skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: An approach to online network monitoring using clustered patterns

Abstract

Network traffic monitoring is a core element in network operations and management for various purposes such as anomaly detection, change detection, and fault/failure detection. In this study, we introduce a new approach to online monitoring using a pattern-based representation of the network traffic. Unlike the past online techniques limited to a single variable to summarize (e.g., sketch), the focus of this study is on capturing the network state from the multivariate attributes under consideration. To this end, we employ clustering with its benefit of the aggregation of multidimensional variables. The clustered result represents the state of the network with regard to the monitored variables, which can also be compared with the previously observed patterns visually and quantitatively. Finally, we demonstrate the proposed method with two popular use cases, one for estimating state changes and the other for identifying anomalous states, to confirm its feasibility.

Authors:
 [1];  [2];  [1];  [3]
  1. Texas A & M Univ., Commerce, TX (United States)
  2. Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
  3. Electronics and Telecommunications Research Inst. (ETRI), Daejeon (Korea, Republic of)
Publication Date:
Research Org.:
Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States); Electronics and Telecommunications Research Inst. (ETRI), Daejeon (Korea, Republic of)
Sponsoring Org.:
USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR) (SC-21); Ministry of Science, ICT and Future Planning (MSIP) of Korea
OSTI Identifier:
1379769
Grant/Contract Number:
AC02-05CH11231; B0101-15-1293
Resource Type:
Journal Article: Accepted Manuscript
Journal Name:
2017 International Conference on Computing, Networking and Communications, ICNC 2017
Additional Journal Information:
Journal Name: 2017 International Conference on Computing, Networking and Communications, ICNC 2017
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING; monitoring; quality of service; systems modeling; electric breakdown; computer crime; timing; system monitoring; computer network management; pattern clustering

