designs to facilitate the development of adaptations beneficial to the task being performed. An experiment was conducted to validate the proposed algorithm, in which the genetic optimizer was tasked with producing a neural network capable of predicting the sound pressure level (SPL) resulting from jet-surface interaction (JSI) noise. The data used for this task was generated at the NASA Glenn Research Center in the Aero-Acoustic Propulsion Laboratory. Starting with an initial population size of 35 randomly generated networks, and evolved over the course of 10 generations, the genetic algorithm produced a design able to predict SPL as a result of JSI noise within 0.272 dB, on average.

Genetic algorithm↗

Distributed simulation using a real-time shared memory network

The Advanced Control Technology Branch of the NASA Lewis Research Center performs research in the area of advanced digital controls for aeronautic and space propulsion systems. This work requires the real-time implementation of both control software and complex dynamical models of the propulsion system. We are implementing these systems in a distributed, multi-vendor computer environment. Therefore, a need exists for real-time communication and synchronization between the distributed multi-vendor computers. A shared memory network is a potential solution which offers several advantages over other real-time communication approaches. A candidate shared memory network was tested for basic performance. The shared memory network was then used to implement a distributed simulation of a ramjet engine. The accuracy and execution time of the distributed simulation was measured and compared to the performance of the non-partitioned simulation. The ease of partitioning the simulation, the minimal time required to develop for communication between the processors and the resulting execution time all indicate that the shared memory network is a real-time communication technique worthy of serious consideration.

Simon, Donald L.↗

Quantitative Assessment of Transportation Network Vulnerability with Dynamic Traffic Simulation Methods

Transportation networks are critical to the social and economic function of nations. Given the continuing increase in the populations of cities throughout the world, the criticality of transportation infrastructure is expected to increase. Thus, it is ever more important to mitigate congestion as well as to assess the impact disruptions would have on individuals who depend on transportation for their work and livelihood. Moreover, several government organizations are responsible for ensuring transportation networks are available despite the constant threat of natural disasters and terrorist activities. Most of the previous transportation network vulnerability research has been performed in the context of static traffic models, many of which are formulated as traditional optimization problems. However, transportation networks are dynamic because their usage varies over time. Thus, more appropriate methods to characterize the vulnerability of transportation networks should consider their dynamic properties. This paper presents a quantitative approach to assess the vulnerability of a transportation network to disruptions with methods from traffic simulation. Our approach can prioritize the critical links over time and is generalizable to the case where both link and node disruptions are of concern. We illustrate the approach through a series of examples. Our results demonstrate that the approach provides quantitative insight into the time varying criticality of links. Such an approach could be used as the objective function of less traditional optimization methods that use simulation and other techniques to evaluate the relative utility of a particular network defense to reduce vulnerability and increase resilience.

Shekar, Venkateswaran↗

Job Scheduling in a Heterogeneous Grid Environment

Computational grids have the potential for solving large-scale scientific problems using heterogeneous and geographically distributed resources. However, a number of major technical hurdles must be overcome before this potential can be realized. One problem that is critical to effective utilization of computational grids is the efficient scheduling of jobs. This work addresses this problem by describing and evaluating a grid scheduling architecture and three job migration algorithms. The architecture is scalable and does not assume control of local site resources. The job migration policies use the availability and performance of computer systems, the network bandwidth available between systems, and the volume of input and output data associated with each job. An extensive performance comparison is presented using real workloads from leading computational centers. The results, based on several key metrics, demonstrate that the performance of our distributed migration algorithms is significantly greater than that of a local scheduling framework and comparable to a non-scalable global scheduling approach.

Shan, Hong-Zhang↗

An evaluation of parallel optimization for OpenSolaris Network Stack

Computing is now shifting towards multiprocessing. The fundamental goal of multiprocessing is improved performance through the introduction of additional hardware threads or cores (referred to as 'cores' for simplicity). Modern network stacks can exploit parallel cores to allow either message-based parallelism or connection-based parallelism as a means to enhance performance. OpenSolaris has redesigned and parallelized to better utilize additional cores. Three special technologies, named Softring Set, Soft ring and Squeue are introduced in OpenSolaris for stack parallelization. In this paper, we study the OpenSolaris packet receiving process and its core parallelism optimization techniques. Experiment results show that these techniques allow OpenSolaris to achieve better network I/O performance in multiprocessing environments; however, network stack parallelization has also brought extra overheads for system. An effective and efficient network I/O optimization in multiprocessing environments is required to cross all levers of the network stack from network interface to application.

