The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Lorsqu'une panne se produit dans un élément du réseau, tel qu'un commutateur, un routeur et un serveur, les opérateurs de réseau doivent reconnaître l'impact sur le service, comme le temps de récupération après une panne ou la gravité de la panne, car l'impact sur le service est une information essentielle pour gérer les pannes. . Dans cet article, nous proposons un système de prédiction de l'impact des services (DeepSIP) basé sur l'apprentissage profond, qui prédit l'impact sur le service d'une défaillance du réseau dans un élément de réseau à l'aide d'un réseau neuronal convolutif (CNN) multimodal temporel. Plus précisément, DeepSIP prédit le temps de récupération après une panne et la perte de volume de trafic due à la panne d'un réseau sur la base des informations provenant des messages Syslog et du volume de trafic. Étant donné que le temps de récupération est une information utile pour un accord de niveau de service (SLA) et que la perte de volume de trafic est directement liée à la gravité de la panne, nous considérons le temps de récupération et la perte de volume de trafic comme l'impact sur le service. L'impact sur le service est difficile à prévoir, car il dépend des types de pannes de réseau et du volume de trafic au moment où la panne se produit. De plus, les éléments du réseau ne contiennent explicitement aucune information sur l’impact du service. Pour extraire le type de pannes de réseau et prédire l'impact sur le service, nous utilisons les messages Syslog et le volume de trafic passé. Cependant, les messages Syslog et le volume de trafic sont également difficiles à analyser car ces données sont multimodales, fortement corrélées et ont des dépendances temporelles. Pour extraire des fonctionnalités utiles à la prédiction, nous développons un CNN multimodal temporel. Nous avons évalué expérimentalement DeepSIP en termes de précision en le comparant à d'autres méthodes basées sur NN en utilisant des ensembles de données synthétiques et réels. Pour les deux ensembles de données, les résultats montrent que DeepSIP a surpassé les références.
Yoichi MATSUO
NTT Corp.
Tatsuaki KIMURA
Osaka University
Ken NISHIMATSU
NTT Corp.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Yoichi MATSUO, Tatsuaki KIMURA, Ken NISHIMATSU, "DeepSIP: A System for Predicting Service Impact of Network Failure by Temporal Multimodal CNN" in IEICE TRANSACTIONS on Communications,
vol. E104-B, no. 10, pp. 1288-1298, October 2021, doi: 10.1587/transcom.2020EBP3177.
Abstract: When a failure occurs in a network element, such as switch, router, and server, network operators need to recognize the service impact, such as time to recovery from the failure or severity of the failure, since service impact is essential information for handling failures. In this paper, we propose Deep learning based Service Impact Prediction system (DeepSIP), which predicts the service impact of network failure in a network element using a temporal multimodal convolutional neural network (CNN). More precisely, DeepSIP predicts the time to recovery from the failure and the loss of traffic volume due to the failure in a network on the basis of information from syslog messages and traffic volume. Since the time to recovery is useful information for a service level agreement (SLA) and the loss of traffic volume is directly related to the severity of the failure, we regard the time to recovery and the loss of traffic volume as the service impact. The service impact is challenging to predict, since it depends on types of network failures and traffic volume when the failure occurs. Moreover, network elements do not explicitly contain any information about the service impact. To extract the type of network failures and predict the service impact, we use syslog messages and past traffic volume. However, syslog messages and traffic volume are also challenging to analyze because these data are multimodal, are strongly correlated, and have temporal dependencies. To extract useful features for prediction, we develop a temporal multimodal CNN. We experimentally evaluated DeepSIP in terms of accuracy by comparing it with other NN-based methods by using synthetic and real datasets. For both datasets, the results show that DeepSIP outperformed the baselines.
URL: https://global.ieice.org/en_transactions/communications/10.1587/transcom.2020EBP3177/_p
Copier
@ARTICLE{e104-b_10_1288,
author={Yoichi MATSUO, Tatsuaki KIMURA, Ken NISHIMATSU, },
journal={IEICE TRANSACTIONS on Communications},
title={DeepSIP: A System for Predicting Service Impact of Network Failure by Temporal Multimodal CNN},
year={2021},
volume={E104-B},
number={10},
pages={1288-1298},
abstract={When a failure occurs in a network element, such as switch, router, and server, network operators need to recognize the service impact, such as time to recovery from the failure or severity of the failure, since service impact is essential information for handling failures. In this paper, we propose Deep learning based Service Impact Prediction system (DeepSIP), which predicts the service impact of network failure in a network element using a temporal multimodal convolutional neural network (CNN). More precisely, DeepSIP predicts the time to recovery from the failure and the loss of traffic volume due to the failure in a network on the basis of information from syslog messages and traffic volume. Since the time to recovery is useful information for a service level agreement (SLA) and the loss of traffic volume is directly related to the severity of the failure, we regard the time to recovery and the loss of traffic volume as the service impact. The service impact is challenging to predict, since it depends on types of network failures and traffic volume when the failure occurs. Moreover, network elements do not explicitly contain any information about the service impact. To extract the type of network failures and predict the service impact, we use syslog messages and past traffic volume. However, syslog messages and traffic volume are also challenging to analyze because these data are multimodal, are strongly correlated, and have temporal dependencies. To extract useful features for prediction, we develop a temporal multimodal CNN. We experimentally evaluated DeepSIP in terms of accuracy by comparing it with other NN-based methods by using synthetic and real datasets. For both datasets, the results show that DeepSIP outperformed the baselines.},
keywords={},
doi={10.1587/transcom.2020EBP3177},
ISSN={1745-1345},
month={October},}
Copier
TY - JOUR
TI - DeepSIP: A System for Predicting Service Impact of Network Failure by Temporal Multimodal CNN
T2 - IEICE TRANSACTIONS on Communications
SP - 1288
EP - 1298
AU - Yoichi MATSUO
AU - Tatsuaki KIMURA
AU - Ken NISHIMATSU
PY - 2021
DO - 10.1587/transcom.2020EBP3177
JO - IEICE TRANSACTIONS on Communications
SN - 1745-1345
VL - E104-B
IS - 10
JA - IEICE TRANSACTIONS on Communications
Y1 - October 2021
AB - When a failure occurs in a network element, such as switch, router, and server, network operators need to recognize the service impact, such as time to recovery from the failure or severity of the failure, since service impact is essential information for handling failures. In this paper, we propose Deep learning based Service Impact Prediction system (DeepSIP), which predicts the service impact of network failure in a network element using a temporal multimodal convolutional neural network (CNN). More precisely, DeepSIP predicts the time to recovery from the failure and the loss of traffic volume due to the failure in a network on the basis of information from syslog messages and traffic volume. Since the time to recovery is useful information for a service level agreement (SLA) and the loss of traffic volume is directly related to the severity of the failure, we regard the time to recovery and the loss of traffic volume as the service impact. The service impact is challenging to predict, since it depends on types of network failures and traffic volume when the failure occurs. Moreover, network elements do not explicitly contain any information about the service impact. To extract the type of network failures and predict the service impact, we use syslog messages and past traffic volume. However, syslog messages and traffic volume are also challenging to analyze because these data are multimodal, are strongly correlated, and have temporal dependencies. To extract useful features for prediction, we develop a temporal multimodal CNN. We experimentally evaluated DeepSIP in terms of accuracy by comparing it with other NN-based methods by using synthetic and real datasets. For both datasets, the results show that DeepSIP outperformed the baselines.
ER -