Citation Formats

Kim, Jinoh, Sim, Alex, Suh, Sang C., and Kim, Ikkyun. An approach to online network monitoring using clustered patterns. United States: N. p., 2017. Web. doi:10.1109/ICCNC.2017.7876207.
Kim, Jinoh, Sim, Alex, Suh, Sang C., & Kim, Ikkyun. An approach to online network monitoring using clustered patterns. United States. doi:10.1109/ICCNC.2017.7876207.
Kim, Jinoh, Sim, Alex, Suh, Sang C., and Kim, Ikkyun. Mon . "An approach to online network monitoring using clustered patterns". United States. doi:10.1109/ICCNC.2017.7876207. https://www.osti.gov/servlets/purl/1379769.
@article{osti_1379769,
title = {An approach to online network monitoring using clustered patterns},
author = {Kim, Jinoh and Sim, Alex and Suh, Sang C. and Kim, Ikkyun},
abstractNote = {Network traffic monitoring is a core element in network operations and management for various purposes such as anomaly detection, change detection, and fault/failure detection. In this study, we introduce a new approach to online monitoring using a pattern-based representation of the network traffic. Unlike the past online techniques limited to a single variable to summarize (e.g., sketch), the focus of this study is on capturing the network state from the multivariate attributes under consideration. To this end, we employ clustering with its benefit of the aggregation of multidimensional variables. The clustered result represents the state of the network with regard to the monitored variables, which can also be compared with the previously observed patterns visually and quantitatively. Finally, we demonstrate the proposed method with two popular use cases, one for estimating state changes and the other for identifying anomalous states, to confirm its feasibility.},
doi = {10.1109/ICCNC.2017.7876207},
journal = {2017 International Conference on Computing, Networking and Communications, ICNC 2017},
number = ,
volume = ,
place = {United States},
year = {Mon Mar 13 00:00:00 EDT 2017},
month = {Mon Mar 13 00:00:00 EDT 2017}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Save / Share:
  • From May 2004, in the Institute of Nuclear Physics is installed Albanian Radiological Monitoring Network, in the framework of emergency monitoring in the territory of Albania. In this network, this is unique monitoring on-line system in our country. are included 5(five) monitoring stations, respectively in Tirane, Shkoder, Kukes, Korce and Vlore. The last four stations are near Albanian borders The network performs measures of ambient dose rate in a range from 5 nSv/h up to 10 Sv/h. For measurements are used detector of type VACUTEC 70045 A, which are calibrated in the Centre of Applied Nuclear Physics, University of Tirana,more » using standard radiation source Cs-137. This monitoring help to warn in real time the relative authorities, in case of radiological accidents of 5th degree (for example accidents in nuclear power plants, near Albanian territory).« less
  • Industrially ground surfaces often have a characteristic surface topography known as chatter marks. The surface finishing is mainly monitored by optical measurement techniques. In this work, the monitoring of an industrial belt grinding process with a light scattering sensor is presented. Although this technique is primarily applied for parametric surface roughness analysis, here it is shown that it enables also the measurement of the surface topography, i.e., the chatter marks occurring during the belt grinding process. In particular, it is proven that the light scattering method is appropriate to measure online the topography of chatter marks. Furthermore, the frequency analysismore » of the data reveals that the wavelength of chatter marks strongly depends on process parameters, such as the grinding speed.« less
  • Purpose: The purpose of this work is to improve the existing methodology to estimate neutron equivalent dose in organs during radiotherapy treatments, based on a Static Random Access Memory neutron detector (SRAMnd) [1]. This is possible thanks to the introduction of a new digital detector with improved characteristics, which is able to measure online the neutron fluence rate in the presence of an intense photon background [2]. Its reduced size, allows the direct estimation of doses in specific points inside an anthropomorphic phantom (NORMA) without using passive detectors as TLD or CR-39. This versatility will allow not only to improvemore » the existing models (generic abdomen and H and N [1]) but to generate more specific ones for any technique. Methods: The new Thermal Neutron Rate Detector (TNRD), based on a diode device sensitized to thermal neutrons, have been inserted in 16 points of the phantom. These points are distributed to infer doses to specific organs. Simultaneous measurements of these devices and a reference one, located in front of the gantry, have been performed for the mentioned generic treatments, in order to improve the existing model. Results: These new devices have shown more precise since they agree better with Monte Carlo simulations. The comparison of the thermal neutron fluence, measured with TNRD, and the existing models, converted from events to fluence, shows an average improvement of (3.90±3.37) % for H and N and (12.61±9.43) % for abdomen, normalized to the maximum value. Conclusion: This work indicates the potential of these new devices for more precise neutron equivalent dose estimation in organs, as a consequence of radiotherapy treatments. The simplicity of the process makes possible to establish more specific models that will provide a better dose estimation. References[1] Phys Med Biol 2012; 57:6167–6191.[2] A new active thermal neutron detector. Radiat. Prot. Dosim. (in press)« less
  • In the application of oil-air lubrication system, electrical capacitance tomography (ECT) provides a promising way for monitoring oil film in the pipelines by reconstructing cross sectional oil distributions in real time. While in the case of small diameter pipe and thin oil film, the thickness of the oil film is hard to be observed visually since the interface of oil and air is not obvious in the reconstructed images. And the existence of artifacts in the reconstructions has seriously influenced the effectiveness of image segmentation techniques such as level set method. Besides, level set method is also unavailable for onlinemore » monitoring due to its low computation speed. To address these problems, a modified level set method is developed: a distance regularized level set evolution formulation is extended to image two-phase flow online using an ECT system, a narrowband image filter is defined to eliminate the influence of artifacts, and considering the continuity of the oil distribution variation, the detected oil-air interface of a former image can be used as the initial contour for the detection of the subsequent frame; thus, the propagation from the initial contour to the boundary can be greatly accelerated, making it possible for real time tracking. To testify the feasibility of the proposed method, an oil-air lubrication facility with 4 mm inner diameter pipe is measured in normal operation using an 8-electrode ECT system. Both simulation and experiment results indicate that the modified level set method is capable of visualizing the oil-air interface accurately online.« less
  • In this paper, the generation of multi-clustered structure of self-organized neural network with different neuronal firing patterns, i.e., bursting or spiking, has been investigated. The initially all-to-all-connected spiking neural network or bursting neural network can be self-organized into clustered structure through the symmetric spike-timing-dependent plasticity learning for both bursting and spiking neurons. However, the time consumption of this clustering procedure of the burst-based self-organized neural network (BSON) is much shorter than the spike-based self-organized neural network (SSON). Our results show that the BSON network has more obvious small-world properties, i.e., higher clustering coefficient and smaller shortest path length than themore » SSON network. Also, the results of larger structure entropy and activity entropy of the BSON network demonstrate that this network has higher topological complexity and dynamical diversity, which benefits for enhancing information transmission of neural circuits. Hence, we conclude that the burst firing can significantly enhance the efficiency of clustering procedure and the emergent clustered structure renders the whole network more synchronous and therefore more sensitive to weak input. This result is further confirmed from its improved performance on stochastic resonance. Therefore, we believe that the multi-clustered neural network which self-organized from the bursting dynamics has high efficiency in information processing.« less