99 GENERAL AND MISCELLANEOUS↗

Understanding sources of organic aerosol during CalNex-2010 using the CMAQ-VBS

Abstract. Community Multiscale Air Quality (CMAQ) model simulations utilizing the traditional organic aerosol (OA) treatment (CMAQ-AE6) and a volatility basis set (VBS) treatment for OA (CMAQ-VBS) were evaluated against measurements collected at routine monitoring networks (Chemical Speciation Network (CSN) and Interagency Monitoring of Protected Visual Environments (IMPROVE)) and those collected during the 2010 California at the Nexus of Air Quality and Climate Change (CalNex) field campaign to examine important sources of OA in southern California. Traditionally, CMAQ treats primary organic aerosol (POA) as nonvolatile and uses a two-product framework to represent secondary organic aerosol (SOA) formation. CMAQ-VBS instead treats POA as semivolatile and lumps OA using volatility bins spaced an order of magnitude apart. The CMAQ-VBS approach underpredicted organic carbon (OC) at IMPROVE and CSN sites to a greater degree than CMAQ-AE6 due to the semivolatile POA treatment. However, comparisons to aerosol mass spectrometer (AMS) measurements collected at Pasadena, CA, indicated that CMAQ-VBS better represented the diurnal profile and primary/secondary split of OA. CMAQ-VBS SOA underpredicted the average measured AMS oxygenated organic aerosol (OOA, a surrogate for SOA) concentration by a factor of 5.2, representing a considerable improvement to CMAQ-AE6 SOA predictions (factor of 24 lower than AMS). We use two new methods, one based on species ratios (SOA/ΔCO and SOA/Ox) and another on a simplified SOA parameterization, to apportion the SOA underprediction for CMAQ-VBS to slow photochemical oxidation (estimated as 1.5 × lower than observed at Pasadena using −log(NOx : NOy)), low intrinsic SOA formation efficiency (low by 1.6 to 2 × for Pasadena), and low emissions or excessive dispersion for the Pasadena site (estimated to be 1.6 to 2.3 × too low/excessive). The first and third factors are common to CMAQ-AE6, while the intrinsic SOA formation efficiency for that model is estimated to be too low by about 7 × . From source-apportioned model results, we found most of the CMAQ-VBS modeled POA at the Pasadena CalNex site was attributable to meat cooking emissions (48 %, consistent with a substantial fraction of cooking OA in the observations). This is compared to 18 % from gasoline vehicle emissions, 13 % from biomass burning (in the form of residential wood combustion), and 8 % from diesel vehicle emissions. All "other" inventoried emission sources (e.g., industrial, point, and area sources) comprised the final 13 %. The CMAQ-VBS semivolatile POA treatment underpredicted AMS hydrocarbon-like OA (HOA) + cooking-influenced OA (CIOA) at Pasadena by a factor of 1.8 compared to a factor of 1.4 overprediction of POA in CMAQ-AE6, but it did capture the AMS diurnal profile of HOA and CIOA well, with the exception of the midday peak. Overall, the CMAQ-VBS with its semivolatile treatment of POA, SOA from intermediate volatility organic compounds (IVOCs), and aging of SOA improves SOA model performance (though SOA formation efficiency is still 1.6–2 × too low). However, continued efforts are needed to better understand assumptions in the parameterization (e.g., SOA aging) and provide additional certainty to how best to apply existing emission inventories in a framework that treats POA as semivolatile, which currently degrades existing model performance at routine monitoring networks. The VBS and other approaches (e.g., AE6) require additional work to appropriately incorporate IVOC emissions and subsequent SOA formation.

54 ENVIRONMENTAL SCIENCES↗

Performance evaluation of the NASA/KSC CAD/CAE and office automation LAN's

This study's objective is the performance evaluation of the existing CAD/CAE (Computer Aided Design/Computer Aided Engineering) network at NASA/KSC. This evaluation also includes a similar study of the Office Automation network, since it is being planned to integrate this network into the CAD/CAE network. The Microsoft mail facility which is presently on the CAD/CAE network was monitored to determine its present usage. This performance evaluation of the various networks will aid the NASA/KSC network managers in planning for the integration of future workload requirements into the CAD/CAE network and determining the effectiveness of the planned FDDI (Fiber Distributed Data Interface) migration.

Zobrist, George W.↗

A Highly Stretchable and Fatigue-Free Transparent Electrode Based on an In-Plane Buckled Au Nanotrough Network

Flexible transparent electrodes are an indispensable component of next-generation soft optoelectronics such as wearable electronics and electronic artificial skins (E-skins). Among the existing candidate materials, metal nanotrough networks exhibit optimal overall optoelectronic performance with impressive bendability. However, their further practical applications are hindered by their limited mechanical stretchability, which is highly desired in biointegrated systems. Here it is demonstrated that superior mechanical stretchability with tensile strains up to 300% can be achieved in Au nanotrough networks by introducing an in-plane sinusoidal wavy structure. For the first time it is shown that with a precisely tuned nanotrough geometry along with an optimized network configuration, the buckled Au nanotrough network can be repeatedly stretched to strains up to 120% for 100 000 cycles, exhibiting excellent fatigue performance. Such highly stretchable and fatigue-free (at a high tensile strain up to 100%) Au nanotrough networks present a competent sheet resistance (R sh ≈ 10 Ω sq -1 ) with high optical transparency (T = 91%), which are demonstrated to be highly compatible with human skins as conformal flexible transparent electrodes.

36 MATERIALS SCIENCE↗

Predicting runtime and resource utilization of jobs on integrated cloud and HPC systems

Recent advances in virtualization technologies used in cloud computing offer performance that closely approaches bare-metal levels. Combined with specialized instance types and high-speed networking services for cluster computing, cloud platforms have become a compelling option for high-performance computing (HPC). However, most current batch job schedulers in HPC systems are designed for homogeneous clusters and make decisions based on limited information about jobs and system status. Scientists typically submit computational jobs to these schedulers with a requested runtime that is often over- or under-estimated. More accurate runtime predictions can help schedulers make better decisions and reduce job turnaround times. Here, they can also support decisions about migrating jobs to the cloud to avoid long queue wait times in HPC systems.

97 MATHEMATICS AND COMPUTING↗

A Flexible Operation of Distributed Generation in Distribution Networks With Dynamic Boundaries

Distributed generators performing black start to form isolated microgrids offer a flexible and resilient solution to service restoration in distribution systems. Employing the dynamic microgrid concept, distributed generators can form isolated microgrids by changing their physical boundaries through smart switches and conventional circuit breakers. However, the flexibility of distributed generators has not been thoroughly investigated and utilised in existing works. To address this issue, this letter presents a new model to the reconfiguration formulation in active distribution networks considering different operation modes of distributed generators. Compared with the existing models, this study provides a new formulation supporting different DGs' operation modes and fully making use of their flexibility. Illustrative results on IEEE 34-test systems verify the effectiveness of the proposed model.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Assessment of the 802.11g Wireless Protocol for Lunar Surface Communications

Future lunar surface missions supporting the NASA Vision for Space Exploration will rely on wireless networks to transmit voice and data. The ad hoc network architecture is of particular interest since it does not require a complex infrastructure. In this report, we looked at data performance over an ad hoc network with varying distances between Apple AirPort wireless cards. We developed a testing program to transmit data packets at precise times and then monitored the receive time to characterize connection delay, packet loss, and data rate. Best results were received for wireless links of less than 75 ft, and marginally acceptable (25-percent) packet loss was received at 150 ft. It is likely that better results will be obtained on the lunar surface because of reduced radiofrequency interference; however, higher power transmitters or receivers will be needed for significant performance gains.

Chelmins, David T.↗

“Understanding Robustness Lottery”: A Geometric Visual Comparative Analysis of Neural Network Pruning Approaches

Deep learning approaches have provided state-of-the-art performance in many applications by relying on large and overparameterized neural networks. However, such networks are very brittle and are difficult to deploy on resource-limited platforms. Model pruning, i.e., reducing the size of the network, is a widely adopted strategy that can lead to a more robust and compact model. Many heuristics exist for model pruning, but our understanding of the pruning process remains limited due to the black-box nature of a neural network model. Empirical studies show that some heuristics improve performance whereas others can make models more brittle. Here, this work aims to shed light on how different pruning methods alter the network’s internal feature representation and the corresponding impact on model performance. To facilitate a comprehensive comparison and characterization of the high-dimensional model feature space, we introduce a visual geometric analysis of feature representations. We evaluated a set of critical geometric concepts decomposed from the commonly adopted classification loss and used them to design a visualization system to compare and highlight the impact of pruning on model performance and feature representation. The proposed tool provides an environment for an in-depth comparison of pruning methods and a comprehensive understanding of how the model responds to common data corruption. By leveraging the proposed visualization, machine learning researchers can reveal the similarities between pruning methods and redundancy in robustness evaluation benchmarks, obtain geometric insights about the differences between pruned models that achieve superior robustness performance, and identify samples that are robust or fragile to model pruning and common data corruption.

Li, Zhimin [Univ. of Utah, Salt Lake City, UT (Uni↗

Modeling the influence of preferential flow on the spatial variability and time-dependence of mineral weathering rates

Inferences of weathering rates from laboratory and field observations suggest significant scale and time-dependence. Preferential flow induced by heterogeneity (manifest as permeability variations or discrete fractures) has been suggested as one potential mechanism causing scale/time-dependence. In this paper, we present a quantitative evaluation of the influence of preferential flow on weathering rates using reactive transport modeling. Simulations were performed in discrete fracture networks (DFNs) and correlated random permeability fields (CRPFs), and compared to simulations in homogeneous permeability fields. The simulations reveal spatial variability in the weathering rate, multidimensional distribution of reactions zones, and the formation of rough weathering interfaces and corestones due to preferential flow. In the homogeneous fields and CRPFs, the domain-averaged weathering rate is initially constant as long as the weathering front is contained within the domain, reflecting equilibrium-controlled behavior. The behavior in the CRPFs was influenced by macrodispersion, with more spread-out weathering profiles, an earlier departure from the initial constant rate and longer persistence of weathering. DFN simulations exhibited a sustained time-dependence resulting from the formation of diffusion-controlled weathering fronts in matrix blocks, which is consistent with the shrinking core mechanism. A significant decrease in the domain-averaged weathering rate is evident despite high remaining mineral volume fractions, but the decline does not follow a math formula dependence, characteristic of diffusion, due to network scale effects and advection-controlled behavior near the inflow boundary. Finally, the DFN simulations also reveal relatively constant horizontally averaged weathering rates over a significant depth range, challenging the very notion of a weathering front.

54 ENVIRONMENTAL SCIENCES↗

CILogon: A federated X.509 certification authority for cyberinfrastructure logon

SUMMARY CILogon provides a federated X.509 certification authority for secure access to cyberinfrastructure such as the Extreme Science and Engineering Discovery Environment. CILogon relies on federated authentication (Security Assertion Markup Language (OASIS, Burlington, MA (USA)) and OpenID) for determining user identities when issuing certificates. Federated authentication enables users to obtain certificates using existing identities (university, Google, etc.). Federated authentication also enables CILogon to serve a national‐scale user community without requiring a large network of registration authorities performing manual user identification. CILogon supports multiple levels of assurance and custom interfaces for specific user communities. In this article, we introduce the CILogon service and describe experiences and lessons learned from the first 3 years of operation. © 2014 The Authors. Concurrency and Computation: Practice and Experience published by John Wiley & Sons Ltd.

Basney, Jim↗